From midori at ebi.ac.uk Thu Oct 1 02:58:08 2009 From: midori at ebi.ac.uk (Midori Harris) Date: Thu, 1 Oct 2009 10:58:08 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <29af5e2d0909291052v3b9821ele28f6229ac2b7232@mail.gmail.com> <29af5e2d0909291139n125fe58bpfdc59b449affa481@mail.gmail.com> Message-ID: Hi Karen, Thanks. I think it'll work nicely if I announce that I'll do the merge when I'm back from vacation -- that'll give everyone three weeks. Watch the GO list ... Midori On Wed, 30 Sep 2009, Karen Christie wrote: > Hi Midori, > > I checked with our GO annotation team and they think two weeks is enough, but > are starting now rather than waiting till it goes through to avoid having no > function annotations for those genes once this term is gone from Function. > > I know the standard email states when there are annotations. Perhaps it could > also explicitly state when there is no comparable replacement term for > annotation. > > thanks for considering annotation needs in this ontology change! > > -Karen > > > > On Wed, 30 Sep 2009, Midori Harris wrote: > >> How long will you need? Especially if it's more than the usual two weeks, >> I'd like to get an alert out soon. >> >> m >> >> On Tue, 29 Sep 2009, Karen Christie wrote: >> >>> These two function terms were created with the intent to represent the two >>> catalytic activites that occur in large spliceosomal complexes. Based on >>> the current understanding, I would ascribe these functions to a complex as >>> a whole. There is not yet evidence that would allow ascribing either >>> catalytic activity to a smaller portion of either relevant complex and >>> they're certainly not catalyzed by a single gene product. >>> >>> We have recently revised the component ontology and now also represent the >>> two specific complexes that carry out each of these two catalytic >>> activities, these two terms and their child terms: >>> - catalytic step 1 spliceosome >>> - catalytic step 2 spliceosome >>> >>> In process, we have these terms: >>> - generation of catalytic spliceosome for first transesterification step >>> - generation of catalytic spliceosome for second transesterification step >>> >>> So, I think we are doing this (splicing) right (finally). In process, we >>> represent the big picture, including some specific stages like generation >>> of particular complexes in the cycle. In complex, we represent the various >>> identifiable complexes that researchers recognize. In function, we >>> represent only the two catalytic activities. >>> >>> As named/phrased, the existing function grouping term directly above the >>> two specific function terms (first/second spliceosomal transesterification >>> activity) basically just means "be involved with splicing", which I agree >>> is equivalent to the existing process term. So, I agree that the current >>> function term (GO:0031202) should be merged into the corresponding process >>> term, as proposed. >>> >>> If we do need a grouping term in the function ontology for these two >>> function terms, it should be a generalized function, something like >>> "transesterification activity", not something process specific as the >>> current one is. >>> >>> -Karen >>> >>> P.S. In terms of proceeding with this, we'll need to make sure we give >>> annotation groups an appropriate heads up since there will no function >>> term that has an equivalent meaning where you could do a direct replace. >>> The two child terms of GO:0031202 are more specific. Thus reannotation >>> will require reading for each gene to determine whether to annotate to >>> either of these two specific terms or to the root node for function. I >>> think it would be appropriate to have a longer lead time than two weeks to >>> allow groups to work on this before essentially removing the term from >>> function. SGD has about 60 genes annotated to this term, excluding >>> computational annotations. >>> >>> >>> >>> On Tue, 29 Sep 2009, Alan Ruttenberg wrote: >>> >>>> On Tue, Sep 29, 2009 at 2:18 PM, Karen Christie >>>> wrote: >>>> >>>>> I'd like to know you deal with multisubunit enzymes, where no one >>>>> particular gene product has the catalytic activity. >>>>> >>>> >>>> I think that the function is then ascribed to either the complex as a >>>> whole, >>>> or to the part of the complex that is responsible for the activity. >>>> >>>> The schema is: >>>> >>>> Complex >>>> has_part protein 1 >>>> has_part protein 2 >>>> ... >>>> >>>> Complex bears function (reverse function inheres in complex) >>>> >>>> function realized in process >>>> >>>> which implies: >>>> >>>> Complex participates in process. >>>> >>>> In addition, typically there will be other participants in the process. >>>> For >>>> example the mRNA that is being spliced. >>>> >>>> If we are at the point where we need to argue about which part of the >>>> complex bears the function then we are in good shape. >>>> >>>> -Alan >>>> >>>> >>>> >>>>> -Karen >>>>> >>>>> >>>>> >>>>> On Tue, 29 Sep 2009, Alan Ruttenberg wrote: >>>>> >>>>> The problem is that this still confuses function with process. >>>>> Functions >>>>>> are >>>>>> the things that the enzymes (using the term broadly) have that make >>>>>> the >>>>>> processes happen. If one goes into more detail, there are a variety of >>>>>> participants in the process (substrates, products, "cofactors"), but >>>>>> only >>>>>> one bearer of the function in a particular process. >>>>>> Every function has a corresponding process, its "realization". >>>>>> >>>>>> Functions are not parts of processes. The realizations of functions >>>>>> are. >>>>>> >>>>>> The functions have a physical basis in the structure of their bearers. >>>>>> >>>>>> The functions exist before any process happens. A molecule may have a >>>>>> function but never actually realize it. >>>>>> >>>>>> These distinctions have important consequences as people such as myself >>>>>> work >>>>>> on adding more detail to the GO. The processes get more detail by >>>>>> adding >>>>>> *participants*. Within those processes, when there is a molecule like >>>>>> an >>>>>> enzyme that functions in a particular way, it needs to be distinguished >>>>>> from >>>>>> the other participants in some way. The way that it is distinguished is >>>>>> by >>>>>> saying it bears a function. >>>>>> >>>>>> The heuristic that a molecular function is a "single step" process is >>>>>> misused, IMO. That it is a "single step" derives from the fact that in >>>>>> such >>>>>> descriptions there is a single molecule that has a function. The step >>>>>> is >>>>>> the >>>>>> process that surrounds execution of that function. >>>>>> >>>>>> As I said, getting this distinction clear is essential for future >>>>>> detailing >>>>>> of GO processes such as I and others do. Can we start to get this right >>>>>> here >>>>>> and now? >>>>>> >>>>>> -Alan >>>>>> >>>>>> >>>>>> >>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>> wrote: >>>>>> >>>>>> Hi, >>>>>>> >>>>>>> We have two SF items in which merging function and process terms looks >>>>>>> like >>>>>>> a viable solution to a problem. One is an overly-specific, single-step >>>>>>> "process" that is essentially equivalent to a "function" (in the GO >>>>>>> sense); >>>>>>> the other is a "function" term that is defined such that it doesn't >>>>>>> describe >>>>>>> a single activity, and is essentially the same as a process term. >>>>>>> >>>>>>> SF 2864212 - merge MF term >>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>> GO:0031202 >>>>>>> >>>>>>> into BP term >>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>> >>>>>>> For this one, we would merge the MF term into the BP term, and ensure >>>>>>> that >>>>>>> the merged term has is_a ancestry in the BP graph. I think that's all >>>>>>> we >>>>>>> would change. >>>>>>> >>>>>>> >>>>>>> >>>>>>> https://sourceforge.net/tracker/?func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>> >>>>>>> >>>>>>> SF 2864271 - merge BP term >>>>>>> >>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>> >>>>>>> into MF term >>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>> >>>>>>> >>>>>>> >>>>>>> https://sourceforge.net/tracker/?func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>> >>>>>>> In this case, the proposal is to merge the process term into the >>>>>>> function >>>>>>> term, ensuring that the merged term has is_a ancestry only in the MF >>>>>>> ontology,and has a part_of link to a process term. Specifically, the >>>>>>> merged >>>>>>> term would retain the name and ID from GO:0004479, and would be >>>>>>> part_of >>>>>>> translational initiation GO:0006413. >>>>>>> >>>>>>> Does anyone have any objections or other comments on these? >>>>>>> >>>>>>> thanks, >>>>>>> m >>>>>>> _______________________________________________ >>>>>>> Annotation mailing list >>>>>>> Annotation at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>> >>>>>>> >>>>>> >>>> >>> >> > From Judith.Blake at jax.org Thu Oct 1 05:33:23 2009 From: Judith.Blake at jax.org (Judith Blake) Date: Thu, 1 Oct 2009 08:33:23 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> Message-ID: Ok Ok I'll use the word 'functioning's' for the 'occurrants actions' represented in the molecular function ontology. Ontology-speak. Yes what is being implemented is what was originally proposed earlier. Clarifying that the gene product has the potential and the MF terms describes the specific action is what I was after. judy On 9/30/09 8:30 PM, "Suzanna Lewis" wrote: Alan has it right. Functions are not occurents, Function*ings* are. But I think that the solution that David, Karen, and Midori have proposed is fine. We've consciously collapsed (to avoid repeating things) the "Function*ings* are part of processes" into a single artifact, a relationship (i.e. their is no information artifact for Function*ings* itself, it is implicit) that is for now called "part_of" (if I remember correctly). The name of this relationship may not be ideal, but we are trying to be clear about its intended usage. Chris & David first proposed this at the St. Croix meeting. And I seem to remember that David and I talked about this (and many other things, such as the nature of the implied relationship between an instance and a type when anyone makes an annotation) again at the relation ontology meeting in Denver. -S On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake wrote: Functions are occurrants. A function describes an action, biochemical, structural, whatever. A function doesn't describe anything. Information artifact describe things. Sorry to be picky about words. It's my job. If you look at the papers on BFO you will see that functions are continuants and occurrents are disjoint from continuants. So functions are not occurrents. If the GO is using terminology at odds with BFO that needs to be fixed - we agreed at the last foundry meeting that we were all going to use a common upper level ontology. Functions are part of processes. Function*ings* (i.e. realizations of functions) are part of processes. A variety of structures, encoded in a molecule, have the potential to realize a function. Type error: processes don't have potential. (concluded by substitution "potential to realize a function" + function is_a process => "potential to realize a process". The error is that it is the functions that are the potentials. Potentials are realized. The realizations are processes. We discussed this very recently with Michael and he agreed. That is a property of the molecule, not of the function. What is the reference of "That"? A molecule can have encoded the potential of engaging in several functions. Absolutely. (as long as you rewrite "functions" -> "functionings" (i.e. processes).) A molecule can bear multiple functions is the way we would say this. The confusion I think is that of the relationship between the gene product and a molecular function, and the molecular function and the biological process. And what do you say these are. I (and BFO) are clear: The relation between a molecular function and a gene product is "inheres". The relation between molecular function and a biological process is "realized_by". -Alan Judy On 9/29/09 1:52 PM, "Alan Ruttenberg" wrote: The problem is that this still confuses function with process. Functions are the things that the enzymes (using the term broadly) have that make the processes happen. If one goes into more detail, there are a variety of participants in the process (substrates, products, "cofactors"), but only one bearer of the function in a particular process. Every function has a corresponding process, its "realization". Functions are not parts of processes. The realizations of functions are. The functions have a physical basis in the structure of their bearers. The functions exist before any process happens. A molecule may have a function but never actually realize it. These distinctions have important consequences as people such as myself work on adding more detail to the GO. The processes get more detail by adding *participants*. Within those processes, when there is a molecule like an enzyme that functions in a particular way, it needs to be distinguished from the other participants in some way. The way that it is distinguished is by saying it bears a function. The heuristic that a molecular function is a "single step" process is misused, IMO. That it is a "single step" derives from the fact that in such descriptions there is a single molecule that has a function. The step is the process that surrounds execution of that function. As I said, getting this distinction clear is essential for future detailing of GO processes such as I and others do. Can we start to get this right here and now? -Alan On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris wrote: Hi, We have two SF items in which merging function and process terms looks like a viable solution to a problem. One is an overly-specific, single-step "process" that is essentially equivalent to a "function" (in the GO sense); the other is a "function" term that is defined such that it doesn't describe a single activity, and is essentially the same as a process term. SF 2864212 - merge MF term RNA splicing factor activity, transesterification mechanism GO:0031202 into BP term RNA splicing, via transesterification reactions GO:0000375 For this one, we would merge the MF term into the BP term, and ensure that the merged term has is_a ancestry in the BP graph. I think that's all we would change. https://sourceforge.net/tracker/?func=detail&aid=2864212&group_id=36855&atid=440764 SF 2864271 - merge BP term conversion of met-tRNAf to fmet-tRNA GO:0001718 into MF term methionyl-tRNA formyltransferase activity GO:0004479 https://sourceforge.net/tracker/?func=detail&aid=2864271&group_id=36855&atid=440764 In this case, the proposal is to merge the process term into the function term, ensuring that the merged term has is_a ancestry only in the MF ontology,and has a part_of link to a process term. Specifically, the merged term would retain the name and ID from GO:0004479, and would be part_of translational initiation GO:0006413. Does anyone have any objections or other comments on these? thanks, m _______________________________________________ Annotation mailing list Annotation at geneontology.org http://fafner.stanford.edu/mailman/listinfo/annotation _______________________________________________ Annotation mailing list Annotation at geneontology.org http://fafner.stanford.edu/mailman/listinfo/annotation -------------- next part -------------- An HTML attachment was scrubbed... URL: From phismith at buffalo.edu Thu Oct 1 05:58:37 2009 From: phismith at buffalo.edu (Barry Smith) Date: Thu, 01 Oct 2009 08:58:37 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> Message-ID: <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> At 08:33 AM 10/1/2009, Judith Blake wrote: >Ok Ok I'll use the word 'functioning's' for the 'occurrants actions' >represented in the molecular function ontology. Ontology-speak. Great. Ideally, MF should be renamed the 'Molecular Functioning Ontology' This is not just ontological persnicketiness. There are already people who are annotating with both 'function' and 'functioning' e.g. using OBI to describe experiments, where it is important to record both the functions e.g. of lab instruments, and the processes in which they participate. The former exist (as potentials) even when the relevant instrument is switched off. There is also parallel work on physiology, distinguishing e.g. the function of the heart (to pump), from realizations of this function in functionings (pumpings), and both of these in turn from non-functional processes such as thumpings, fibrillatings, and so forth. If the GO were to entrench, now, a confusion of function and functioning it will not work work well with these activities in the future. Barry >Yes what is being implemented is what was originally proposed >earlier. Clarifying that the gene product has the potential and the >MF terms describes the specific action is what I was after. >judy > > >On 9/30/09 8:30 PM, "Suzanna Lewis" ><suzi at fruitfly.org> wrote: > >Alan has it right. Functions are not occurents, Function*ings* are. > >But I think that the solution that David, Karen, and Midori have >proposed is fine. We've consciously collapsed (to avoid repeating >things) the "Function*ings* are part of processes" into a single >artifact, a relationship (i.e. their is no information artifact for >Function*ings* itself, it is implicit) that is for now called >"part_of" (if I remember correctly). The name of this relationship >may not be ideal, but we are trying to be clear about its intended usage. > >Chris & David first proposed this at the St. Croix meeting. And I >seem to remember that David and I talked about this (and many other >things, such as the nature of the implied relationship between an >instance and a type when anyone makes an annotation) again at the >relation ontology meeting in Denver. > >-S > >On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: > > > >On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake ><Judith.Blake at jax.org> wrote: > >Functions are occurrants. A function describes an action, >biochemical, structural, whatever. > >A function doesn't describe anything. Information artifact describe things. >Sorry to be picky about words. It's my job. >If you look at the papers on BFO you will see that functions are >continuants and occurrents are disjoint from continuants. So >functions are not occurrents. If the GO is using terminology at odds >with BFO that needs to be fixed - we agreed at the last foundry >meeting that we were all going to use a common upper level ontology. > > >Functions are part of processes. > >Function*ings* (i.e. realizations of functions) are part of processes. > >A variety of structures, encoded in a molecule, have the potential >to realize a function. > >Type error: processes don't have potential. (concluded by >substitution "potential to realize a function" + function is_a >process => "potential to realize a process". > >The error is that it is the functions that are the potentials. >Potentials are realized. The realizations are processes. > >We discussed this very recently with Michael and he agreed. > >That is a property of the molecule, not of the function. > >What is the reference of "That"? > >A molecule can have encoded the potential of engaging in several functions. > >Absolutely. (as long as you rewrite "functions" -> "functionings" >(i.e. processes).) > >A molecule can bear multiple functions is the way we would say this. > >The confusion I think is that of the relationship between the gene >product and a molecular function, and the molecular function and the >biological process. > >And what do you say these are. I (and BFO) are clear: The relation >between a molecular function and a gene product is "inheres". The >relation between molecular function and a biological process is "realized_by". > >-Alan > > > >Judy > > > > >On 9/29/09 1:52 PM, "Alan Ruttenberg" ><alanruttenberg at gmail.com> wrote: > >The problem is that this still confuses function with process. >Functions are the things that the enzymes (using the term >broadly) have that make the processes happen. If one goes into more >detail, there are a variety of participants in the process >(substrates, products, "cofactors"), but only one bearer of the >function in a particular process. > >Every function has a corresponding process, its "realization". > >Functions are not parts of processes. The realizations of functions are. > >The functions have a physical basis in the structure of their bearers. > >The functions exist before any process happens. A molecule may have >a function but never actually realize it. > >These distinctions have important consequences as people such as >myself work on adding more detail to the GO. The processes get more >detail by adding *participants*. Within those processes, when there >is a molecule like an enzyme that functions in a particular way, it >needs to be distinguished from the other participants in some way. >The way that it is distinguished is by saying it bears a function. > >The heuristic that a molecular function is a "single step" process >is misused, IMO. That it is a "single step" derives from the fact >that in such descriptions there is a single molecule that has a >function. The step is the process that surrounds execution of that function. > >As I said, getting this distinction clear is essential for future >detailing of GO processes such as I and others do. Can we start to >get this right here and now? > >-Alan > > > >On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris ><midori at ebi.ac.uk> wrote: >Hi, > >We have two SF items in which merging function and process terms >looks like a viable solution to a problem. One is an >overly-specific, single-step "process" that is essentially >equivalent to a "function" (in the GO sense); the other is a >"function" term that is defined such that it doesn't describe a >single activity, and is essentially the same as a process term. > >SF 2864212 - merge MF term > RNA splicing factor activity, transesterification mechanism GO:0031202 > >into BP term > RNA splicing, via transesterification reactions GO:0000375 > >For this one, we would merge the MF term into the BP term, and >ensure that the merged term has is_a ancestry in the BP graph. I >think that's all we would change. > >https://sourceforge.net/tracker/?func=detail&aid=2864212&group_id=36855&atid=440764 > > >SF 2864271 - merge BP term > > conversion of met-tRNAf to fmet-tRNA GO:0001718 > >into MF term > methionyl-tRNA formyltransferase activity GO:0004479 > >https://sourceforge.net/tracker/?func=detail&aid=2864271&group_id=36855&atid=440764 > >In this case, the proposal is to merge the process term into the >function term, ensuring that the merged term has is_a ancestry only >in the MF ontology,and has a part_of link to a process term. >Specifically, the merged term would retain the name and ID from >GO:0004479, and would be part_of translational initiation GO:0006413. > >Does anyone have any objections or other comments on these? > >thanks, >m >_______________________________________________ >Annotation mailing list >Annotation at geneontology.org >http://fafner.stanford.edu/mailman/listinfo/annotation > > > >_______________________________________________ >Annotation mailing list >Annotation at geneontology.org >http://fafner.stanford.edu/mailman/listinfo/annotation > > From cjm at yuri.lbl.gov Sat Oct 3 07:52:16 2009 From: cjm at yuri.lbl.gov (Chris Mungall) Date: Sat, 03 Oct 2009 07:52:16 -0700 Subject: [Ontology-editors] problem with obo file Message-ID: ERROR: namespace-different-from-is_a-parent: GO:0070973 in , GO:0070972 in biological_process From cjm at berkeleybop.org Sat Oct 3 13:20:17 2009 From: cjm at berkeleybop.org (Chris Mungall) Date: Sat, 3 Oct 2009 13:20:17 -0700 Subject: [Ontology-editors] transport/secretion documentation & help with XPs Message-ID: http://www.geneontology.org/GO.process.guidelines.shtml#tpt States: x localization [p] establishment of x localization ---[i] establishment of x orientation ---[i] x movement ---[i] x secretion ---[i] x transport This is inconsistent with the current structure of GO, in which we consistently have x secretion is_a x transport E.g. serotonin secretion is_a serotonin transport I think the documentation should be changed to x localization [p] establishment of x localization ---[i] establishment of x orientation ---[i] x movement ---[i] x transport -----[i] x secretion ((A side comment: does anyone else find the use of the term "x movement" as being disjoint from "x transport" as unintuitive as I do? If fedex transport a package from the UK to the US, the package moves. Would "x propulsion" or "x motility" or "x directed movement" or "x movement" be better? It seems that some of these synonyms are already favoured in the GO.)) There are also 4 missing links from the rule: GO:0015774-polysaccharide transport GO:0033037- polysaccharide localization GO:0006869-lipid transport GO:0010876-lipid localization GO:0030050-vesicle transport along actin filament GO: 0051648-vesicle localization GO:0047496-vesicle transport along microtubule GO:0051648- vesicle localization (there may be more, the CHEBI xps are highly incomplete) I think we need to coordinate the documentation with the XPs and running the reasoner. The is_a rules are easy to check from within oboedit (just use assert-implied-links). The part_of rules are a little harder though. Also, whilst this documentation is very good it's actually quite incomplete with respect to the subtleties of the location terms. The XPs for this part of the ontology are in need of some work. I have a little exercise for us all to help improve these. In any localization process there may be multiple participants. Each participant plays a different role. Can you help come up with intuitive names for some of these roles? And in addition, names for the relationship between the process and the participant. At the moment, we have relations like results_in_transport_of. But we would like a more general relation that encompasses movement and localization. For example, in "ovarian nurse cell to oocyte transport" we have two named participants - an ovarian nurse cell and an oocyte. We have to distinguish between the roles played by each of the participants, so we can say that the ONC has the role of "starting point", the oocyte has the role of "destination". If we look at the destination we can also identify another participant - the ring canal acts as the "conduit". We could also say that the gene product(s) plays the role of transporter (or more generally, the "molecular agent"). in biological_process_xp_cell we have: [Term] id: GO:0007300 ! ovarian nurse cell to oocyte transport intersection_of: GO:0006810 ! transport intersection_of: OBO_REL:results_in_transport_from CL:0000026 ! nurse cell intersection_of: OBO_REL:results_in_transport_to CL:0000023 ! oocyte (we could also add a relationship to FBbt:00005409 but let's keep it simple for now) The definitions and relation hierarchy for these relations are in the xp/ directory Here is a summary of them. The general form is * * transportee: P results_in_transport_of X "X transport" X is in a different location at the end of P * destination: P results_in_transport_to X "__ transport into X" the transportee is located in X at the end of P * start location: P results_in_transport_from X "X to __ transport" the transportee is located in X at the start of P * obstacle : P results_in_transport_across X "transX transport" (e.g. transmembrane transport) the transportee is located on one side of X at the beginning of P, and the other side at the end * conduit : P results_in_transport_through X ? the transportee is located within X throughout P, and X connects the start location with the destination * guide : P results_in_transport_along X "__ transport along X" the transportee is located in the vicinity of X throughout P, and X connects the start location with the destination * container : P results_in_transport_within X "X-mediated transport" (?? do we use the term this way consistently) the transportee is located within X throughout P, and X is a transportee in the same process My questions are: * Does this make sense? * Can we think of better names for the roles? * Better names for the relationships? Yes they are long, this is because the sentence "(every) P R (a) X" should be semi-english. Another alternative would be names like has_destination, has_transportee, ... Thoughts? * Which of the above are appropriate for localization terms? I think just transportee and destination. Is there a more general name for the 2 relations here? From aji at ebi.ac.uk Mon Oct 5 04:32:25 2009 From: aji at ebi.ac.uk (Amelia Ireland) Date: Mon, 5 Oct 2009 12:32:25 +0100 Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: References: Message-ID: On Oct 3, 2009, at 9:20 PM, Chris Mungall wrote: > http://www.geneontology.org/GO.process.guidelines.shtml#tpt > > States: > > x localization > [p] establishment of x localization > ---[i] establishment of x orientation > ---[i] x movement > ---[i] x secretion > ---[i] x transport > > This is inconsistent with the current structure of GO, in which we > consistently have [snip] I doubt very much that the documentation in question has been updated since it was first written, which was many aeons ago (and unfortunately I think that is the case for most of the ontology docs)... > (A side comment: does anyone else find the use of the term "x > movement" as being disjoint from "x transport" as unintuitive as I > do? If fedex transport a package from the UK to the US, the package > moves. Would "x propulsion" or "x motility" or "x directed movement" > or "x movement" be better? It seems that some of these synonyms are > already favoured in the GO.) [snip] I presume that the transport / movement split was made to differentiate between changes in the location of X mediated by something versus changes in location where X moves under its own steam. The trouble is that we colloquially use 'movement' to apply to both situations, i.e. transport is a narrow synonym of movement. If there is a better word or phrase that could be employed in the place of 'movement'... maybe something involving 'self-propelled' / 'self- directed' / 'self-...'? -- Amelia Ireland GO Editorial Office http://www.berkeleybop.org || http://www.ebi.ac.uk Boycott Trader Joe's Red List seafood: http://traitorjoe.com From midori at ebi.ac.uk Mon Oct 5 06:30:22 2009 From: midori at ebi.ac.uk (Midori Harris) Date: Mon, 5 Oct 2009 14:30:22 +0100 (BST) Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: References: Message-ID: some responses interspersed ... On Sat, 3 Oct 2009, Chris Mungall wrote: > http://www.geneontology.org/GO.process.guidelines.shtml#tpt > > States: > > x localization > [p] establishment of x localization > ---[i] establishment of x orientation > ---[i] x movement > ---[i] x secretion > ---[i] x transport > > This is inconsistent with the current structure of GO, in which we > consistently have > > x secretion is_a x transport > > E.g. > > serotonin secretion is_a serotonin transport > > I think the documentation should be changed to > > x localization > [p] establishment of x localization > ---[i] establishment of x orientation > ---[i] x movement > ---[i] x transport > -----[i] x secretion I've made this small correction. No doubt the docs would benefit from much more of an overhaul ... > ((A side comment: does anyone else find the use of the term "x movement" as > being disjoint from "x transport" as unintuitive as I do? If fedex transport > a package from the UK to the US, the package moves. Would "x propulsion" or > "x motility" or "x directed movement" or "x movement" be better? It seems > that some of these synonyms are already favoured in the GO.)) There is no standard definition of 'movement' in GO, and there are only 38 'movement' BP terms (not counting regulation terms and 'perception of rate of movement'). I'm not sure why movement was included in the documentation, since it's not really used in a standardized structure as the documentation implies. I think transport is_a movement -- the GO standard definition is 'directed movement of ...', i.e. transport is movement that results in something going from point A to point B. Not all movement results in/is equivalent to transport. If I take a train from here to London, I'm being transported, but if I just jump up and down in one place I'm not. Furthermore, it doesn't look like the terms with 'movement' in the names consistently capture a distinction between movement and transport. From a spot check of parents and ancestors, some of them could simply be renamed to use 'transport', and others do mean movement that isn't transport. We could do a bit of cleanup to make everything consistent with transport is_a movement (although we would still have very few 'x movement' terms). > There are also 4 missing links from the > rule: > > GO:0015774-polysaccharide transport GO:0033037-polysaccharide > localization > GO:0006869-lipid transport GO:0010876-lipid localization > GO:0030050-vesicle transport along actin filament > GO:0051648-vesicle localization > GO:0047496-vesicle transport along microtubule GO:0051648-vesicle > localization I've added these. > (there may be more, the CHEBI xps are highly incomplete) > > I think we need to coordinate the documentation with the XPs and running the > reasoner. The is_a rules are easy to check from within oboedit (just use Sounds good, apart from the challenge of finding time to add documentation to our plates ... I'll have to think longer about the rest ... I like the relationship names, despite the length, because they're fairly close to self-explanatory. I'm less sure about the words for the roles ... 'transportee' is a horrible coinage, but I haven't come up with anything good yet ('cargo' has essentially the desired meaning, but would almost certainly be perceived as odd, if not wrong, for some transported things, e.g. ions). m [snip] From midori at ebi.ac.uk Mon Oct 5 06:33:32 2009 From: midori at ebi.ac.uk (Midori Harris) Date: Mon, 5 Oct 2009 14:33:32 +0100 (BST) Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: References: Message-ID: On Mon, 5 Oct 2009, Amelia Ireland wrote: > On Oct 3, 2009, at 9:20 PM, Chris Mungall wrote: > >> (A side comment: does anyone else find the use of the term "x movement" as >> being disjoint from "x transport" as unintuitive as I do? If fedex >> transport a package from the UK to the US, the package moves. Would "x >> propulsion" or "x motility" or "x directed movement" or "x movement" be >> better? It seems that some of these synonyms are already favoured in the >> GO.) > [snip] > > I presume that the transport / movement split was made to differentiate > between changes in the location of X mediated by something versus changes in > location where X moves under its own steam. The actual 'movement' terms and their definitions are, for the most part, not consistent with this distinction. (See my previous email for comments on how movement isn't used consistently anyway, but the movement terms don't describe self-propelled movement.) m From dph at informatics.jax.org Mon Oct 5 07:11:40 2009 From: dph at informatics.jax.org (David Hill) Date: Mon, 05 Oct 2009 10:11:40 -0400 Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: References: Message-ID: <4AC9FE9C.4050102@informatics.jax.org> We just need to keep in mind that movement also includes the movement of an organism or a cell. Midori Harris wrote: > some responses interspersed ... > > On Sat, 3 Oct 2009, Chris Mungall wrote: > >> http://www.geneontology.org/GO.process.guidelines.shtml#tpt >> >> States: >> >> x localization >> [p] establishment of x localization >> ---[i] establishment of x orientation >> ---[i] x movement >> ---[i] x secretion >> ---[i] x transport >> >> This is inconsistent with the current structure of GO, in which we >> consistently have >> >> x secretion is_a x transport >> >> E.g. >> >> serotonin secretion is_a serotonin transport >> >> I think the documentation should be changed to >> >> x localization >> [p] establishment of x localization >> ---[i] establishment of x orientation >> ---[i] x movement >> ---[i] x transport >> -----[i] x secretion > > I've made this small correction. No doubt the docs would benefit from > much more of an overhaul ... > >> ((A side comment: does anyone else find the use of the term "x >> movement" as being disjoint from "x transport" as unintuitive as I >> do? If fedex transport a package from the UK to the US, the package >> moves. Would "x propulsion" or "x motility" or "x directed movement" >> or "x movement" be better? It seems that some of these synonyms are >> already favoured in the GO.)) > > There is no standard definition of 'movement' in GO, and there are > only 38 'movement' BP terms (not counting regulation terms and > 'perception of rate of movement'). I'm not sure why movement was > included in the documentation, since it's not really used in a > standardized structure as the documentation implies. > > I think transport is_a movement -- the GO standard definition is > 'directed movement of ...', i.e. transport is movement that results in > something going from point A to point B. Not all movement results > in/is equivalent to transport. If I take a train from here to London, > I'm being transported, but if I just jump up and down in one place I'm > not. > > Furthermore, it doesn't look like the terms with 'movement' in the > names consistently capture a distinction between movement and > transport. From a spot check of parents and ancestors, some of them > could simply be renamed to use 'transport', and others do mean > movement that isn't transport. We could do a bit of cleanup to make > everything consistent with transport is_a movement (although we would > still have very few 'x movement' terms). > >> There are also 4 missing links from the > localization> rule: >> >> GO:0015774-polysaccharide transport GO:0033037-polysaccharide >> localization >> GO:0006869-lipid transport GO:0010876-lipid localization >> GO:0030050-vesicle transport along actin filament >> GO:0051648-vesicle localization >> GO:0047496-vesicle transport along microtubule GO:0051648-vesicle >> localization > > I've added these. > >> (there may be more, the CHEBI xps are highly incomplete) >> >> I think we need to coordinate the documentation with the XPs and >> running the reasoner. The is_a rules are easy to check from within >> oboedit (just use > > Sounds good, apart from the challenge of finding time to add > documentation to our plates ... > > I'll have to think longer about the rest ... I like the relationship > names, despite the length, because they're fairly close to > self-explanatory. I'm less sure about the words for the roles ... > 'transportee' is a horrible coinage, but I haven't come up with > anything good yet ('cargo' has essentially the desired meaning, but > would almost certainly be perceived as odd, if not wrong, for some > transported things, e.g. ions). > > m > > [snip] > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From jane at ebi.ac.uk Tue Oct 6 06:24:51 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Tue, 06 Oct 2009 14:24:51 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> References: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> Message-ID: <4ACB4523.1040706@ebi.ac.uk> I think the problem here is with trying to characterise MF as either *all* functions or *all* processes (i.e functionings) when in fact MF contains a combination of both. We've been back and forth over this many times and this seems to be the crux of it. Some MF terms fit happily as *parts* of processes (functionings or little processes as I prefer to call them ;-)) e.g 'allene-oxide cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein tag' does not unfold over time whichever way you look at it. I think the strategy of ignoring the existence of this latter class and saying everything in MF is a BFO process isn't really satisfactory in the long term. As I see it we have the following options to fix this: 1. Take away the word 'activity' from the term names and call everything in MF a BFO function. Make realised_in (or participates_in?) relations between MF and BP. We'd need to think carefully about how annotations would be propagated over realised_in. 2. Call everything in MF a BFO process and remove any terms that don't fit as processes e.g. protein tag. We lose a lot of annotations this way. We could possibly move them to a different ontology but would mean lots of re-annotation. 3. Divide MF into BFO processes and BFO functions, either both under MF or move the BFO processes into BP and keep MF for functions. The disadvantage here is that we wouldn't have the function for e.g. allene-oxide cyclase - we'd only have the process. You could I suppose make 'allene-oxide cyclase function' and 'allene-oxide cyclase activity' but that gets pretty confusing. 4. Ignore the problem and hope it goes away (it won't). I don't really like 2 or 4. Jane Barry Smith wrote: > At 08:33 AM 10/1/2009, Judith Blake wrote: >> Ok Ok I'll use the word 'functioning's' for the 'occurrants actions' >> represented in the molecular function ontology. Ontology-speak. > > Great. > Ideally, MF should be renamed the 'Molecular Functioning Ontology' > This is not just ontological persnicketiness. There are already people > who are annotating with both 'function' and 'functioning' e.g. using > OBI to describe experiments, where it is important to record both the > functions e.g. of lab instruments, and the processes in which they > participate. The former exist (as potentials) even when the relevant > instrument is switched off. > There is also parallel work on physiology, distinguishing e.g. the > function of the heart (to pump), from realizations of this function in > functionings (pumpings), and both of these in turn from non-functional > processes such as thumpings, fibrillatings, and so forth. > If the GO were to entrench, now, a confusion of function and > functioning it will not work work well with these activities in the > future. > Barry > >> Yes what is being implemented is what was originally proposed >> earlier. Clarifying that the gene product has the potential and the >> MF terms describes the specific action is what I was after. >> judy >> >> >> On 9/30/09 8:30 PM, "Suzanna Lewis" >> <suzi at fruitfly.org> wrote: >> >> Alan has it right. Functions are not occurents, Function*ings* are. >> >> But I think that the solution that David, Karen, and Midori have >> proposed is fine. We've consciously collapsed (to avoid repeating >> things) the "Function*ings* are part of processes" into a single >> artifact, a relationship (i.e. their is no information artifact for >> Function*ings* itself, it is implicit) that is for now called >> "part_of" (if I remember correctly). The name of this relationship >> may not be ideal, but we are trying to be clear about its intended >> usage. >> >> Chris & David first proposed this at the St. Croix meeting. And I >> seem to remember that David and I talked about this (and many other >> things, such as the nature of the implied relationship between an >> instance and a type when anyone makes an annotation) again at the >> relation ontology meeting in Denver. >> >> -S >> >> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >> >> >> >> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >> <Judith.Blake at jax.org> wrote: >> >> Functions are occurrants. A function describes an action, >> biochemical, structural, whatever. >> >> A function doesn't describe anything. Information artifact describe >> things. >> Sorry to be picky about words. It's my job. >> If you look at the papers on BFO you will see that functions are >> continuants and occurrents are disjoint from continuants. So >> functions are not occurrents. If the GO is using terminology at odds >> with BFO that needs to be fixed - we agreed at the last foundry >> meeting that we were all going to use a common upper level ontology. >> >> >> Functions are part of processes. >> >> Function*ings* (i.e. realizations of functions) are part of processes. >> >> A variety of structures, encoded in a molecule, have the potential to >> realize a function. >> >> Type error: processes don't have potential. (concluded by >> substitution "potential to realize a function" + function is_a >> process => "potential to realize a process". >> >> The error is that it is the functions that are the potentials. >> Potentials are realized. The realizations are processes. >> >> We discussed this very recently with Michael and he agreed. >> >> That is a property of the molecule, not of the function. >> >> What is the reference of "That"? >> >> A molecule can have encoded the potential of engaging in several >> functions. >> >> Absolutely. (as long as you rewrite "functions" -> "functionings" >> (i.e. processes).) >> >> A molecule can bear multiple functions is the way we would say this. >> >> The confusion I think is that of the relationship between the gene >> product and a molecular function, and the molecular function and the >> biological process. >> >> And what do you say these are. I (and BFO) are clear: The relation >> between a molecular function and a gene product is "inheres". The >> relation between molecular function and a biological process is >> "realized_by". >> >> -Alan >> >> >> >> Judy >> >> >> >> >> On 9/29/09 1:52 PM, "Alan Ruttenberg" >> <alanruttenberg at gmail.com> wrote: >> >> The problem is that this still confuses function with process. >> Functions are the things that the enzymes (using the term broadly) >> have that make the processes happen. If one goes into more detail, >> there are a variety of participants in the process (substrates, >> products, "cofactors"), but only one bearer of the function in a >> particular process. >> >> Every function has a corresponding process, its "realization". >> >> Functions are not parts of processes. The realizations of functions are. >> >> The functions have a physical basis in the structure of their bearers. >> >> The functions exist before any process happens. A molecule may have a >> function but never actually realize it. >> >> These distinctions have important consequences as people such as >> myself work on adding more detail to the GO. The processes get more >> detail by adding *participants*. Within those processes, when there >> is a molecule like an enzyme that functions in a particular way, it >> needs to be distinguished from the other participants in some way. >> The way that it is distinguished is by saying it bears a function. >> >> The heuristic that a molecular function is a "single step" process is >> misused, IMO. That it is a "single step" derives from the fact that >> in such descriptions there is a single molecule that has a function. >> The step is the process that surrounds execution of that function. >> >> As I said, getting this distinction clear is essential for future >> detailing of GO processes such as I and others do. Can we start to >> get this right here and now? >> >> -Alan >> >> >> >> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >> <midori at ebi.ac.uk> wrote: >> Hi, >> >> We have two SF items in which merging function and process terms >> looks like a viable solution to a problem. One is an overly-specific, >> single-step "process" that is essentially equivalent to a "function" >> (in the GO sense); the other is a "function" term that is defined >> such that it doesn't describe a single activity, and is essentially >> the same as a process term. >> >> SF 2864212 - merge MF term >> RNA splicing factor activity, transesterification mechanism GO:0031202 >> >> into BP term >> RNA splicing, via transesterification reactions GO:0000375 >> >> For this one, we would merge the MF term into the BP term, and ensure >> that the merged term has is_a ancestry in the BP graph. I think >> that's all we would change. >> >> https://sourceforge.net/tracker/?func=detail&aid=2864212&group_id=36855&atid=440764 >> >> >> >> SF 2864271 - merge BP term >> >> conversion of met-tRNAf to fmet-tRNA GO:0001718 >> >> into MF term >> methionyl-tRNA formyltransferase activity GO:0004479 >> >> https://sourceforge.net/tracker/?func=detail&aid=2864271&group_id=36855&atid=440764 >> >> >> In this case, the proposal is to merge the process term into the >> function term, ensuring that the merged term has is_a ancestry only >> in the MF ontology,and has a part_of link to a process term. >> Specifically, the merged term would retain the name and ID from >> GO:0004479, and would be part_of translational initiation GO:0006413. >> >> Does anyone have any objections or other comments on these? >> >> thanks, >> m >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> >> >> >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> >> > > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors -- Dr Jane Lomax GO Editorial Office EMBL-EBI Wellcome Trust Genome Campus Hinxton Cambridgeshire, UK CB10 1SD p: +44 1223 492516 f: +44 1223 494468 From jane at ebi.ac.uk Tue Oct 6 08:14:31 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Tue, 06 Oct 2009 16:14:31 +0100 Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: References: Message-ID: <4ACB5ED7.2040604@ebi.ac.uk> Hi Chris - * Does this make sense? Yes - I think this is great. * Can we think of better names for the roles? transportee, destination, start location all fine. obstacle - may be better as 'barrier'? conduit - I'm being a bit slow here, what are examples of this? guide - ok container - or vehicle? * Better names for the relationships? Yes they are long, this is because the sentence "(every) P R (a) X" should be semi-english. Another alternative would be names like has_destination, has_transportee, ... Thoughts? I find them quite intuitive. The only one that was confusing for me was: * container : P results_in_transport_within X "X-mediated transport" (?? do we use the term this way consistently) the transportee is located within X throughout P, and X is a transportee in the same process because I read it as intra-X transport (i.e. transport with an X). I'm having a hard time thinking of an alternative though - 'P results_in_transport_by X' doesn't convey that the transportee is enclosed by X. * Which of the above are appropriate for localization terms? I think just transportee and destination. Is there a more general name for the 2 relations here? I agree - just transportee and destination. Also think you can use the same relationship names because the def of localization is "Any process by which a cell, a substance, or a cellular entity, such as a protein complex or organelle, is transported to, and/or maintained in a specific location." Jane Chris Mungall wrote: > http://www.geneontology.org/GO.process.guidelines.shtml#tpt > > States: > > x localization > [p] establishment of x localization > ---[i] establishment of x orientation > ---[i] x movement > ---[i] x secretion > ---[i] x transport > > This is inconsistent with the current structure of GO, in which we > consistently have > > x secretion is_a x transport > > E.g. > > serotonin secretion is_a serotonin transport > > I think the documentation should be changed to > > x localization > [p] establishment of x localization > ---[i] establishment of x orientation > ---[i] x movement > ---[i] x transport > -----[i] x secretion > > ((A side comment: does anyone else find the use of the term "x > movement" as being disjoint from "x transport" as unintuitive as I do? > If fedex transport a package from the UK to the US, the package moves. > Would "x propulsion" or "x motility" or "x directed movement" or "x > movement" be better? It seems that some of these synonyms are already > favoured in the GO.)) > > There are also 4 missing links from the localization> rule: > > GO:0015774-polysaccharide transport GO:0033037-polysaccharide > localization > GO:0006869-lipid transport GO:0010876-lipid localization > GO:0030050-vesicle transport along actin filament > GO:0051648-vesicle localization > GO:0047496-vesicle transport along microtubule > GO:0051648-vesicle localization > > (there may be more, the CHEBI xps are highly incomplete) > > I think we need to coordinate the documentation with the XPs and > running the reasoner. The is_a rules are easy to check from within > oboedit (just use assert-implied-links). The part_of rules are a > little harder though. > > Also, whilst this documentation is very good it's actually quite > incomplete with respect to the subtleties of the location terms. > > The XPs for this part of the ontology are in need of some work. I have > a little exercise for us all to help improve these. In any > localization process there may be multiple participants. Each > participant plays a different role. Can you help come up with > intuitive names for some of these roles? And in addition, names for > the relationship between the process and the participant. At the > moment, we have relations like results_in_transport_of. But we would > like a more general relation that encompasses movement and localization. > > For example, in "ovarian nurse cell to oocyte transport" we have two > named participants - an ovarian nurse cell and an oocyte. We have to > distinguish between the roles played by each of the participants, so > we can say that the ONC has the role of "starting point", the oocyte > has the role of "destination". If we look at the destination we can > also identify another participant - the ring canal acts as the > "conduit". We could also say that the gene product(s) plays the role > of transporter (or more generally, the "molecular agent"). > > in biological_process_xp_cell we have: > > [Term] > id: GO:0007300 ! ovarian nurse cell to oocyte transport > intersection_of: GO:0006810 ! transport > intersection_of: OBO_REL:results_in_transport_from CL:0000026 ! nurse > cell > intersection_of: OBO_REL:results_in_transport_to CL:0000023 ! oocyte > > (we could also add a relationship to FBbt:00005409 but let's keep it > simple for now) > > The definitions and relation hierarchy for these relations are in the > xp/ directory > > Here is a summary of them. The general form is > > * > > > > * transportee: P results_in_transport_of X > "X transport" > X is in a different location at the end of P > > * destination: P results_in_transport_to X > "__ transport into X" > the transportee is located in X at the end of P > > * start location: P results_in_transport_from X > "X to __ transport" > the transportee is located in X at the start of P > > * obstacle : P results_in_transport_across X > "transX transport" (e.g. transmembrane transport) > the transportee is located on one side of X at the beginning of P, > and the other side at the end > > * conduit : P results_in_transport_through X > ? > the transportee is located within X throughout P, and X connects the > start location with the destination > > * guide : P results_in_transport_along X > "__ transport along X" > the transportee is located in the vicinity of X throughout P, and X > connects the start location with the destination > > * container : P results_in_transport_within X > "X-mediated transport" (?? do we use the term this way consistently) > the transportee is located within X throughout P, and X is a > transportee in the same process > > My questions are: > > * Does this make sense? > > * Can we think of better names for the roles? > > * Better names for the relationships? Yes they are long, this is > because the sentence "(every) P R (a) X" should be semi-english. > Another alternative would be names like has_destination, > has_transportee, ... Thoughts? > > * Which of the above are appropriate for localization terms? I think > just transportee and destination. Is there a more general name for the > 2 relations here? > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors -- Dr Jane Lomax GO Editorial Office EMBL-EBI Wellcome Trust Genome Campus Hinxton Cambridgeshire, UK CB10 1SD p: +44 1223 492516 f: +44 1223 494468 From cjm at berkeleybop.org Tue Oct 6 08:38:48 2009 From: cjm at berkeleybop.org (Chris Mungall) Date: Tue, 6 Oct 2009 08:38:48 -0700 Subject: [Ontology-editors] transport/secretion documentation & help with XPs In-Reply-To: <4ACB5ED7.2040604@ebi.ac.uk> References: <4ACB5ED7.2040604@ebi.ac.uk> Message-ID: On Oct 6, 2009, at 8:14 AM, Jane Lomax wrote: > Hi Chris - > > * Does this make sense? > > Yes - I think this is great. > > * Can we think of better names for the roles? > > transportee, destination, start location all fine. > > obstacle - may be better as 'barrier'? > > conduit - I'm being a bit slow here, what are examples of this? I was originally thinking of ring canal as an example. The synapse or space between cells in signaling fulfils the definition, but 'conduit' would be an odd word here. So perhaps 'conduit' if the cargo is traveling through the lumen or interior space of an enclosed component, and medium if it's open..? > guide - ok > > container - or vehicle? > > * Better names for the relationships? Yes they are long, this is > because the sentence "(every) P R (a) X" should be semi-english. > Another alternative would be names like has_destination, > has_transportee, ... Thoughts? > > I find them quite intuitive. The only one that was confusing for me > was: > > * container : P results_in_transport_within X > "X-mediated transport" (?? do we use the term this way consistently) > the transportee is located within X throughout P, and X is a > transportee in the same process > > because I read it as intra-X transport (i.e. transport with an X). > I'm having a hard time thinking of an alternative though - 'P > results_in_transport_by X' doesn't convey that the transportee is > enclosed by X. yes, within isn't so good. I guess results_in_transport_whilst_contained_within?? > * Which of the above are appropriate for localization terms? I think > just transportee and destination. Is there a more general name for > the 2 relations here? > > I agree - just transportee and destination. Also think you can use > the same relationship names because the def of localization is "Any > process by which a cell, a substance, or a cellular entity, such as > a protein complex or organelle, is transported to, and/or maintained > in a specific location." > > Jane > > > > Chris Mungall wrote: >> http://www.geneontology.org/GO.process.guidelines.shtml#tpt >> >> States: >> >> x localization >> [p] establishment of x localization >> ---[i] establishment of x orientation >> ---[i] x movement >> ---[i] x secretion >> ---[i] x transport >> >> This is inconsistent with the current structure of GO, in which we >> consistently have >> >> x secretion is_a x transport >> >> E.g. >> >> serotonin secretion is_a serotonin transport >> >> I think the documentation should be changed to >> >> x localization >> [p] establishment of x localization >> ---[i] establishment of x orientation >> ---[i] x movement >> ---[i] x transport >> -----[i] x secretion >> >> ((A side comment: does anyone else find the use of the term "x >> movement" as being disjoint from "x transport" as unintuitive as I >> do? If fedex transport a package from the UK to the US, the package >> moves. Would "x propulsion" or "x motility" or "x directed >> movement" or "x movement" be better? It seems that some of these >> synonyms are already favoured in the GO.)) >> >> There are also 4 missing links from the > localization> rule: >> >> GO:0015774-polysaccharide transport GO:0033037- >> polysaccharide localization >> GO:0006869-lipid transport GO:0010876-lipid localization >> GO:0030050-vesicle transport along actin filament GO: >> 0051648-vesicle localization >> GO:0047496-vesicle transport along microtubule GO:0051648- >> vesicle localization >> >> (there may be more, the CHEBI xps are highly incomplete) >> >> I think we need to coordinate the documentation with the XPs and >> running the reasoner. The is_a rules are easy to check from within >> oboedit (just use assert-implied-links). The part_of rules are a >> little harder though. >> >> Also, whilst this documentation is very good it's actually quite >> incomplete with respect to the subtleties of the location terms. >> >> The XPs for this part of the ontology are in need of some work. I >> have a little exercise for us all to help improve these. In any >> localization process there may be multiple participants. Each >> participant plays a different role. Can you help come up with >> intuitive names for some of these roles? And in addition, names for >> the relationship between the process and the participant. At the >> moment, we have relations like results_in_transport_of. But we >> would like a more general relation that encompasses movement and >> localization. >> >> For example, in "ovarian nurse cell to oocyte transport" we have >> two named participants - an ovarian nurse cell and an oocyte. We >> have to distinguish between the roles played by each of the >> participants, so we can say that the ONC has the role of "starting >> point", the oocyte has the role of "destination". If we look at the >> destination we can also identify another participant - the ring >> canal acts as the "conduit". We could also say that the gene >> product(s) plays the role of transporter (or more generally, the >> "molecular agent"). >> >> in biological_process_xp_cell we have: >> >> [Term] >> id: GO:0007300 ! ovarian nurse cell to oocyte transport >> intersection_of: GO:0006810 ! transport >> intersection_of: OBO_REL:results_in_transport_from CL:0000026 ! >> nurse cell >> intersection_of: OBO_REL:results_in_transport_to CL:0000023 ! oocyte >> >> (we could also add a relationship to FBbt:00005409 but let's keep >> it simple for now) >> >> The definitions and relation hierarchy for these relations are in >> the xp/ directory >> >> Here is a summary of them. The general form is >> >> * >> >> >> >> * transportee: P results_in_transport_of X >> "X transport" >> X is in a different location at the end of P >> >> * destination: P results_in_transport_to X >> "__ transport into X" >> the transportee is located in X at the end of P >> >> * start location: P results_in_transport_from X >> "X to __ transport" >> the transportee is located in X at the start of P >> >> * obstacle : P results_in_transport_across X >> "transX transport" (e.g. transmembrane transport) >> the transportee is located on one side of X at the beginning of P, >> and the other side at the end >> >> * conduit : P results_in_transport_through X >> ? >> the transportee is located within X throughout P, and X connects >> the start location with the destination >> >> * guide : P results_in_transport_along X >> "__ transport along X" >> the transportee is located in the vicinity of X throughout P, and >> X connects the start location with the destination >> >> * container : P results_in_transport_within X >> "X-mediated transport" (?? do we use the term this way consistently) >> the transportee is located within X throughout P, and X is a >> transportee in the same process >> >> My questions are: >> >> * Does this make sense? >> >> * Can we think of better names for the roles? >> >> * Better names for the relationships? Yes they are long, this is >> because the sentence "(every) P R (a) X" should be semi-english. >> Another alternative would be names like has_destination, >> has_transportee, ... Thoughts? >> >> * Which of the above are appropriate for localization terms? I >> think just transportee and destination. Is there a more general >> name for the 2 relations here? >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > > > -- > Dr Jane Lomax > GO Editorial Office > EMBL-EBI > Wellcome Trust Genome Campus > Hinxton > Cambridgeshire, UK > CB10 1SD > > p: +44 1223 492516 > f: +44 1223 494468 > > From suzi at fruitfly.org Tue Oct 6 11:45:52 2009 From: suzi at fruitfly.org (Suzanna Lewis) Date: Tue, 6 Oct 2009 11:45:52 -0700 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACB4523.1040706@ebi.ac.uk> References: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> <4ACB4523.1040706@ebi.ac.uk> Message-ID: On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: > I think the problem here is with trying to characterise MF as either > *all* functions or *all* processes (i.e functionings) > when in fact MF contains a combination of both. > > We've been back and forth over this many times and this seems to be > the crux of it. > > Some MF terms fit happily as *parts* of processes (functionings or > little processes as I prefer to call them ;-)) e.g 'allene-oxide > cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein > tag' does not unfold over time whichever way you look at it. I think > the strategy of ignoring the existence of this latter class and > saying everything in MF is a BFO process isn't really satisfactory > in the long term. > > As I see it we have the following options to fix this: > > 1. Take away the word 'activity' from the term names and call > everything in MF a BFO function. Make realised_in (or > participates_in?) relations between MF and BP. We'd need to think > carefully about how annotations would be propagated over realised_in. This is what was proposed in St. Croix and it makes sense to me, but it does lead us into some unnecessarily confusing areas (i.e. having yet another relationship to be added). That is why I am now convinced that #3 the best approach. > > 2. Call everything in MF a BFO process and remove any terms that > don't fit as processes e.g. protein tag. We lose a lot of > annotations this way. We could possibly move them to a different > ontology but would mean lots of re-annotation. Strongly against this one. > > 3. Divide MF into BFO processes and BFO functions, either both under > MF or move the BFO processes into BP and keep MF for functions. The > disadvantage here is that we wouldn't have the function for e.g. > allene-oxide cyclase - we'd only have the process. You could I > suppose make 'allene-oxide cyclase function' and 'allene-oxide > cyclase activity' but that gets pretty confusing. I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep both under MF. Least amount of change for users, while still solving the problem. There is also 3 (c) which Amelia proposed which is to divide and put the structural terms in an entirely separate new 4th branch, but I think a simple divide at the top of MF is enough. In any case can we please, please please, get rid of the string "activity". It adds nothing and is quite gratuitous. > > 4. Ignore the problem and hope it goes away (it won't). Nope, we've got to fix this. > > I don't really like 2 or 4. > > Jane > > > Barry Smith wrote: >> At 08:33 AM 10/1/2009, Judith Blake wrote: >>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>> actions' represented in the molecular function ontology. Ontology- >>> speak. >> >> Great. >> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >> This is not just ontological persnicketiness. There are already >> people who are annotating with both 'function' and 'functioning' >> e.g. using OBI to describe experiments, where it is important to >> record both the functions e.g. of lab instruments, and the >> processes in which they participate. The former exist (as >> potentials) even when the relevant instrument is switched off. >> There is also parallel work on physiology, distinguishing e.g. the >> function of the heart (to pump), from realizations of this function >> in functionings (pumpings), and both of these in turn from non- >> functional processes such as thumpings, fibrillatings, and so forth. >> If the GO were to entrench, now, a confusion of function and >> functioning it will not work work well with these activities in the >> future. >> Barry >> >>> Yes what is being implemented is what was originally proposed >>> earlier. Clarifying that the gene product has the potential and >>> the MF terms describes the specific action is what I was after. >>> judy >>> >>> >>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>> <suzi at fruitfly.org> wrote: >>> >>> Alan has it right. Functions are not occurents, Function*ings* are. >>> >>> But I think that the solution that David, Karen, and Midori have >>> proposed is fine. We've consciously collapsed (to avoid repeating >>> things) the "Function*ings* are part of processes" into a single >>> artifact, a relationship (i.e. their is no information artifact >>> for Function*ings* itself, it is implicit) that is for now called >>> "part_of" (if I remember correctly). The name of this relationship >>> may not be ideal, but we are trying to be clear about its intended >>> usage. >>> >>> Chris & David first proposed this at the St. Croix meeting. And I >>> seem to remember that David and I talked about this (and many >>> other things, such as the nature of the implied relationship >>> between an instance and a type when anyone makes an annotation) >>> again at the relation ontology meeting in Denver. >>> >>> -S >>> >>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>> >>> >>> >>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>> <Judith.Blake at jax.org> wrote: >>> >>> Functions are occurrants. A function describes an action, >>> biochemical, structural, whatever. >>> >>> A function doesn't describe anything. Information artifact >>> describe things. >>> Sorry to be picky about words. It's my job. >>> If you look at the papers on BFO you will see that functions are >>> continuants and occurrents are disjoint from continuants. So >>> functions are not occurrents. If the GO is using terminology at >>> odds with BFO that needs to be fixed - we agreed at the last >>> foundry meeting that we were all going to use a common upper level >>> ontology. >>> >>> >>> Functions are part of processes. >>> >>> Function*ings* (i.e. realizations of functions) are part of >>> processes. >>> >>> A variety of structures, encoded in a molecule, have the potential >>> to realize a function. >>> >>> Type error: processes don't have potential. (concluded by >>> substitution "potential to realize a function" + function is_a >>> process => "potential to realize a process". >>> >>> The error is that it is the functions that are the potentials. >>> Potentials are realized. The realizations are processes. >>> >>> We discussed this very recently with Michael and he agreed. >>> >>> That is a property of the molecule, not of the function. >>> >>> What is the reference of "That"? >>> >>> A molecule can have encoded the potential of engaging in several >>> functions. >>> >>> Absolutely. (as long as you rewrite "functions" -> >>> "functionings" (i.e. processes).) >>> >>> A molecule can bear multiple functions is the way we would say this. >>> >>> The confusion I think is that of the relationship between the gene >>> product and a molecular function, and the molecular function and >>> the biological process. >>> >>> And what do you say these are. I (and BFO) are clear: The relation >>> between a molecular function and a gene product is "inheres". The >>> relation between molecular function and a biological process is >>> "realized_by". >>> >>> -Alan >>> >>> >>> >>> Judy >>> >>> >>> >>> >>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>> <alanruttenberg at gmail.com> wrote: >>> >>> The problem is that this still confuses function with process. >>> Functions are the things that the enzymes (using the term >>> broadly) have that make the processes happen. If one goes into >>> more detail, there are a variety of participants in the process >>> (substrates, products, "cofactors"), but only one bearer of the >>> function in a particular process. >>> >>> Every function has a corresponding process, its "realization". >>> >>> Functions are not parts of processes. The realizations of >>> functions are. >>> >>> The functions have a physical basis in the structure of their >>> bearers. >>> >>> The functions exist before any process happens. A molecule may >>> have a function but never actually realize it. >>> >>> These distinctions have important consequences as people such as >>> myself work on adding more detail to the GO. The processes get >>> more detail by adding *participants*. Within those processes, when >>> there is a molecule like an enzyme that functions in a particular >>> way, it needs to be distinguished from the other participants in >>> some way. The way that it is distinguished is by saying it bears a >>> function. >>> >>> The heuristic that a molecular function is a "single step" process >>> is misused, IMO. That it is a "single step" derives from the fact >>> that in such descriptions there is a single molecule that has a >>> function. The step is the process that surrounds execution of that >>> function. >>> >>> As I said, getting this distinction clear is essential for future >>> detailing of GO processes such as I and others do. Can we start to >>> get this right here and now? >>> >>> -Alan >>> >>> >>> >>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>> <midori at ebi.ac.uk> wrote: >>> Hi, >>> >>> We have two SF items in which merging function and process terms >>> looks like a viable solution to a problem. One is an overly- >>> specific, single-step "process" that is essentially equivalent to >>> a "function" (in the GO sense); the other is a "function" term >>> that is defined such that it doesn't describe a single activity, >>> and is essentially the same as a process term. >>> >>> SF 2864212 - merge MF term >>> RNA splicing factor activity, transesterification mechanism GO: >>> 0031202 >>> >>> into BP term >>> RNA splicing, via transesterification reactions GO:0000375 >>> >>> For this one, we would merge the MF term into the BP term, and >>> ensure that the merged term has is_a ancestry in the BP graph. I >>> think that's all we would change. >>> >>> >> >https://sourceforge.net/tracker/? >>> func=detail&aid=2864212&group_id=36855&atid=440764 >>> >>> >>> SF 2864271 - merge BP term >>> >>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>> >>> into MF term >>> methionyl-tRNA formyltransferase activity GO:0004479 >>> >>> >> >https://sourceforge.net/tracker/? >>> func=detail&aid=2864271&group_id=36855&atid=440764 >>> >>> In this case, the proposal is to merge the process term into the >>> function term, ensuring that the merged term has is_a ancestry >>> only in the MF ontology,and has a part_of link to a process term. >>> Specifically, the merged term would retain the name and ID from GO: >>> 0004479, and would be part_of translational initiation GO:0006413. >>> >>> Does anyone have any objections or other comments on these? >>> >>> thanks, >>> m >>> _______________________________________________ >>> Annotation mailing list >>> Annotation at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/annotation >>> >>> >>> >>> _______________________________________________ >>> Annotation mailing list >>> Annotation at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/annotation >>> >>> >> >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > > > -- > Dr Jane Lomax > GO Editorial Office > EMBL-EBI > Wellcome Trust Genome Campus > Hinxton > Cambridgeshire, UK > CB10 1SD > > p: +44 1223 492516 > f: +44 1223 494468 > > From aji at ebi.ac.uk Tue Oct 6 13:38:01 2009 From: aji at ebi.ac.uk (Amelia Ireland) Date: Tue, 6 Oct 2009 21:38:01 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> <4ACB4523.1040706@ebi.ac.uk> Message-ID: On Oct 6, 2009, at 7:45 PM, Suzanna Lewis wrote: > On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: [snip] > There is also 3 (c) which Amelia proposed which is to divide and put > the structural terms in an entirely separate new 4th branch, but I > think a simple divide at the top of MF is enough. I was actually asking whether the 'structural constituent of XXX' could be put in the cellular component ontology, as the term name suggests an element of a cell. Apologies if I didn't make that very clear! -- Amelia Ireland GO Editorial Office http://www.berkeleybop.org || http://www.ebi.ac.uk Boycott Trader Joe's Red List seafood: http://traitorjoe.com From kchris at genome.stanford.edu Tue Oct 6 15:06:35 2009 From: kchris at genome.stanford.edu (Karen Christie) Date: Tue, 6 Oct 2009 15:06:35 -0700 (PDT) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4AE49577-109D-4104-B40C-5E10FA5CE1B2@fruitfly.org> <20091001130015.284C65B0013@mweb1.acsu.buffalo.edu> <4ACB4523.1040706@ebi.ac.uk> Message-ID: To me, it would be redundant to put 'structural constituent of XXX' in the component ontology when XXX (e.g. flagellum, cell wall) is already in the component ontology. The cellular component is the thing itself. The 'structural constituent' portion only seems relevant in terms of representing a function of gene products that make up certain structures, so I don't see how it would fit to put 'structural constituent of xxx' in component. -Karen On Tue, 6 Oct 2009, Amelia Ireland wrote: > On Oct 6, 2009, at 7:45 PM, Suzanna Lewis wrote: >> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: > [snip] >> There is also 3 (c) which Amelia proposed which is to divide and put the >> structural terms in an entirely separate new 4th branch, but I think a >> simple divide at the top of MF is enough. > > I was actually asking whether the 'structural constituent of XXX' could be > put in the cellular component ontology, as the term name suggests an element > of a cell. Apologies if I didn't make that very clear! > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors From dph at informatics.jax.org Wed Oct 7 05:54:17 2009 From: dph at informatics.jax.org (David Hill) Date: Wed, 07 Oct 2009 08:54:17 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: Message-ID: <4ACC8F79.2060808@informatics.jax.org> Hi Everyone, I think the issue with the problematic function terms is that we described them as roles because we didn't take enough time to carefully think about what the gene products were actually doing. When I look at them, they are almost always cases where we wanted to assign a function because we knew the product functioned in a process. I think we can just define them as functions involved in processes, sometimes the functions are generic and sometimes they are not. Transcription factor activity is and example of one that is not. It is_a 'DNA binding activity' involved in 'regulation of transcription'. Here is what a biologist wants to know: Gene product X has the ability to carry out some kind of action (function), either passively (like actin) or actively (like histidine kinase). In the proper biological context, the gene product will carry out that action and that will be part of a biological process. In some cases, there is a requirement for more than one gene product for this elemental action to be carried out. The biological process describes the overall objective of the action. I cannot think of any case where a single gene product acts and has a biological effect, that is succeeds in achieving a biological objective. If that were the case, I think biology wouldn't work, because there is always some background activity as long as a molecule and substrates are present. This is the distinction between a function and a process in the misunderstood 'single step processes'. Even phosphorylation as a biological objective (that has some effect on a biological system) is the result of the action of the activity of many kinase molecules, or the repeated activity of individual kinase molecules. So there are three ways we can represent this: 1) gene product MFing biological process In this case the gene product is a continuant. The relationship describes the relationship between the continuant and the MF which is an occurent. The biological process is an occurent. 2) gene product MF MFing biological process In this case the gene product is a continuant. The MF is a continuant that is an enduring quality of the gene product. The MFing is an occurent that is the realization of the MF. The biological process is an occurent. 3) gene product MF BP In this case the gene product is a continuant. The MF is a continuant that is an enduring quality of the gene product. The BP is the occurent that results from the realization of the MF. I think that the most rigorous way to represent this would be #2. But, I think the most practical way would be #1. When I talk to biologists, both annotators and non-annotators, their eyes glaze over as soon as I mention functions and functionings. The bottom line is they don't care about it, they don't care to think about it, and they are confused by it. They are confused because in the absence of the gene product <>MF relationship, all they care about is a function being executed as part of a biological process. They could care less about the neurotransmitters that never bind a receptor and have a potential to execute a function, but never execute it. However, if we get rid of the MF<>BP side, they care about what the potential of the gene products are. So #1 works for this view. The MF being an occurent is the intuitive way that we think about this. MFs being parts of processes fits with the view I expressed above. #3 is not practical because it would require putting every MF into the BP ontology as an occurent. This would essentially recapitulate the entire MF ontology in BP. It is much easier for us to use #1. The key is in the relationship which combines the ideas of 'has potential' and 'carries out'. My 2 cents. David Judith Blake wrote: > I agree with Suzi, I support the continuation of MF, always have. > I don't have an issue with retaining the word 'activity'. > I find utility in describing the MFs, annotating the gp as having the potential of realizing a MF. > > I see the utility of providing explicit relation between some MF terms and biological processes. This will be a long journey. > I see continued utility of providing 'function' and 'process' sets for the many applications and users currently use MF and BP in that way. > > I don't think we will ever fully provide the relations between MF and BP if only because so much isn't known about how molecular function(ings) contribute to many biological processes. > > Judy > > On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: > > > > On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: > > >> I think the problem here is with trying to characterise MF as either >> *all* functions or *all* processes (i.e functionings) >> > > >> when in fact MF contains a combination of both. >> >> We've been back and forth over this many times and this seems to be >> the crux of it. >> >> Some MF terms fit happily as *parts* of processes (functionings or >> little processes as I prefer to call them ;-)) e.g 'allene-oxide >> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >> tag' does not unfold over time whichever way you look at it. I think >> the strategy of ignoring the existence of this latter class and >> saying everything in MF is a BFO process isn't really satisfactory >> in the long term. >> >> As I see it we have the following options to fix this: >> >> 1. Take away the word 'activity' from the term names and call >> everything in MF a BFO function. Make realised_in (or >> participates_in?) relations between MF and BP. We'd need to think >> carefully about how annotations would be propagated over realised_in. >> > > This is what was proposed in St. Croix and it makes sense to me, but > it does lead us into some unnecessarily confusing areas (i.e. having > yet another relationship to be added). That is why I am now convinced > that #3 the best approach. > > >> 2. Call everything in MF a BFO process and remove any terms that >> don't fit as processes e.g. protein tag. We lose a lot of >> annotations this way. We could possibly move them to a different >> ontology but would mean lots of re-annotation. >> > > Strongly against this one. > > >> 3. Divide MF into BFO processes and BFO functions, either both under >> MF or move the BFO processes into BP and keep MF for functions. The >> disadvantage here is that we wouldn't have the function for e.g. >> allene-oxide cyclase - we'd only have the process. You could I >> suppose make 'allene-oxide cyclase function' and 'allene-oxide >> cyclase activity' but that gets pretty confusing. >> > > I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep > both under MF. Least amount of change for users, while still solving > the problem. > > There is also 3 (c) which Amelia proposed which is to divide and put > the structural terms in an entirely separate new 4th branch, but I > think a simple divide at the top of MF is enough. > > In any case can we please, please please, get rid of the string > "activity". It adds nothing and is quite gratuitous. > > >> 4. Ignore the problem and hope it goes away (it won't). >> > > Nope, we've got to fix this. > > >> I don't really like 2 or 4. >> >> Jane >> >> >> Barry Smith wrote: >> >>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>> >>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>> actions' represented in the molecular function ontology. Ontology- >>>> speak. >>>> >>> Great. >>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>> This is not just ontological persnicketiness. There are already >>> people who are annotating with both 'function' and 'functioning' >>> e.g. using OBI to describe experiments, where it is important to >>> record both the functions e.g. of lab instruments, and the >>> processes in which they participate. The former exist (as >>> potentials) even when the relevant instrument is switched off. >>> There is also parallel work on physiology, distinguishing e.g. the >>> function of the heart (to pump), from realizations of this function >>> in functionings (pumpings), and both of these in turn from non- >>> functional processes such as thumpings, fibrillatings, and so forth. >>> If the GO were to entrench, now, a confusion of function and >>> functioning it will not work work well with these activities in the >>> future. >>> Barry >>> >>> >>>> Yes what is being implemented is what was originally proposed >>>> earlier. Clarifying that the gene product has the potential and >>>> the MF terms describes the specific action is what I was after. >>>> judy >>>> >>>> >>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>> <suzi at fruitfly.org> wrote: >>>> >>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>> >>>> But I think that the solution that David, Karen, and Midori have >>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>> things) the "Function*ings* are part of processes" into a single >>>> artifact, a relationship (i.e. their is no information artifact >>>> for Function*ings* itself, it is implicit) that is for now called >>>> "part_of" (if I remember correctly). The name of this relationship >>>> may not be ideal, but we are trying to be clear about its intended >>>> usage. >>>> >>>> Chris & David first proposed this at the St. Croix meeting. And I >>>> seem to remember that David and I talked about this (and many >>>> other things, such as the nature of the implied relationship >>>> between an instance and a type when anyone makes an annotation) >>>> again at the relation ontology meeting in Denver. >>>> >>>> -S >>>> >>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>> >>>> >>>> >>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>> <Judith.Blake at jax.org> wrote: >>>> >>>> Functions are occurrants. A function describes an action, >>>> biochemical, structural, whatever. >>>> >>>> A function doesn't describe anything. Information artifact >>>> describe things. >>>> Sorry to be picky about words. It's my job. >>>> If you look at the papers on BFO you will see that functions are >>>> continuants and occurrents are disjoint from continuants. So >>>> functions are not occurrents. If the GO is using terminology at >>>> odds with BFO that needs to be fixed - we agreed at the last >>>> foundry meeting that we were all going to use a common upper level >>>> ontology. >>>> >>>> >>>> Functions are part of processes. >>>> >>>> Function*ings* (i.e. realizations of functions) are part of >>>> processes. >>>> >>>> A variety of structures, encoded in a molecule, have the potential >>>> to realize a function. >>>> >>>> Type error: processes don't have potential. (concluded by >>>> substitution "potential to realize a function" + function is_a >>>> process => "potential to realize a process". >>>> >>>> The error is that it is the functions that are the potentials. >>>> Potentials are realized. The realizations are processes. >>>> >>>> We discussed this very recently with Michael and he agreed. >>>> >>>> That is a property of the molecule, not of the function. >>>> >>>> What is the reference of "That"? >>>> >>>> A molecule can have encoded the potential of engaging in several >>>> functions. >>>> >>>> Absolutely. (as long as you rewrite "functions" -> >>>> "functionings" (i.e. processes).) >>>> >>>> A molecule can bear multiple functions is the way we would say this. >>>> >>>> The confusion I think is that of the relationship between the gene >>>> product and a molecular function, and the molecular function and >>>> the biological process. >>>> >>>> And what do you say these are. I (and BFO) are clear: The relation >>>> between a molecular function and a gene product is "inheres". The >>>> relation between molecular function and a biological process is >>>> "realized_by". >>>> >>>> -Alan >>>> >>>> >>>> >>>> Judy >>>> >>>> >>>> >>>> >>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>> <alanruttenberg at gmail.com> wrote: >>>> >>>> The problem is that this still confuses function with process. >>>> Functions are the things that the enzymes (using the term >>>> broadly) have that make the processes happen. If one goes into >>>> more detail, there are a variety of participants in the process >>>> (substrates, products, "cofactors"), but only one bearer of the >>>> function in a particular process. >>>> >>>> Every function has a corresponding process, its "realization". >>>> >>>> Functions are not parts of processes. The realizations of >>>> functions are. >>>> >>>> The functions have a physical basis in the structure of their >>>> bearers. >>>> >>>> The functions exist before any process happens. A molecule may >>>> have a function but never actually realize it. >>>> >>>> These distinctions have important consequences as people such as >>>> myself work on adding more detail to the GO. The processes get >>>> more detail by adding *participants*. Within those processes, when >>>> there is a molecule like an enzyme that functions in a particular >>>> way, it needs to be distinguished from the other participants in >>>> some way. The way that it is distinguished is by saying it bears a >>>> function. >>>> >>>> The heuristic that a molecular function is a "single step" process >>>> is misused, IMO. That it is a "single step" derives from the fact >>>> that in such descriptions there is a single molecule that has a >>>> function. The step is the process that surrounds execution of that >>>> function. >>>> >>>> As I said, getting this distinction clear is essential for future >>>> detailing of GO processes such as I and others do. Can we start to >>>> get this right here and now? >>>> >>>> -Alan >>>> >>>> >>>> >>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>> <midori at ebi.ac.uk> wrote: >>>> Hi, >>>> >>>> We have two SF items in which merging function and process terms >>>> looks like a viable solution to a problem. One is an overly- >>>> specific, single-step "process" that is essentially equivalent to >>>> a "function" (in the GO sense); the other is a "function" term >>>> that is defined such that it doesn't describe a single activity, >>>> and is essentially the same as a process term. >>>> >>>> SF 2864212 - merge MF term >>>> RNA splicing factor activity, transesterification mechanism GO: >>>> 0031202 >>>> >>>> into BP term >>>> RNA splicing, via transesterification reactions GO:0000375 >>>> >>>> For this one, we would merge the MF term into the BP term, and >>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>> think that's all we would change. >>>> >>>> >>> >>>>> https://sourceforge.net/tracker/? >>>>> >>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>> >>>> >>>> SF 2864271 - merge BP term >>>> >>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>> >>>> into MF term >>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>> >>>> >>> >>>>> https://sourceforge.net/tracker/? >>>>> >>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>> >>>> In this case, the proposal is to merge the process term into the >>>> function term, ensuring that the merged term has is_a ancestry >>>> only in the MF ontology,and has a part_of link to a process term. >>>> Specifically, the merged term would retain the name and ID from GO: >>>> 0004479, and would be part_of translational initiation GO:0006413. >>>> >>>> Does anyone have any objections or other comments on these? >>>> >>>> thanks, >>>> m >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>>> >>>> >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>>> >>>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>> >> -- >> Dr Jane Lomax >> GO Editorial Office >> EMBL-EBI >> Wellcome Trust Genome Campus >> Hinxton >> Cambridgeshire, UK >> CB10 1SD >> >> p: +44 1223 492516 >> f: +44 1223 494468 >> >> >> > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From shimoyama at mcw.edu Wed Oct 7 06:43:37 2009 From: shimoyama at mcw.edu (Shimoyama, Mary) Date: Wed, 7 Oct 2009 08:43:37 -0500 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACC8F79.2060808@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> Message-ID: <1448A38A42714048B9C53E473E13CCF0028DD0D0@davis.hmgc.mcw.edu> As always David, you have provided a very succinct and intelligent assessment. Mary Shimoyama Program Manager Rat Genome Database Human and Molecular Genetics Center Medical College of Wisconsin shimoyama at mcw.edu Tel: 414-456-7505 Fax: 414-456-6516 http://rgd.mcw.edu -----Original Message----- From: annotation-bounces at genome.stanford.edu [mailto:annotation-bounces at genome.stanford.edu] On Behalf Of David Hill Sent: Wednesday, October 07, 2009 7:54 AM To: Judith Blake Cc: Suzanna Lewis; Barry Smith; Ontology at genome.stanford.edu; Editors; GO Annotation list; Michael Ashburner Subject: Re: [Annotation] [Ontology-editors] merging MF and BP terms Hi Everyone, I think the issue with the problematic function terms is that we described them as roles because we didn't take enough time to carefully think about what the gene products were actually doing. When I look at them, they are almost always cases where we wanted to assign a function because we knew the product functioned in a process. I think we can just define them as functions involved in processes, sometimes the functions are generic and sometimes they are not. Transcription factor activity is and example of one that is not. It is_a 'DNA binding activity' involved in 'regulation of transcription'. Here is what a biologist wants to know: Gene product X has the ability to carry out some kind of action (function), either passively (like actin) or actively (like histidine kinase). In the proper biological context, the gene product will carry out that action and that will be part of a biological process. In some cases, there is a requirement for more than one gene product for this elemental action to be carried out. The biological process describes the overall objective of the action. I cannot think of any case where a single gene product acts and has a biological effect, that is succeeds in achieving a biological objective. If that were the case, I think biology wouldn't work, because there is always some background activity as long as a molecule and substrates are present. This is the distinction between a function and a process in the misunderstood 'single step processes'. Even phosphorylation as a biological objective (that has some effect on a biological system) is the result of the action of the activity of many kinase molecules, or the repeated activity of individual kinase molecules. So there are three ways we can represent this: 1) gene product MFing biological process In this case the gene product is a continuant. The relationship describes the relationship between the continuant and the MF which is an occurent. The biological process is an occurent. 2) gene product MF MFing biological process In this case the gene product is a continuant. The MF is a continuant that is an enduring quality of the gene product. The MFing is an occurent that is the realization of the MF. The biological process is an occurent. 3) gene product MF BP In this case the gene product is a continuant. The MF is a continuant that is an enduring quality of the gene product. The BP is the occurent that results from the realization of the MF. I think that the most rigorous way to represent this would be #2. But, I think the most practical way would be #1. When I talk to biologists, both annotators and non-annotators, their eyes glaze over as soon as I mention functions and functionings. The bottom line is they don't care about it, they don't care to think about it, and they are confused by it. They are confused because in the absence of the gene product <>MF relationship, all they care about is a function being executed as part of a biological process. They could care less about the neurotransmitters that never bind a receptor and have a potential to execute a function, but never execute it. However, if we get rid of the MF<>BP side, they care about what the potential of the gene products are. So #1 works for this view. The MF being an occurent is the intuitive way that we think about this. MFs being parts of processes fits with the view I expressed above. #3 is not practical because it would require putting every MF into the BP ontology as an occurent. This would essentially recapitulate the entire MF ontology in BP. It is much easier for us to use #1. The key is in the relationship which combines the ideas of 'has potential' and 'carries out'. My 2 cents. David Judith Blake wrote: > I agree with Suzi, I support the continuation of MF, always have. > I don't have an issue with retaining the word 'activity'. > I find utility in describing the MFs, annotating the gp as having the potential of realizing a MF. > > I see the utility of providing explicit relation between some MF terms and biological processes. This will be a long journey. > I see continued utility of providing 'function' and 'process' sets for the many applications and users currently use MF and BP in that way. > > I don't think we will ever fully provide the relations between MF and BP if only because so much isn't known about how molecular function(ings) contribute to many biological processes. > > Judy > > On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: > > > > On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: > > >> I think the problem here is with trying to characterise MF as either >> *all* functions or *all* processes (i.e functionings) >> > > >> when in fact MF contains a combination of both. >> >> We've been back and forth over this many times and this seems to be >> the crux of it. >> >> Some MF terms fit happily as *parts* of processes (functionings or >> little processes as I prefer to call them ;-)) e.g 'allene-oxide >> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >> tag' does not unfold over time whichever way you look at it. I think >> the strategy of ignoring the existence of this latter class and >> saying everything in MF is a BFO process isn't really satisfactory >> in the long term. >> >> As I see it we have the following options to fix this: >> >> 1. Take away the word 'activity' from the term names and call >> everything in MF a BFO function. Make realised_in (or >> participates_in?) relations between MF and BP. We'd need to think >> carefully about how annotations would be propagated over realised_in. >> > > This is what was proposed in St. Croix and it makes sense to me, but > it does lead us into some unnecessarily confusing areas (i.e. having > yet another relationship to be added). That is why I am now convinced > that #3 the best approach. > > >> 2. Call everything in MF a BFO process and remove any terms that >> don't fit as processes e.g. protein tag. We lose a lot of >> annotations this way. We could possibly move them to a different >> ontology but would mean lots of re-annotation. >> > > Strongly against this one. > > >> 3. Divide MF into BFO processes and BFO functions, either both under >> MF or move the BFO processes into BP and keep MF for functions. The >> disadvantage here is that we wouldn't have the function for e.g. >> allene-oxide cyclase - we'd only have the process. You could I >> suppose make 'allene-oxide cyclase function' and 'allene-oxide >> cyclase activity' but that gets pretty confusing. >> > > I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep > both under MF. Least amount of change for users, while still solving > the problem. > > There is also 3 (c) which Amelia proposed which is to divide and put > the structural terms in an entirely separate new 4th branch, but I > think a simple divide at the top of MF is enough. > > In any case can we please, please please, get rid of the string > "activity". It adds nothing and is quite gratuitous. > > >> 4. Ignore the problem and hope it goes away (it won't). >> > > Nope, we've got to fix this. > > >> I don't really like 2 or 4. >> >> Jane >> >> >> Barry Smith wrote: >> >>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>> >>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>> actions' represented in the molecular function ontology. Ontology- >>>> speak. >>>> >>> Great. >>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>> This is not just ontological persnicketiness. There are already >>> people who are annotating with both 'function' and 'functioning' >>> e.g. using OBI to describe experiments, where it is important to >>> record both the functions e.g. of lab instruments, and the >>> processes in which they participate. The former exist (as >>> potentials) even when the relevant instrument is switched off. >>> There is also parallel work on physiology, distinguishing e.g. the >>> function of the heart (to pump), from realizations of this function >>> in functionings (pumpings), and both of these in turn from non- >>> functional processes such as thumpings, fibrillatings, and so forth. >>> If the GO were to entrench, now, a confusion of function and >>> functioning it will not work work well with these activities in the >>> future. >>> Barry >>> >>> >>>> Yes what is being implemented is what was originally proposed >>>> earlier. Clarifying that the gene product has the potential and >>>> the MF terms describes the specific action is what I was after. >>>> judy >>>> >>>> >>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>> <suzi at fruitfly.org> wrote: >>>> >>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>> >>>> But I think that the solution that David, Karen, and Midori have >>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>> things) the "Function*ings* are part of processes" into a single >>>> artifact, a relationship (i.e. their is no information artifact >>>> for Function*ings* itself, it is implicit) that is for now called >>>> "part_of" (if I remember correctly). The name of this relationship >>>> may not be ideal, but we are trying to be clear about its intended >>>> usage. >>>> >>>> Chris & David first proposed this at the St. Croix meeting. And I >>>> seem to remember that David and I talked about this (and many >>>> other things, such as the nature of the implied relationship >>>> between an instance and a type when anyone makes an annotation) >>>> again at the relation ontology meeting in Denver. >>>> >>>> -S >>>> >>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>> >>>> >>>> >>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>> <Judith.Blake at jax.org> wrote: >>>> >>>> Functions are occurrants. A function describes an action, >>>> biochemical, structural, whatever. >>>> >>>> A function doesn't describe anything. Information artifact >>>> describe things. >>>> Sorry to be picky about words. It's my job. >>>> If you look at the papers on BFO you will see that functions are >>>> continuants and occurrents are disjoint from continuants. So >>>> functions are not occurrents. If the GO is using terminology at >>>> odds with BFO that needs to be fixed - we agreed at the last >>>> foundry meeting that we were all going to use a common upper level >>>> ontology. >>>> >>>> >>>> Functions are part of processes. >>>> >>>> Function*ings* (i.e. realizations of functions) are part of >>>> processes. >>>> >>>> A variety of structures, encoded in a molecule, have the potential >>>> to realize a function. >>>> >>>> Type error: processes don't have potential. (concluded by >>>> substitution "potential to realize a function" + function is_a >>>> process => "potential to realize a process". >>>> >>>> The error is that it is the functions that are the potentials. >>>> Potentials are realized. The realizations are processes. >>>> >>>> We discussed this very recently with Michael and he agreed. >>>> >>>> That is a property of the molecule, not of the function. >>>> >>>> What is the reference of "That"? >>>> >>>> A molecule can have encoded the potential of engaging in several >>>> functions. >>>> >>>> Absolutely. (as long as you rewrite "functions" -> >>>> "functionings" (i.e. processes).) >>>> >>>> A molecule can bear multiple functions is the way we would say this. >>>> >>>> The confusion I think is that of the relationship between the gene >>>> product and a molecular function, and the molecular function and >>>> the biological process. >>>> >>>> And what do you say these are. I (and BFO) are clear: The relation >>>> between a molecular function and a gene product is "inheres". The >>>> relation between molecular function and a biological process is >>>> "realized_by". >>>> >>>> -Alan >>>> >>>> >>>> >>>> Judy >>>> >>>> >>>> >>>> >>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>> <alanruttenberg at gmail.com> wrote: >>>> >>>> The problem is that this still confuses function with process. >>>> Functions are the things that the enzymes (using the term >>>> broadly) have that make the processes happen. If one goes into >>>> more detail, there are a variety of participants in the process >>>> (substrates, products, "cofactors"), but only one bearer of the >>>> function in a particular process. >>>> >>>> Every function has a corresponding process, its "realization". >>>> >>>> Functions are not parts of processes. The realizations of >>>> functions are. >>>> >>>> The functions have a physical basis in the structure of their >>>> bearers. >>>> >>>> The functions exist before any process happens. A molecule may >>>> have a function but never actually realize it. >>>> >>>> These distinctions have important consequences as people such as >>>> myself work on adding more detail to the GO. The processes get >>>> more detail by adding *participants*. Within those processes, when >>>> there is a molecule like an enzyme that functions in a particular >>>> way, it needs to be distinguished from the other participants in >>>> some way. The way that it is distinguished is by saying it bears a >>>> function. >>>> >>>> The heuristic that a molecular function is a "single step" process >>>> is misused, IMO. That it is a "single step" derives from the fact >>>> that in such descriptions there is a single molecule that has a >>>> function. The step is the process that surrounds execution of that >>>> function. >>>> >>>> As I said, getting this distinction clear is essential for future >>>> detailing of GO processes such as I and others do. Can we start to >>>> get this right here and now? >>>> >>>> -Alan >>>> >>>> >>>> >>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>> <midori at ebi.ac.uk> wrote: >>>> Hi, >>>> >>>> We have two SF items in which merging function and process terms >>>> looks like a viable solution to a problem. One is an overly- >>>> specific, single-step "process" that is essentially equivalent to >>>> a "function" (in the GO sense); the other is a "function" term >>>> that is defined such that it doesn't describe a single activity, >>>> and is essentially the same as a process term. >>>> >>>> SF 2864212 - merge MF term >>>> RNA splicing factor activity, transesterification mechanism GO: >>>> 0031202 >>>> >>>> into BP term >>>> RNA splicing, via transesterification reactions GO:0000375 >>>> >>>> For this one, we would merge the MF term into the BP term, and >>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>> think that's all we would change. >>>> >>>> >>> >>>>> https://sourceforge.net/tracker/? >>>>> >>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>> >>>> >>>> SF 2864271 - merge BP term >>>> >>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>> >>>> into MF term >>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>> >>>> >>> >>>>> https://sourceforge.net/tracker/? >>>>> >>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>> >>>> In this case, the proposal is to merge the process term into the >>>> function term, ensuring that the merged term has is_a ancestry >>>> only in the MF ontology,and has a part_of link to a process term. >>>> Specifically, the merged term would retain the name and ID from GO: >>>> 0004479, and would be part_of translational initiation GO:0006413. >>>> >>>> Does anyone have any objections or other comments on these? >>>> >>>> thanks, >>>> m >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>>> >>>> >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>>> >>>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>> >> -- >> Dr Jane Lomax >> GO Editorial Office >> EMBL-EBI >> Wellcome Trust Genome Campus >> Hinxton >> Cambridgeshire, UK >> CB10 1SD >> >> p: +44 1223 492516 >> f: +44 1223 494468 >> >> >> > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 _______________________________________________ Annotation mailing list Annotation at geneontology.org http://fafner.stanford.edu/mailman/listinfo/annotation From val at sanger.ac.uk Wed Oct 7 07:39:25 2009 From: val at sanger.ac.uk (Valerie Wood) Date: Wed, 07 Oct 2009 15:39:25 +0100 Subject: [Ontology-editors] small molecule metabolism In-Reply-To: <49E46DD1.8050600@sanger.ac.uk> References: <583876A4-B091-494E-8CA3-C94F341F9463@berkeleybop.org> <49DDCF81.3050605@sanger.ac.uk> <8e22ab960904090917w776600a5g51e2c940c9efe889@mail.gmail.com> <49DE23D3.5060208@sanger.ac.uk> <49DF4F2F.5000205@informatics.jax.org> <49DF5100.2030402@informatics.jax.org> <49E46DD1.8050600@sanger.ac.uk> Message-ID: <4ACCA81D.6070303@sanger.ac.uk> I am resurrecting this SF item 21006200 as 1. I have come across some users studying the "metabolome" who potentially have a problem with the absence of a term which describes "small molecule metabolism". They were expecting an enrichment for primary metabolism, but this includes over half of S. cerevisiae genes because it includes DNA metabolism, protein metabolism etc. It is not possible to see enrichment for "metabolic pathways" in general . This means that if you only see the enrichment from the culmulative annotations to all of the "small molecule biochmemical pathways" you would miss it. Jane and also think it would be a handy term in the slim, otherwise we would need to include a very broad term (primary metabolism < 50% of S.c ) val Valerie Wood wrote: > > Do we need to include these di-polymers in the small molecule > pathways? if not, it could be defined as non-polymeric molecules as > Harold suggested? > > There do not appear to be many metabolic pathways which involve the > disaccharides which do not involve their catabolism into their > component mono-saccharides, or their attachment to a macromolecule? > in which case I don't think we would lose anything be defining > > "small molecule metabolism, or metabolite metabolism" as > > The chemical reactions and pathways, including anabolism and > catabolism, by which living organisms transform non-polymeric organic > molecules. > > and groupling all of small molecule metabolism under here ( excluding > disaccharide metabolic process) > > a di-polymer is therefore neither a small molecule of a macromolecule > by current def (the macromolecule def is actually vaguer because it > refers to "high" relative molecular mass....) > > Val > > > Harold Drabkin wrote: > >> arrgh! Alex reminded me of sucrose and lactose. >> maybe we want to cut it off above 2? >> >> There goes my epiphany! >> >> h >> >> >> Harold Drabkin wrote: >> >>> a thought.. >>> >>> How about saying that these are the non-polymeric molecules found >>> within a cell? (assuming we do not want to count di-peptides, >>> di-nucleotides, or di-saccharides )? >>> >>> hd >>> >>> Chris Mungall wrote: >>> >>>> >>>> On Apr 9, 2009, at 9:35 AM, Valerie Wood wrote: >>>> >>>>> Tanya Berardini wrote: >>>>> >>>>>> >>>>>> >>>>>> On Thu, Apr 9, 2009 at 8:48 AM, Chris Mungall >>>>>> > wrote: >>>>>> >>>>>> >>>>>> On Apr 9, 2009, at 3:35 AM, Valerie Wood wrote: >>>>>> >>>>>> It seems like there is a gap in the terminology of biology to >>>>>> decribe "everything that is not a macromolecule molecule". >>>>>> Maybe we should make one up.... >>>>>> Perhaps "small molecule metabolism" would be acceptable if it >>>>>> is defined as "everything that is not a macromolucule" but >>>>>> that is not an acceptable way of defining something is it? >>>>>> >>>>>> >>>>>> Do we really need a term for it? Why not just ask for non-X >>>>>> metabolism any time you're interested in metabolism of Ys >>>>>> where Ys >>>>>> are not Xs >>>>>> >>>>>> Granted tools can't do this yet but it's not hard given the >>>>>> correct structures in the ontology, and we should perhaps be >>>>>> working towards a situation where tools do support this >>>>>> >>>>>> >>>>>> I am partial to this approach. Defining 'small molecule >>>>>> metabolism' as everything that is not 'macromolecule metabolism' >>>>>> violates the ontology design principle of positivity. Why not >>>>>> just combine the annotations from the terms that do cover what is >>>>>> desired and then analyze those results? >>>>>> >>>>>> Tanya >>>>>> >>>>> >>>>> Its quite difficult to do this during enrichment analysis. >>>>> I have seen a number of times that terms which would be >>>>> classically termed "biochemical pathways" are enriched, bacause I >>>>> see the annotations in my data individually. >>>>> The enrichment tools don't show this because the number of >>>>> annotations to the individual terms are not large enough. tThe >>>>> parent term "cellular metabolic process" is not enriched because >>>>> the effect is masked by all of the other 3000 annotations to this >>>>> term generated mainly by the variouse types of macromolecule >>>>> metabolic process i.e.DNA metabolic process, protein metabolic >>>>> process etc. >>>>> >>>>> When you are analysisng whole genome datasets it isn't really >>>>> practical to add and subtract processes and repeat enrichment (it >>>>> gets way too complicated to process and report the results, as you >>>>> would have to fiddle with the P-values for everything you did >>>>> manually and then reintegrate it into your whole genome analysis) >>>>> >>>>> This isn't really a problem for me now because I worked around it, >>>>> but I could only do this because I know what the problem was. >>>>> >>>>> I thought that other users may appreciate some sort of grouping >>>>> term here for similar analyses- it just seems that there should be >>>>> a term to group these processes in the same way that >>>>> macromolecular metabolic processes are grouped. The fact is that >>>>> if you have a bunch of genes enriched for low numbers of various >>>>> small metabolism/canonical biochemical pathway terms, this >>>>> enrichment would most likely be overlooked. >>>> >>>> >>>> I agree with Tanya but understand the practical need for enrichment >>>> analysis. >>>> >>>> I am envisioning a partial solution along the following lines: >>>> >>>> Just as we have goslims, we can have gofats. a gofat would live >>>> outside the ontology and contain statements like >>>> >>>> GOFAT:1 small molecule metabolism = metabolism and not >>>> has_participant chebi:macromolecule >>>> >>>> The reasoner would compute is_a parentage to the fat terms and >>>> create a derived obo file. If the tool accepts OBOFs+GAFs then just >>>> give the tool this obo file instead of the regular one. >>>> >>>> (note this only works for tools that allow you to input an obo >>>> file. Some web-based tools may not give you the flexibility to >>>> substitute anything other than the regular GO) >>>> >>>> This should work in your particular case, no need to keep >>>> re-analyzing once you've defined your fat. We could even make the >>>> fat-derived obo files available on the website, as we do for slims >>>> >>>> This solution isn't perfect as it requires the analyzer to know a >>>> priori which may be useful grouping categories. Really the tool >>>> should be able to do this. For example, for a particular dataset, >>>> "metabolism with a molecule with an X side chain, missing a Y" may >>>> be enriched. There are strategies for dealing with this - rule >>>> mining, or pre-computing every possible cross-product. This will >>>> require a little more know-how from tools developers. >>>> >>>> >>>>> >>>>> val >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> The Wellcome Trust Sanger Institute is operated by Genome Research >>>>> Limited, a charity registered in England with number 1021457 and a >>>>> company registered in England with number 2742969, whose >>>>> registered office is 215 Euston Road, London, NW1 2BE. >>>> >>>> >>>> _______________________________________________ >>>> Ontology-editors mailing list >>>> Ontology-editors at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>> >>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >> >> >> > > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From val at sanger.ac.uk Wed Oct 7 07:43:06 2009 From: val at sanger.ac.uk (val at sanger.ac.uk) Date: Wed, 7 Oct 2009 15:43:06 +0100 (BST) Subject: [Ontology-editors] small molecule metabolism In-Reply-To: <4ACCA81D.6070303@sanger.ac.uk> References: <583876A4-B091-494E-8CA3-C94F341F9463@berkeleybop.org> <49DDCF81.3050605@sanger.ac.uk> <8e22ab960904090917w776600a5g51e2c940c9efe889@mail.gmail.com> <49DE23D3.5060208@sanger.ac.uk> <49DF4F2F.5000205@informatics.jax.org> <49DF5100.2030402@informatics.jax.org> <49E46DD1.8050600@sanger.ac.uk> <4ACCA81D.6070303@sanger.ac.uk> Message-ID: sorry 2106200 val > I am resurrecting this SF item 21006200 > > as > > 1. I have come across some users studying the "metabolome" > who potentially have a problem with the absence of a term which > describes "small molecule metabolism". They were expecting an enrichment > for primary metabolism, but this includes over half of S. cerevisiae > genes because it includes DNA metabolism, protein metabolism etc. It is > not possible to see enrichment for "metabolic pathways" in general . > This means that if you only see the enrichment from the culmulative > annotations to all of the "small molecule biochmemical pathways" you > would miss it. > > Jane and also think it would be a handy term in the slim, otherwise we > would need to include a very broad term > (primary metabolism < 50% of S.c ) > > val > > > > Valerie Wood wrote: >> >> Do we need to include these di-polymers in the small molecule >> pathways? if not, it could be defined as non-polymeric molecules as >> Harold suggested? >> >> There do not appear to be many metabolic pathways which involve the >> disaccharides which do not involve their catabolism into their >> component mono-saccharides, or their attachment to a macromolecule? >> in which case I don't think we would lose anything be defining >> >> "small molecule metabolism, or metabolite metabolism" as >> >> The chemical reactions and pathways, including anabolism and >> catabolism, by which living organisms transform non-polymeric organic >> molecules. >> >> and groupling all of small molecule metabolism under here ( excluding >> disaccharide metabolic process) >> >> a di-polymer is therefore neither a small molecule of a macromolecule >> by current def (the macromolecule def is actually vaguer because it >> refers to "high" relative molecular mass....) >> >> Val >> >> >> Harold Drabkin wrote: >> >>> arrgh! Alex reminded me of sucrose and lactose. >>> maybe we want to cut it off above 2? >>> >>> There goes my epiphany! >>> >>> h >>> >>> >>> Harold Drabkin wrote: >>> >>>> a thought.. >>>> >>>> How about saying that these are the non-polymeric molecules found >>>> within a cell? (assuming we do not want to count di-peptides, >>>> di-nucleotides, or di-saccharides )? >>>> >>>> hd >>>> >>>> Chris Mungall wrote: >>>> >>>>> >>>>> On Apr 9, 2009, at 9:35 AM, Valerie Wood wrote: >>>>> >>>>>> Tanya Berardini wrote: >>>>>> >>>>>>> >>>>>>> >>>>>>> On Thu, Apr 9, 2009 at 8:48 AM, Chris Mungall >>>>>>> > wrote: >>>>>>> >>>>>>> >>>>>>> On Apr 9, 2009, at 3:35 AM, Valerie Wood wrote: >>>>>>> >>>>>>> It seems like there is a gap in the terminology of biology >>>>>>> to >>>>>>> decribe "everything that is not a macromolecule molecule". >>>>>>> Maybe we should make one up.... >>>>>>> Perhaps "small molecule metabolism" would be acceptable if >>>>>>> it >>>>>>> is defined as "everything that is not a macromolucule" but >>>>>>> that is not an acceptable way of defining something is it? >>>>>>> >>>>>>> >>>>>>> Do we really need a term for it? Why not just ask for non-X >>>>>>> metabolism any time you're interested in metabolism of Ys >>>>>>> where Ys >>>>>>> are not Xs >>>>>>> >>>>>>> Granted tools can't do this yet but it's not hard given the >>>>>>> correct structures in the ontology, and we should perhaps be >>>>>>> working towards a situation where tools do support this >>>>>>> >>>>>>> >>>>>>> I am partial to this approach. Defining 'small molecule >>>>>>> metabolism' as everything that is not 'macromolecule metabolism' >>>>>>> violates the ontology design principle of positivity. Why not >>>>>>> just combine the annotations from the terms that do cover what is >>>>>>> desired and then analyze those results? >>>>>>> >>>>>>> Tanya >>>>>>> >>>>>> >>>>>> Its quite difficult to do this during enrichment analysis. >>>>>> I have seen a number of times that terms which would be >>>>>> classically termed "biochemical pathways" are enriched, bacause I >>>>>> see the annotations in my data individually. >>>>>> The enrichment tools don't show this because the number of >>>>>> annotations to the individual terms are not large enough. tThe >>>>>> parent term "cellular metabolic process" is not enriched because >>>>>> the effect is masked by all of the other 3000 annotations to this >>>>>> term generated mainly by the variouse types of macromolecule >>>>>> metabolic process i.e.DNA metabolic process, protein metabolic >>>>>> process etc. >>>>>> >>>>>> When you are analysisng whole genome datasets it isn't really >>>>>> practical to add and subtract processes and repeat enrichment (it >>>>>> gets way too complicated to process and report the results, as you >>>>>> would have to fiddle with the P-values for everything you did >>>>>> manually and then reintegrate it into your whole genome analysis) >>>>>> >>>>>> This isn't really a problem for me now because I worked around it, >>>>>> but I could only do this because I know what the problem was. >>>>>> >>>>>> I thought that other users may appreciate some sort of grouping >>>>>> term here for similar analyses- it just seems that there should be >>>>>> a term to group these processes in the same way that >>>>>> macromolecular metabolic processes are grouped. The fact is that >>>>>> if you have a bunch of genes enriched for low numbers of various >>>>>> small metabolism/canonical biochemical pathway terms, this >>>>>> enrichment would most likely be overlooked. >>>>> >>>>> >>>>> I agree with Tanya but understand the practical need for enrichment >>>>> analysis. >>>>> >>>>> I am envisioning a partial solution along the following lines: >>>>> >>>>> Just as we have goslims, we can have gofats. a gofat would live >>>>> outside the ontology and contain statements like >>>>> >>>>> GOFAT:1 small molecule metabolism = metabolism and not >>>>> has_participant chebi:macromolecule >>>>> >>>>> The reasoner would compute is_a parentage to the fat terms and >>>>> create a derived obo file. If the tool accepts OBOFs+GAFs then just >>>>> give the tool this obo file instead of the regular one. >>>>> >>>>> (note this only works for tools that allow you to input an obo >>>>> file. Some web-based tools may not give you the flexibility to >>>>> substitute anything other than the regular GO) >>>>> >>>>> This should work in your particular case, no need to keep >>>>> re-analyzing once you've defined your fat. We could even make the >>>>> fat-derived obo files available on the website, as we do for slims >>>>> >>>>> This solution isn't perfect as it requires the analyzer to know a >>>>> priori which may be useful grouping categories. Really the tool >>>>> should be able to do this. For example, for a particular dataset, >>>>> "metabolism with a molecule with an X side chain, missing a Y" may >>>>> be enriched. There are strategies for dealing with this - rule >>>>> mining, or pre-computing every possible cross-product. This will >>>>> require a little more know-how from tools developers. >>>>> >>>>> >>>>>> >>>>>> val >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> The Wellcome Trust Sanger Institute is operated by Genome Research >>>>>> Limited, a charity registered in England with number 1021457 and a >>>>>> company registered in England with number 2742969, whose >>>>>> registered office is 215 Euston Road, London, NW1 2BE. >>>>> >>>>> >>>>> _______________________________________________ >>>>> Ontology-editors mailing list >>>>> Ontology-editors at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>> >>>> >>>> _______________________________________________ >>>> Ontology-editors mailing list >>>> Ontology-editors at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>> >>> >>> >> >> > > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From jane at ebi.ac.uk Wed Oct 7 13:12:34 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Wed, 7 Oct 2009 21:12:34 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACC8F79.2060808@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> Message-ID: Hi David - got a bit confused for a minute there because I think your #1 is equivalent to my #3. So to be clear, I'm talking about this one: gene product MFing biological process The downside of this option is what we do with the terms in MF that aren't occurents - protein tag and their ilk. Are you up for the two nodes option? As an experiment I tried splitting MF into MFings and true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. I can send the file if anyone's interested. Jane On Wed, 7 Oct 2009, David Hill wrote: > Hi Everyone, > > I think the issue with the problematic function terms is that we described > them as roles because we didn't take enough time to carefully think about > what the gene products were actually doing. When I look at them, they are > almost always cases where we wanted to assign a function because we knew the > product functioned in a process. I think we can just define them as functions > involved in processes, sometimes the functions are generic and sometimes they > are not. Transcription factor activity is and example of one that is not. It > is_a 'DNA binding activity' involved in 'regulation of transcription'. > > Here is what a biologist wants to know: > > Gene product X has the ability to carry out some kind of action (function), > either passively (like actin) or actively (like histidine kinase). In the > proper biological context, the gene product will carry out that action and > that will be part of a biological process. > > > In some cases, there is a requirement for more than one gene product for this > elemental action to be carried out. The biological process describes the > overall objective of the action. I cannot think of any case where a single > gene product acts and has a biological effect, that is succeeds in achieving > a biological objective. If that were the case, I think biology wouldn't work, > because there is always some background activity as long as a molecule and > substrates are present. This is the distinction between a function and a > process in the misunderstood 'single step processes'. Even phosphorylation as > a biological objective (that has some effect on a biological system) is the > result of the action of the activity of many kinase molecules, or the > repeated activity of individual kinase molecules. > > So there are three ways we can represent this: > > 1) gene product MFing biological > process > > In this case the gene product is a continuant. The out> relationship describes the relationship between the continuant and the > MF which is an occurent. The biological process is an occurent. > > 2) gene product MF MFing biological > process > > In this case the gene product is a continuant. The MF is a continuant that is > an enduring quality of the gene product. The MFing is an occurent that is the > realization of the MF. The biological process is an occurent. > > 3) gene product MF BP > > In this case the gene product is a continuant. The MF is a continuant that is > an enduring quality of the gene product. The BP is the occurent that results > from the realization of the MF. > > I think that the most rigorous way to represent this would be #2. But, I > think the most practical way would be #1. When I talk to biologists, both > annotators and non-annotators, their eyes glaze over as soon as I mention > functions and functionings. The bottom line is they don't care about it, they > don't care to think about it, and they are confused by it. They are confused > because in the absence of the gene product <>MF relationship, all they care > about is a function being executed as part of a biological process. They > could care less about the neurotransmitters that never bind a receptor and > have a potential to execute a function, but never execute it. > > However, if we get rid of the MF<>BP side, they care about what the potential > of the gene products are. So #1 works for this view. The MF being an occurent > is the intuitive way that we think about this. MFs being parts of processes > fits with the view I expressed above. > > #3 is not practical because it would require putting every MF into the BP > ontology as an occurent. This would essentially recapitulate the entire MF > ontology in BP. It is much easier for us to use #1. The key is in the > relationship which combines the ideas of 'has > potential' and 'carries out'. > > My 2 cents. > > David > > Judith Blake wrote: >> I agree with Suzi, I support the continuation of MF, always have. >> I don't have an issue with retaining the word 'activity'. >> I find utility in describing the MFs, annotating the gp as having the >> potential of realizing a MF. >> >> I see the utility of providing explicit relation between some MF terms and >> biological processes. This will be a long journey. >> I see continued utility of providing 'function' and 'process' sets for the >> many applications and users currently use MF and BP in that way. >> >> I don't think we will ever fully provide the relations between MF and BP if >> only because so much isn't known about how molecular function(ings) >> contribute to many biological processes. >> >> Judy >> >> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >> >> >> >> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >> >> >>> I think the problem here is with trying to characterise MF as either >>> *all* functions or *all* processes (i.e functionings) >>> >> >> >>> when in fact MF contains a combination of both. >>> >>> We've been back and forth over this many times and this seems to be >>> the crux of it. >>> >>> Some MF terms fit happily as *parts* of processes (functionings or >>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>> tag' does not unfold over time whichever way you look at it. I think >>> the strategy of ignoring the existence of this latter class and >>> saying everything in MF is a BFO process isn't really satisfactory >>> in the long term. >>> >>> As I see it we have the following options to fix this: >>> >>> 1. Take away the word 'activity' from the term names and call >>> everything in MF a BFO function. Make realised_in (or >>> participates_in?) relations between MF and BP. We'd need to think >>> carefully about how annotations would be propagated over realised_in. >>> >> >> This is what was proposed in St. Croix and it makes sense to me, but >> it does lead us into some unnecessarily confusing areas (i.e. having >> yet another relationship to be added). That is why I am now convinced >> that #3 the best approach. >> >> >>> 2. Call everything in MF a BFO process and remove any terms that >>> don't fit as processes e.g. protein tag. We lose a lot of >>> annotations this way. We could possibly move them to a different >>> ontology but would mean lots of re-annotation. >>> >> >> Strongly against this one. >> >> >>> 3. Divide MF into BFO processes and BFO functions, either both under >>> MF or move the BFO processes into BP and keep MF for functions. The >>> disadvantage here is that we wouldn't have the function for e.g. >>> allene-oxide cyclase - we'd only have the process. You could I >>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>> cyclase activity' but that gets pretty confusing. >>> >> >> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep >> both under MF. Least amount of change for users, while still solving >> the problem. >> >> There is also 3 (c) which Amelia proposed which is to divide and put >> the structural terms in an entirely separate new 4th branch, but I >> think a simple divide at the top of MF is enough. >> >> In any case can we please, please please, get rid of the string >> "activity". It adds nothing and is quite gratuitous. >> >> >>> 4. Ignore the problem and hope it goes away (it won't). >>> >> >> Nope, we've got to fix this. >> >> >>> I don't really like 2 or 4. >>> >>> Jane >>> >>> >>> Barry Smith wrote: >>> >>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>> >>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>> actions' represented in the molecular function ontology. Ontology- >>>>> speak. >>>>> >>>> Great. >>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>> This is not just ontological persnicketiness. There are already >>>> people who are annotating with both 'function' and 'functioning' >>>> e.g. using OBI to describe experiments, where it is important to >>>> record both the functions e.g. of lab instruments, and the >>>> processes in which they participate. The former exist (as >>>> potentials) even when the relevant instrument is switched off. >>>> There is also parallel work on physiology, distinguishing e.g. the >>>> function of the heart (to pump), from realizations of this function >>>> in functionings (pumpings), and both of these in turn from non- >>>> functional processes such as thumpings, fibrillatings, and so forth. >>>> If the GO were to entrench, now, a confusion of function and >>>> functioning it will not work work well with these activities in the >>>> future. >>>> Barry >>>> >>>> >>>>> Yes what is being implemented is what was originally proposed >>>>> earlier. Clarifying that the gene product has the potential and >>>>> the MF terms describes the specific action is what I was after. >>>>> judy >>>>> >>>>> >>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>> <suzi at fruitfly.org> wrote: >>>>> >>>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>>> >>>>> But I think that the solution that David, Karen, and Midori have >>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>> things) the "Function*ings* are part of processes" into a single >>>>> artifact, a relationship (i.e. their is no information artifact >>>>> for Function*ings* itself, it is implicit) that is for now called >>>>> "part_of" (if I remember correctly). The name of this relationship >>>>> may not be ideal, but we are trying to be clear about its intended >>>>> usage. >>>>> >>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>> seem to remember that David and I talked about this (and many >>>>> other things, such as the nature of the implied relationship >>>>> between an instance and a type when anyone makes an annotation) >>>>> again at the relation ontology meeting in Denver. >>>>> >>>>> -S >>>>> >>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>> >>>>> >>>>> >>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>> <Judith.Blake at jax.org> wrote: >>>>> >>>>> Functions are occurrants. A function describes an action, >>>>> biochemical, structural, whatever. >>>>> >>>>> A function doesn't describe anything. Information artifact >>>>> describe things. >>>>> Sorry to be picky about words. It's my job. >>>>> If you look at the papers on BFO you will see that functions are >>>>> continuants and occurrents are disjoint from continuants. So >>>>> functions are not occurrents. If the GO is using terminology at >>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>> foundry meeting that we were all going to use a common upper level >>>>> ontology. >>>>> >>>>> >>>>> Functions are part of processes. >>>>> >>>>> Function*ings* (i.e. realizations of functions) are part of >>>>> processes. >>>>> >>>>> A variety of structures, encoded in a molecule, have the potential >>>>> to realize a function. >>>>> >>>>> Type error: processes don't have potential. (concluded by >>>>> substitution "potential to realize a function" + function is_a >>>>> process => "potential to realize a process". >>>>> >>>>> The error is that it is the functions that are the potentials. >>>>> Potentials are realized. The realizations are processes. >>>>> >>>>> We discussed this very recently with Michael and he agreed. >>>>> >>>>> That is a property of the molecule, not of the function. >>>>> >>>>> What is the reference of "That"? >>>>> >>>>> A molecule can have encoded the potential of engaging in several >>>>> functions. >>>>> >>>>> Absolutely. (as long as you rewrite "functions" -> >>>>> "functionings" (i.e. processes).) >>>>> >>>>> A molecule can bear multiple functions is the way we would say this. >>>>> >>>>> The confusion I think is that of the relationship between the gene >>>>> product and a molecular function, and the molecular function and >>>>> the biological process. >>>>> >>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>> between a molecular function and a gene product is "inheres". The >>>>> relation between molecular function and a biological process is >>>>> "realized_by". >>>>> >>>>> -Alan >>>>> >>>>> >>>>> >>>>> Judy >>>>> >>>>> >>>>> >>>>> >>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>> <alanruttenberg at gmail.com> wrote: >>>>> >>>>> The problem is that this still confuses function with process. >>>>> Functions are the things that the enzymes (using the term >>>>> broadly) have that make the processes happen. If one goes into >>>>> more detail, there are a variety of participants in the process >>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>> function in a particular process. >>>>> >>>>> Every function has a corresponding process, its "realization". >>>>> >>>>> Functions are not parts of processes. The realizations of >>>>> functions are. >>>>> >>>>> The functions have a physical basis in the structure of their >>>>> bearers. >>>>> >>>>> The functions exist before any process happens. A molecule may >>>>> have a function but never actually realize it. >>>>> >>>>> These distinctions have important consequences as people such as >>>>> myself work on adding more detail to the GO. The processes get >>>>> more detail by adding *participants*. Within those processes, when >>>>> there is a molecule like an enzyme that functions in a particular >>>>> way, it needs to be distinguished from the other participants in >>>>> some way. The way that it is distinguished is by saying it bears a >>>>> function. >>>>> >>>>> The heuristic that a molecular function is a "single step" process >>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>> that in such descriptions there is a single molecule that has a >>>>> function. The step is the process that surrounds execution of that >>>>> function. >>>>> >>>>> As I said, getting this distinction clear is essential for future >>>>> detailing of GO processes such as I and others do. Can we start to >>>>> get this right here and now? >>>>> >>>>> -Alan >>>>> >>>>> >>>>> >>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>> <midori at ebi.ac.uk> wrote: >>>>> Hi, >>>>> >>>>> We have two SF items in which merging function and process terms >>>>> looks like a viable solution to a problem. One is an overly- >>>>> specific, single-step "process" that is essentially equivalent to >>>>> a "function" (in the GO sense); the other is a "function" term >>>>> that is defined such that it doesn't describe a single activity, >>>>> and is essentially the same as a process term. >>>>> >>>>> SF 2864212 - merge MF term >>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>> 0031202 >>>>> >>>>> into BP term >>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>> >>>>> For this one, we would merge the MF term into the BP term, and >>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>> think that's all we would change. >>>>> >>>>> >>>> >>>>>> https://sourceforge.net/tracker/? >>>>>> >>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>> >>>>> >>>>> SF 2864271 - merge BP term >>>>> >>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>> >>>>> into MF term >>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>> >>>>> >>>> >>>>>> https://sourceforge.net/tracker/? >>>>>> >>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>> >>>>> In this case, the proposal is to merge the process term into the >>>>> function term, ensuring that the merged term has is_a ancestry >>>>> only in the MF ontology,and has a part_of link to a process term. >>>>> Specifically, the merged term would retain the name and ID from GO: >>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>> >>>>> Does anyone have any objections or other comments on these? >>>>> >>>>> thanks, >>>>> m >>>>> _______________________________________________ >>>>> Annotation mailing list >>>>> Annotation at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> Annotation mailing list >>>>> Annotation at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>> >>>>> >>>>> >>>> _______________________________________________ >>>> Ontology-editors mailing list >>>> Ontology-editors at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>> >>> -- >>> Dr Jane Lomax >>> GO Editorial Office >>> EMBL-EBI >>> Wellcome Trust Genome Campus >>> Hinxton >>> Cambridgeshire, UK >>> CB10 1SD >>> >>> p: +44 1223 492516 >>> f: +44 1223 494468 >>> >>> >>> >> >> >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> > > -- Dr Jane Lomax GO Editorial Office EMBL-EBI Wellcome Trust Genome Campus Hinxton Cambridgeshire, UK CB10 1SD p: +44 1223 492516 f: +44 1223 494468 From dph at informatics.jax.org Wed Oct 7 13:14:14 2009 From: dph at informatics.jax.org (David Hill) Date: Wed, 07 Oct 2009 16:14:14 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> Message-ID: <4ACCF696.2040408@informatics.jax.org> I think a protein tag has a function as a signal. D Jane Lomax wrote: > Hi David - got a bit confused for a minute there because I think your > #1 is equivalent to my #3. So to be clear, I'm talking about this one: > > gene product MFing biological > process > > The downside of this option is what we do with the terms in MF that > aren't occurents - protein tag and their ilk. Are you up for the two > nodes option? As an experiment I tried splitting MF into MFings and > true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. I > can send the file if anyone's interested. > > Jane > > On Wed, 7 Oct 2009, David Hill wrote: > >> Hi Everyone, >> >> I think the issue with the problematic function terms is that we >> described them as roles because we didn't take enough time to >> carefully think about what the gene products were actually doing. >> When I look at them, they are almost always cases where we wanted to >> assign a function because we knew the product functioned in a >> process. I think we can just define them as functions involved in >> processes, sometimes the functions are generic and sometimes they are >> not. Transcription factor activity is and example of one that is not. >> It is_a 'DNA binding activity' involved in 'regulation of >> transcription'. >> >> Here is what a biologist wants to know: >> >> Gene product X has the ability to carry out some kind of action >> (function), either passively (like actin) or actively (like histidine >> kinase). In the proper biological context, the gene product will >> carry out that action and that will be part of a biological process. >> >> >> In some cases, there is a requirement for more than one gene product >> for this elemental action to be carried out. The biological process >> describes the overall objective of the action. I cannot think of any >> case where a single gene product acts and has a biological effect, >> that is succeeds in achieving a biological objective. If that were >> the case, I think biology wouldn't work, because there is always some >> background activity as long as a molecule and substrates are present. >> This is the distinction between a function and a process in the >> misunderstood 'single step processes'. Even phosphorylation as a >> biological objective (that has some effect on a biological system) is >> the result of the action of the activity of many kinase molecules, or >> the repeated activity of individual kinase molecules. >> >> So there are three ways we can represent this: >> >> 1) gene product MFing >> biological process >> >> In this case the gene product is a continuant. The > carry out> relationship describes the relationship between the >> continuant and the MF which is an occurent. The biological process is >> an occurent. >> >> 2) gene product MF MFing >> biological process >> >> In this case the gene product is a continuant. The MF is a continuant >> that is an enduring quality of the gene product. The MFing is an >> occurent that is the realization of the MF. The biological process is >> an occurent. >> >> 3) gene product MF BP >> >> In this case the gene product is a continuant. The MF is a continuant >> that is an enduring quality of the gene product. The BP is the >> occurent that results from the realization of the MF. >> >> I think that the most rigorous way to represent this would be #2. >> But, I think the most practical way would be #1. When I talk to >> biologists, both annotators and non-annotators, their eyes glaze over >> as soon as I mention functions and functionings. The bottom line is >> they don't care about it, they don't care to think about it, and they >> are confused by it. They are confused because in the absence of the >> gene product <>MF relationship, all they care about is a function >> being executed as part of a biological process. They could care less >> about the neurotransmitters that never bind a receptor and have a >> potential to execute a function, but never execute it. >> >> However, if we get rid of the MF<>BP side, they care about what the >> potential of the gene products are. So #1 works for this view. The MF >> being an occurent is the intuitive way that we think about this. MFs >> being parts of processes fits with the view I expressed above. >> >> #3 is not practical because it would require putting every MF into >> the BP ontology as an occurent. This would essentially recapitulate >> the entire MF ontology in BP. It is much easier for us to use #1. The >> key is in the relationship which >> combines the ideas of 'has potential' and 'carries out'. >> >> My 2 cents. >> >> David >> >> Judith Blake wrote: >>> I agree with Suzi, I support the continuation of MF, always have. >>> I don't have an issue with retaining the word 'activity'. >>> I find utility in describing the MFs, annotating the gp as having >>> the potential of realizing a MF. >>> >>> I see the utility of providing explicit relation between some MF >>> terms and biological processes. This will be a long journey. >>> I see continued utility of providing 'function' and 'process' sets >>> for the many applications and users currently use MF and BP in that >>> way. >>> >>> I don't think we will ever fully provide the relations between MF >>> and BP if only because so much isn't known about how molecular >>> function(ings) contribute to many biological processes. >>> >>> Judy >>> >>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>> >>> >>> >>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>> >>> >>>> I think the problem here is with trying to characterise MF as either >>>> *all* functions or *all* processes (i.e functionings) >>>> >>> >>> >>>> when in fact MF contains a combination of both. >>>> >>>> We've been back and forth over this many times and this seems to be >>>> the crux of it. >>>> >>>> Some MF terms fit happily as *parts* of processes (functionings or >>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>>> tag' does not unfold over time whichever way you look at it. I think >>>> the strategy of ignoring the existence of this latter class and >>>> saying everything in MF is a BFO process isn't really satisfactory >>>> in the long term. >>>> >>>> As I see it we have the following options to fix this: >>>> >>>> 1. Take away the word 'activity' from the term names and call >>>> everything in MF a BFO function. Make realised_in (or >>>> participates_in?) relations between MF and BP. We'd need to think >>>> carefully about how annotations would be propagated over realised_in. >>>> >>> >>> This is what was proposed in St. Croix and it makes sense to me, but >>> it does lead us into some unnecessarily confusing areas (i.e. having >>> yet another relationship to be added). That is why I am now convinced >>> that #3 the best approach. >>> >>> >>>> 2. Call everything in MF a BFO process and remove any terms that >>>> don't fit as processes e.g. protein tag. We lose a lot of >>>> annotations this way. We could possibly move them to a different >>>> ontology but would mean lots of re-annotation. >>>> >>> >>> Strongly against this one. >>> >>> >>>> 3. Divide MF into BFO processes and BFO functions, either both under >>>> MF or move the BFO processes into BP and keep MF for functions. The >>>> disadvantage here is that we wouldn't have the function for e.g. >>>> allene-oxide cyclase - we'd only have the process. You could I >>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>> cyclase activity' but that gets pretty confusing. >>>> >>> >>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep >>> both under MF. Least amount of change for users, while still solving >>> the problem. >>> >>> There is also 3 (c) which Amelia proposed which is to divide and put >>> the structural terms in an entirely separate new 4th branch, but I >>> think a simple divide at the top of MF is enough. >>> >>> In any case can we please, please please, get rid of the string >>> "activity". It adds nothing and is quite gratuitous. >>> >>> >>>> 4. Ignore the problem and hope it goes away (it won't). >>>> >>> >>> Nope, we've got to fix this. >>> >>> >>>> I don't really like 2 or 4. >>>> >>>> Jane >>>> >>>> >>>> Barry Smith wrote: >>>> >>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>> >>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>> actions' represented in the molecular function ontology. Ontology- >>>>>> speak. >>>>>> >>>>> Great. >>>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>>> This is not just ontological persnicketiness. There are already >>>>> people who are annotating with both 'function' and 'functioning' >>>>> e.g. using OBI to describe experiments, where it is important to >>>>> record both the functions e.g. of lab instruments, and the >>>>> processes in which they participate. The former exist (as >>>>> potentials) even when the relevant instrument is switched off. >>>>> There is also parallel work on physiology, distinguishing e.g. the >>>>> function of the heart (to pump), from realizations of this function >>>>> in functionings (pumpings), and both of these in turn from non- >>>>> functional processes such as thumpings, fibrillatings, and so forth. >>>>> If the GO were to entrench, now, a confusion of function and >>>>> functioning it will not work work well with these activities in the >>>>> future. >>>>> Barry >>>>> >>>>> >>>>>> Yes what is being implemented is what was originally proposed >>>>>> earlier. Clarifying that the gene product has the potential and >>>>>> the MF terms describes the specific action is what I was after. >>>>>> judy >>>>>> >>>>>> >>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>> <suzi at fruitfly.org> wrote: >>>>>> >>>>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>>>> >>>>>> But I think that the solution that David, Karen, and Midori have >>>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>>> things) the "Function*ings* are part of processes" into a single >>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>> for Function*ings* itself, it is implicit) that is for now called >>>>>> "part_of" (if I remember correctly). The name of this relationship >>>>>> may not be ideal, but we are trying to be clear about its intended >>>>>> usage. >>>>>> >>>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>>> seem to remember that David and I talked about this (and many >>>>>> other things, such as the nature of the implied relationship >>>>>> between an instance and a type when anyone makes an annotation) >>>>>> again at the relation ontology meeting in Denver. >>>>>> >>>>>> -S >>>>>> >>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>> >>>>>> >>>>>> >>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>> <Judith.Blake at jax.org> wrote: >>>>>> >>>>>> Functions are occurrants. A function describes an action, >>>>>> biochemical, structural, whatever. >>>>>> >>>>>> A function doesn't describe anything. Information artifact >>>>>> describe things. >>>>>> Sorry to be picky about words. It's my job. >>>>>> If you look at the papers on BFO you will see that functions are >>>>>> continuants and occurrents are disjoint from continuants. So >>>>>> functions are not occurrents. If the GO is using terminology at >>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>> foundry meeting that we were all going to use a common upper level >>>>>> ontology. >>>>>> >>>>>> >>>>>> Functions are part of processes. >>>>>> >>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>> processes. >>>>>> >>>>>> A variety of structures, encoded in a molecule, have the potential >>>>>> to realize a function. >>>>>> >>>>>> Type error: processes don't have potential. (concluded by >>>>>> substitution "potential to realize a function" + function is_a >>>>>> process => "potential to realize a process". >>>>>> >>>>>> The error is that it is the functions that are the potentials. >>>>>> Potentials are realized. The realizations are processes. >>>>>> >>>>>> We discussed this very recently with Michael and he agreed. >>>>>> >>>>>> That is a property of the molecule, not of the function. >>>>>> >>>>>> What is the reference of "That"? >>>>>> >>>>>> A molecule can have encoded the potential of engaging in several >>>>>> functions. >>>>>> >>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>> "functionings" (i.e. processes).) >>>>>> >>>>>> A molecule can bear multiple functions is the way we would say this. >>>>>> >>>>>> The confusion I think is that of the relationship between the gene >>>>>> product and a molecular function, and the molecular function and >>>>>> the biological process. >>>>>> >>>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>>> between a molecular function and a gene product is "inheres". The >>>>>> relation between molecular function and a biological process is >>>>>> "realized_by". >>>>>> >>>>>> -Alan >>>>>> >>>>>> >>>>>> >>>>>> Judy >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>> <alanruttenberg at gmail.com> wrote: >>>>>> >>>>>> The problem is that this still confuses function with process. >>>>>> Functions are the things that the enzymes (using the term >>>>>> broadly) have that make the processes happen. If one goes into >>>>>> more detail, there are a variety of participants in the process >>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>> function in a particular process. >>>>>> >>>>>> Every function has a corresponding process, its "realization". >>>>>> >>>>>> Functions are not parts of processes. The realizations of >>>>>> functions are. >>>>>> >>>>>> The functions have a physical basis in the structure of their >>>>>> bearers. >>>>>> >>>>>> The functions exist before any process happens. A molecule may >>>>>> have a function but never actually realize it. >>>>>> >>>>>> These distinctions have important consequences as people such as >>>>>> myself work on adding more detail to the GO. The processes get >>>>>> more detail by adding *participants*. Within those processes, when >>>>>> there is a molecule like an enzyme that functions in a particular >>>>>> way, it needs to be distinguished from the other participants in >>>>>> some way. The way that it is distinguished is by saying it bears a >>>>>> function. >>>>>> >>>>>> The heuristic that a molecular function is a "single step" process >>>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>>> that in such descriptions there is a single molecule that has a >>>>>> function. The step is the process that surrounds execution of that >>>>>> function. >>>>>> >>>>>> As I said, getting this distinction clear is essential for future >>>>>> detailing of GO processes such as I and others do. Can we start to >>>>>> get this right here and now? >>>>>> >>>>>> -Alan >>>>>> >>>>>> >>>>>> >>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>> <midori at ebi.ac.uk> wrote: >>>>>> Hi, >>>>>> >>>>>> We have two SF items in which merging function and process terms >>>>>> looks like a viable solution to a problem. One is an overly- >>>>>> specific, single-step "process" that is essentially equivalent to >>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>> that is defined such that it doesn't describe a single activity, >>>>>> and is essentially the same as a process term. >>>>>> >>>>>> SF 2864212 - merge MF term >>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>> 0031202 >>>>>> >>>>>> into BP term >>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>> >>>>>> For this one, we would merge the MF term into the BP term, and >>>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>>> think that's all we would change. >>>>>> >>>>>> >>>>> >>>>>> >>>>>>> https://sourceforge.net/tracker/? >>>>>>> >>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>> >>>>>> >>>>>> SF 2864271 - merge BP term >>>>>> >>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>> >>>>>> into MF term >>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>> >>>>>> >>>>> >>>>>> >>>>>>> https://sourceforge.net/tracker/? >>>>>>> >>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>> >>>>>> In this case, the proposal is to merge the process term into the >>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>> only in the MF ontology,and has a part_of link to a process term. >>>>>> Specifically, the merged term would retain the name and ID from GO: >>>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>>> >>>>>> Does anyone have any objections or other comments on these? >>>>>> >>>>>> thanks, >>>>>> m >>>>>> _______________________________________________ >>>>>> Annotation mailing list >>>>>> Annotation at geneontology.org >>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>> >>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Annotation mailing list >>>>>> Annotation at geneontology.org >>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>> >>>>>> >>>>>> >>>>> _______________________________________________ >>>>> Ontology-editors mailing list >>>>> Ontology-editors at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>> >>>> -- >>>> Dr Jane Lomax >>>> GO Editorial Office >>>> EMBL-EBI >>>> Wellcome Trust Genome Campus >>>> Hinxton >>>> Cambridgeshire, UK >>>> CB10 1SD >>>> >>>> p: +44 1223 492516 >>>> f: +44 1223 494468 >>>> >>>> >>>> >>> >>> >>> >>> >>> ------------------------------------------------------------------------ >>> >>> >>> _______________________________________________ >>> Annotation mailing list >>> Annotation at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/annotation >>> >> >> > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From jane at ebi.ac.uk Wed Oct 7 13:19:32 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Wed, 7 Oct 2009 21:19:32 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACCF696.2040408@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> Message-ID: Right - but it's not an occurent On Wed, 7 Oct 2009, David Hill wrote: > I think a protein tag has a function as a signal. > > D > > Jane Lomax wrote: >> Hi David - got a bit confused for a minute there because I think your #1 is >> equivalent to my #3. So to be clear, I'm talking about this one: >> >> gene product MFing biological >> process >> >> The downside of this option is what we do with the terms in MF that aren't >> occurents - protein tag and their ilk. Are you up for the two nodes option? >> As an experiment I tried splitting MF into MFings and true MFs. It comes >> out roughly as about 1/8 true MFs v/s 7/8 MFings. I can send the file if >> anyone's interested. >> >> Jane >> >> On Wed, 7 Oct 2009, David Hill wrote: >> >>> Hi Everyone, >>> >>> I think the issue with the problematic function terms is that we described >>> them as roles because we didn't take enough time to carefully think about >>> what the gene products were actually doing. When I look at them, they are >>> almost always cases where we wanted to assign a function because we knew >>> the product functioned in a process. I think we can just define them as >>> functions involved in processes, sometimes the functions are generic and >>> sometimes they are not. Transcription factor activity is and example of >>> one that is not. It is_a 'DNA binding activity' involved in 'regulation of >>> transcription'. >>> >>> Here is what a biologist wants to know: >>> >>> Gene product X has the ability to carry out some kind of action >>> (function), either passively (like actin) or actively (like histidine >>> kinase). In the proper biological context, the gene product will carry out >>> that action and that will be part of a biological process. >>> >>> >>> In some cases, there is a requirement for more than one gene product for >>> this elemental action to be carried out. The biological process describes >>> the overall objective of the action. I cannot think of any case where a >>> single gene product acts and has a biological effect, that is succeeds in >>> achieving a biological objective. If that were the case, I think biology >>> wouldn't work, because there is always some background activity as long as >>> a molecule and substrates are present. This is the distinction between a >>> function and a process in the misunderstood 'single step processes'. Even >>> phosphorylation as a biological objective (that has some effect on a >>> biological system) is the result of the action of the activity of many >>> kinase molecules, or the repeated activity of individual kinase molecules. >>> >>> So there are three ways we can represent this: >>> >>> 1) gene product MFing biological >>> process >>> >>> In this case the gene product is a continuant. The >> out> relationship describes the relationship between the continuant and >>> the MF which is an occurent. The biological process is an occurent. >>> >>> 2) gene product MF MFing >>> biological process >>> >>> In this case the gene product is a continuant. The MF is a continuant that >>> is an enduring quality of the gene product. The MFing is an occurent that >>> is the realization of the MF. The biological process is an occurent. >>> >>> 3) gene product MF BP >>> >>> In this case the gene product is a continuant. The MF is a continuant that >>> is an enduring quality of the gene product. The BP is the occurent that >>> results from the realization of the MF. >>> >>> I think that the most rigorous way to represent this would be #2. But, I >>> think the most practical way would be #1. When I talk to biologists, both >>> annotators and non-annotators, their eyes glaze over as soon as I mention >>> functions and functionings. The bottom line is they don't care about it, >>> they don't care to think about it, and they are confused by it. They are >>> confused because in the absence of the gene product <>MF relationship, all >>> they care about is a function being executed as part of a biological >>> process. They could care less about the neurotransmitters that never bind >>> a receptor and have a potential to execute a function, but never execute >>> it. >>> >>> However, if we get rid of the MF<>BP side, they care about what the >>> potential of the gene products are. So #1 works for this view. The MF >>> being an occurent is the intuitive way that we think about this. MFs being >>> parts of processes fits with the view I expressed above. >>> >>> #3 is not practical because it would require putting every MF into the BP >>> ontology as an occurent. This would essentially recapitulate the entire MF >>> ontology in BP. It is much easier for us to use #1. The key is in the >>> relationship which combines the ideas of 'has >>> potential' and 'carries out'. >>> >>> My 2 cents. >>> >>> David >>> >>> Judith Blake wrote: >>>> I agree with Suzi, I support the continuation of MF, always have. >>>> I don't have an issue with retaining the word 'activity'. >>>> I find utility in describing the MFs, annotating the gp as having the >>>> potential of realizing a MF. >>>> >>>> I see the utility of providing explicit relation between some MF terms >>>> and biological processes. This will be a long journey. >>>> I see continued utility of providing 'function' and 'process' sets for >>>> the many applications and users currently use MF and BP in that way. >>>> >>>> I don't think we will ever fully provide the relations between MF and BP >>>> if only because so much isn't known about how molecular function(ings) >>>> contribute to many biological processes. >>>> >>>> Judy >>>> >>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>> >>>> >>>> >>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>> >>>> >>>>> I think the problem here is with trying to characterise MF as either >>>>> *all* functions or *all* processes (i.e functionings) >>>>> >>>> >>>> >>>>> when in fact MF contains a combination of both. >>>>> >>>>> We've been back and forth over this many times and this seems to be >>>>> the crux of it. >>>>> >>>>> Some MF terms fit happily as *parts* of processes (functionings or >>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>>>> tag' does not unfold over time whichever way you look at it. I think >>>>> the strategy of ignoring the existence of this latter class and >>>>> saying everything in MF is a BFO process isn't really satisfactory >>>>> in the long term. >>>>> >>>>> As I see it we have the following options to fix this: >>>>> >>>>> 1. Take away the word 'activity' from the term names and call >>>>> everything in MF a BFO function. Make realised_in (or >>>>> participates_in?) relations between MF and BP. We'd need to think >>>>> carefully about how annotations would be propagated over realised_in. >>>>> >>>> >>>> This is what was proposed in St. Croix and it makes sense to me, but >>>> it does lead us into some unnecessarily confusing areas (i.e. having >>>> yet another relationship to be added). That is why I am now convinced >>>> that #3 the best approach. >>>> >>>> >>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>> annotations this way. We could possibly move them to a different >>>>> ontology but would mean lots of re-annotation. >>>>> >>>> >>>> Strongly against this one. >>>> >>>> >>>>> 3. Divide MF into BFO processes and BFO functions, either both under >>>>> MF or move the BFO processes into BP and keep MF for functions. The >>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>> cyclase activity' but that gets pretty confusing. >>>>> >>>> >>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep >>>> both under MF. Least amount of change for users, while still solving >>>> the problem. >>>> >>>> There is also 3 (c) which Amelia proposed which is to divide and put >>>> the structural terms in an entirely separate new 4th branch, but I >>>> think a simple divide at the top of MF is enough. >>>> >>>> In any case can we please, please please, get rid of the string >>>> "activity". It adds nothing and is quite gratuitous. >>>> >>>> >>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>> >>>> >>>> Nope, we've got to fix this. >>>> >>>> >>>>> I don't really like 2 or 4. >>>>> >>>>> Jane >>>>> >>>>> >>>>> Barry Smith wrote: >>>>> >>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>> >>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>> actions' represented in the molecular function ontology. Ontology- >>>>>>> speak. >>>>>>> >>>>>> Great. >>>>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>>>> This is not just ontological persnicketiness. There are already >>>>>> people who are annotating with both 'function' and 'functioning' >>>>>> e.g. using OBI to describe experiments, where it is important to >>>>>> record both the functions e.g. of lab instruments, and the >>>>>> processes in which they participate. The former exist (as >>>>>> potentials) even when the relevant instrument is switched off. >>>>>> There is also parallel work on physiology, distinguishing e.g. the >>>>>> function of the heart (to pump), from realizations of this function >>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>> functional processes such as thumpings, fibrillatings, and so forth. >>>>>> If the GO were to entrench, now, a confusion of function and >>>>>> functioning it will not work work well with these activities in the >>>>>> future. >>>>>> Barry >>>>>> >>>>>> >>>>>>> Yes what is being implemented is what was originally proposed >>>>>>> earlier. Clarifying that the gene product has the potential and >>>>>>> the MF terms describes the specific action is what I was after. >>>>>>> judy >>>>>>> >>>>>>> >>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>> <suzi at fruitfly.org> wrote: >>>>>>> >>>>>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>>>>> >>>>>>> But I think that the solution that David, Karen, and Midori have >>>>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>>>> things) the "Function*ings* are part of processes" into a single >>>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>>> for Function*ings* itself, it is implicit) that is for now called >>>>>>> "part_of" (if I remember correctly). The name of this relationship >>>>>>> may not be ideal, but we are trying to be clear about its intended >>>>>>> usage. >>>>>>> >>>>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>>>> seem to remember that David and I talked about this (and many >>>>>>> other things, such as the nature of the implied relationship >>>>>>> between an instance and a type when anyone makes an annotation) >>>>>>> again at the relation ontology meeting in Denver. >>>>>>> >>>>>>> -S >>>>>>> >>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>> >>>>>>> Functions are occurrants. A function describes an action, >>>>>>> biochemical, structural, whatever. >>>>>>> >>>>>>> A function doesn't describe anything. Information artifact >>>>>>> describe things. >>>>>>> Sorry to be picky about words. It's my job. >>>>>>> If you look at the papers on BFO you will see that functions are >>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>> functions are not occurrents. If the GO is using terminology at >>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>> foundry meeting that we were all going to use a common upper level >>>>>>> ontology. >>>>>>> >>>>>>> >>>>>>> Functions are part of processes. >>>>>>> >>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>> processes. >>>>>>> >>>>>>> A variety of structures, encoded in a molecule, have the potential >>>>>>> to realize a function. >>>>>>> >>>>>>> Type error: processes don't have potential. (concluded by >>>>>>> substitution "potential to realize a function" + function is_a >>>>>>> process => "potential to realize a process". >>>>>>> >>>>>>> The error is that it is the functions that are the potentials. >>>>>>> Potentials are realized. The realizations are processes. >>>>>>> >>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>> >>>>>>> That is a property of the molecule, not of the function. >>>>>>> >>>>>>> What is the reference of "That"? >>>>>>> >>>>>>> A molecule can have encoded the potential of engaging in several >>>>>>> functions. >>>>>>> >>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>> "functionings" (i.e. processes).) >>>>>>> >>>>>>> A molecule can bear multiple functions is the way we would say this. >>>>>>> >>>>>>> The confusion I think is that of the relationship between the gene >>>>>>> product and a molecular function, and the molecular function and >>>>>>> the biological process. >>>>>>> >>>>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>>>> between a molecular function and a gene product is "inheres". The >>>>>>> relation between molecular function and a biological process is >>>>>>> "realized_by". >>>>>>> >>>>>>> -Alan >>>>>>> >>>>>>> >>>>>>> >>>>>>> Judy >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>> >>>>>>> The problem is that this still confuses function with process. >>>>>>> Functions are the things that the enzymes (using the term >>>>>>> broadly) have that make the processes happen. If one goes into >>>>>>> more detail, there are a variety of participants in the process >>>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>>> function in a particular process. >>>>>>> >>>>>>> Every function has a corresponding process, its "realization". >>>>>>> >>>>>>> Functions are not parts of processes. The realizations of >>>>>>> functions are. >>>>>>> >>>>>>> The functions have a physical basis in the structure of their >>>>>>> bearers. >>>>>>> >>>>>>> The functions exist before any process happens. A molecule may >>>>>>> have a function but never actually realize it. >>>>>>> >>>>>>> These distinctions have important consequences as people such as >>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>> more detail by adding *participants*. Within those processes, when >>>>>>> there is a molecule like an enzyme that functions in a particular >>>>>>> way, it needs to be distinguished from the other participants in >>>>>>> some way. The way that it is distinguished is by saying it bears a >>>>>>> function. >>>>>>> >>>>>>> The heuristic that a molecular function is a "single step" process >>>>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>>>> that in such descriptions there is a single molecule that has a >>>>>>> function. The step is the process that surrounds execution of that >>>>>>> function. >>>>>>> >>>>>>> As I said, getting this distinction clear is essential for future >>>>>>> detailing of GO processes such as I and others do. Can we start to >>>>>>> get this right here and now? >>>>>>> >>>>>>> -Alan >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>> Hi, >>>>>>> >>>>>>> We have two SF items in which merging function and process terms >>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>> specific, single-step "process" that is essentially equivalent to >>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>> that is defined such that it doesn't describe a single activity, >>>>>>> and is essentially the same as a process term. >>>>>>> >>>>>>> SF 2864212 - merge MF term >>>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>>> 0031202 >>>>>>> >>>>>>> into BP term >>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>> >>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>>>> think that's all we would change. >>>>>>> >>>>>>> >>>>>> >>>>>>>> https://sourceforge.net/tracker/? >>>>>>>> >>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>> >>>>>>> >>>>>>> SF 2864271 - merge BP term >>>>>>> >>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>> >>>>>>> into MF term >>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>> >>>>>>> >>>>>> >>>>>>>> https://sourceforge.net/tracker/? >>>>>>>> >>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>> >>>>>>> In this case, the proposal is to merge the process term into the >>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>> only in the MF ontology,and has a part_of link to a process term. >>>>>>> Specifically, the merged term would retain the name and ID from GO: >>>>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>>>> >>>>>>> Does anyone have any objections or other comments on these? >>>>>>> >>>>>>> thanks, >>>>>>> m >>>>>>> _______________________________________________ >>>>>>> Annotation mailing list >>>>>>> Annotation at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>> >>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Annotation mailing list >>>>>>> Annotation at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>> >>>>>>> >>>>>>> >>>>>> _______________________________________________ >>>>>> Ontology-editors mailing list >>>>>> Ontology-editors at geneontology.org >>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>> >>>>> -- >>>>> Dr Jane Lomax >>>>> GO Editorial Office >>>>> EMBL-EBI >>>>> Wellcome Trust Genome Campus >>>>> Hinxton >>>>> Cambridgeshire, UK >>>>> CB10 1SD >>>>> >>>>> p: +44 1223 492516 >>>>> f: +44 1223 494468 >>>>> >>>>> >>>>> >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>> >>> >> > > -- Dr Jane Lomax GO Editorial Office EMBL-EBI Wellcome Trust Genome Campus Hinxton Cambridgeshire, UK CB10 1SD p: +44 1223 492516 f: +44 1223 494468 From dph at informatics.jax.org Wed Oct 7 13:24:03 2009 From: dph at informatics.jax.org (David Hill) Date: Wed, 07 Oct 2009 16:24:03 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> Message-ID: <4ACCF8E3.4080708@informatics.jax.org> When it executes its role as a signal, it is signaling. D Jane Lomax wrote: > Right - but it's not an occurent > > On Wed, 7 Oct 2009, David Hill wrote: > >> I think a protein tag has a function as a signal. >> >> D >> >> Jane Lomax wrote: >>> Hi David - got a bit confused for a minute there because I think >>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>> this one: >>> >>> gene product MFing biological >>> process >>> >>> The downside of this option is what we do with the terms in MF that >>> aren't occurents - protein tag and their ilk. Are you up for the two >>> nodes option? As an experiment I tried splitting MF into MFings and >>> true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. >>> I can send the file if anyone's interested. >>> >>> Jane >>> >>> On Wed, 7 Oct 2009, David Hill wrote: >>> >>>> Hi Everyone, >>>> >>>> I think the issue with the problematic function terms is that we >>>> described them as roles because we didn't take enough time to >>>> carefully think about what the gene products were actually doing. >>>> When I look at them, they are almost always cases where we wanted >>>> to assign a function because we knew the product functioned in a >>>> process. I think we can just define them as functions involved in >>>> processes, sometimes the functions are generic and sometimes they >>>> are not. Transcription factor activity is and example of one that >>>> is not. It is_a 'DNA binding activity' involved in 'regulation of >>>> transcription'. >>>> >>>> Here is what a biologist wants to know: >>>> >>>> Gene product X has the ability to carry out some kind of action >>>> (function), either passively (like actin) or actively (like >>>> histidine kinase). In the proper biological context, the gene >>>> product will carry out that action and that will be part of a >>>> biological process. >>>> >>>> >>>> In some cases, there is a requirement for more than one gene >>>> product for this elemental action to be carried out. The biological >>>> process describes the overall objective of the action. I cannot >>>> think of any case where a single gene product acts and has a >>>> biological effect, that is succeeds in achieving a biological >>>> objective. If that were the case, I think biology wouldn't work, >>>> because there is always some background activity as long as a >>>> molecule and substrates are present. This is the distinction >>>> between a function and a process in the misunderstood 'single step >>>> processes'. Even phosphorylation as a biological objective (that >>>> has some effect on a biological system) is the result of the action >>>> of the activity of many kinase molecules, or the repeated activity >>>> of individual kinase molecules. >>>> >>>> So there are three ways we can represent this: >>>> >>>> 1) gene product MFing >>>> biological process >>>> >>>> In this case the gene product is a continuant. The >>> to carry out> relationship describes the relationship between the >>>> continuant and the MF which is an occurent. The biological process >>>> is an occurent. >>>> >>>> 2) gene product MF MFing >>>> biological process >>>> >>>> In this case the gene product is a continuant. The MF is a >>>> continuant that is an enduring quality of the gene product. The >>>> MFing is an occurent that is the realization of the MF. The >>>> biological process is an occurent. >>>> >>>> 3) gene product MF BP >>>> >>>> In this case the gene product is a continuant. The MF is a >>>> continuant that is an enduring quality of the gene product. The BP >>>> is the occurent that results from the realization of the MF. >>>> >>>> I think that the most rigorous way to represent this would be #2. >>>> But, I think the most practical way would be #1. When I talk to >>>> biologists, both annotators and non-annotators, their eyes glaze >>>> over as soon as I mention functions and functionings. The bottom >>>> line is they don't care about it, they don't care to think about >>>> it, and they are confused by it. They are confused because in the >>>> absence of the gene product <>MF relationship, all they care about >>>> is a function being executed as part of a biological process. They >>>> could care less about the neurotransmitters that never bind a >>>> receptor and have a potential to execute a function, but never >>>> execute it. >>>> >>>> However, if we get rid of the MF<>BP side, they care about what the >>>> potential of the gene products are. So #1 works for this view. The >>>> MF being an occurent is the intuitive way that we think about this. >>>> MFs being parts of processes fits with the view I expressed above. >>>> >>>> #3 is not practical because it would require putting every MF into >>>> the BP ontology as an occurent. This would essentially recapitulate >>>> the entire MF ontology in BP. It is much easier for us to use #1. >>>> The key is in the relationship which >>>> combines the ideas of 'has potential' and 'carries out'. >>>> >>>> My 2 cents. >>>> >>>> David >>>> >>>> Judith Blake wrote: >>>>> I agree with Suzi, I support the continuation of MF, always have. >>>>> I don't have an issue with retaining the word 'activity'. >>>>> I find utility in describing the MFs, annotating the gp as having >>>>> the potential of realizing a MF. >>>>> >>>>> I see the utility of providing explicit relation between some MF >>>>> terms and biological processes. This will be a long journey. >>>>> I see continued utility of providing 'function' and 'process' sets >>>>> for the many applications and users currently use MF and BP in >>>>> that way. >>>>> >>>>> I don't think we will ever fully provide the relations between MF >>>>> and BP if only because so much isn't known about how molecular >>>>> function(ings) contribute to many biological processes. >>>>> >>>>> Judy >>>>> >>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>> >>>>> >>>>> >>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>> >>>>> >>>>>> I think the problem here is with trying to characterise MF as either >>>>>> *all* functions or *all* processes (i.e functionings) >>>>>> >>>>> >>>>> >>>>>> when in fact MF contains a combination of both. >>>>>> >>>>>> We've been back and forth over this many times and this seems to be >>>>>> the crux of it. >>>>>> >>>>>> Some MF terms fit happily as *parts* of processes (functionings or >>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>>>>> tag' does not unfold over time whichever way you look at it. I think >>>>>> the strategy of ignoring the existence of this latter class and >>>>>> saying everything in MF is a BFO process isn't really satisfactory >>>>>> in the long term. >>>>>> >>>>>> As I see it we have the following options to fix this: >>>>>> >>>>>> 1. Take away the word 'activity' from the term names and call >>>>>> everything in MF a BFO function. Make realised_in (or >>>>>> participates_in?) relations between MF and BP. We'd need to think >>>>>> carefully about how annotations would be propagated over >>>>>> realised_in. >>>>>> >>>>> >>>>> This is what was proposed in St. Croix and it makes sense to me, but >>>>> it does lead us into some unnecessarily confusing areas (i.e. having >>>>> yet another relationship to be added). That is why I am now convinced >>>>> that #3 the best approach. >>>>> >>>>> >>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>> annotations this way. We could possibly move them to a different >>>>>> ontology but would mean lots of re-annotation. >>>>>> >>>>> >>>>> Strongly against this one. >>>>> >>>>> >>>>>> 3. Divide MF into BFO processes and BFO functions, either both under >>>>>> MF or move the BFO processes into BP and keep MF for functions. The >>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>> cyclase activity' but that gets pretty confusing. >>>>>> >>>>> >>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and >>>>> keep >>>>> both under MF. Least amount of change for users, while still solving >>>>> the problem. >>>>> >>>>> There is also 3 (c) which Amelia proposed which is to divide and put >>>>> the structural terms in an entirely separate new 4th branch, but I >>>>> think a simple divide at the top of MF is enough. >>>>> >>>>> In any case can we please, please please, get rid of the string >>>>> "activity". It adds nothing and is quite gratuitous. >>>>> >>>>> >>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>> >>>>> >>>>> Nope, we've got to fix this. >>>>> >>>>> >>>>>> I don't really like 2 or 4. >>>>>> >>>>>> Jane >>>>>> >>>>>> >>>>>> Barry Smith wrote: >>>>>> >>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>> >>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>> actions' represented in the molecular function ontology. >>>>>>>> Ontology- >>>>>>>> speak. >>>>>>>> >>>>>>> Great. >>>>>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>>>>> This is not just ontological persnicketiness. There are already >>>>>>> people who are annotating with both 'function' and 'functioning' >>>>>>> e.g. using OBI to describe experiments, where it is important to >>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>> processes in which they participate. The former exist (as >>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>> There is also parallel work on physiology, distinguishing e.g. the >>>>>>> function of the heart (to pump), from realizations of this function >>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>> forth. >>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>> functioning it will not work work well with these activities in the >>>>>>> future. >>>>>>> Barry >>>>>>> >>>>>>> >>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>> earlier. Clarifying that the gene product has the potential and >>>>>>>> the MF terms describes the specific action is what I was after. >>>>>>>> judy >>>>>>>> >>>>>>>> >>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>> >>>>>>>> Alan has it right. Functions are not occurents, Function*ings* >>>>>>>> are. >>>>>>>> >>>>>>>> But I think that the solution that David, Karen, and Midori have >>>>>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>>>>> things) the "Function*ings* are part of processes" into a single >>>>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>>>> for Function*ings* itself, it is implicit) that is for now called >>>>>>>> "part_of" (if I remember correctly). The name of this relationship >>>>>>>> may not be ideal, but we are trying to be clear about its intended >>>>>>>> usage. >>>>>>>> >>>>>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>> other things, such as the nature of the implied relationship >>>>>>>> between an instance and a type when anyone makes an annotation) >>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>> >>>>>>>> -S >>>>>>>> >>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>> >>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>> biochemical, structural, whatever. >>>>>>>> >>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>> describe things. >>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>> If you look at the papers on BFO you will see that functions are >>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>> functions are not occurrents. If the GO is using terminology at >>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>> foundry meeting that we were all going to use a common upper level >>>>>>>> ontology. >>>>>>>> >>>>>>>> >>>>>>>> Functions are part of processes. >>>>>>>> >>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>> processes. >>>>>>>> >>>>>>>> A variety of structures, encoded in a molecule, have the potential >>>>>>>> to realize a function. >>>>>>>> >>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>> process => "potential to realize a process". >>>>>>>> >>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>> >>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>> >>>>>>>> That is a property of the molecule, not of the function. >>>>>>>> >>>>>>>> What is the reference of "That"? >>>>>>>> >>>>>>>> A molecule can have encoded the potential of engaging in several >>>>>>>> functions. >>>>>>>> >>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>> "functionings" (i.e. processes).) >>>>>>>> >>>>>>>> A molecule can bear multiple functions is the way we would say >>>>>>>> this. >>>>>>>> >>>>>>>> The confusion I think is that of the relationship between the gene >>>>>>>> product and a molecular function, and the molecular function and >>>>>>>> the biological process. >>>>>>>> >>>>>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>>>>> between a molecular function and a gene product is "inheres". The >>>>>>>> relation between molecular function and a biological process is >>>>>>>> "realized_by". >>>>>>>> >>>>>>>> -Alan >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> Judy >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>> >>>>>>>> The problem is that this still confuses function with process. >>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>> broadly) have that make the processes happen. If one goes into >>>>>>>> more detail, there are a variety of participants in the process >>>>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>>>> function in a particular process. >>>>>>>> >>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>> >>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>> functions are. >>>>>>>> >>>>>>>> The functions have a physical basis in the structure of their >>>>>>>> bearers. >>>>>>>> >>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>> have a function but never actually realize it. >>>>>>>> >>>>>>>> These distinctions have important consequences as people such as >>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>> more detail by adding *participants*. Within those processes, when >>>>>>>> there is a molecule like an enzyme that functions in a particular >>>>>>>> way, it needs to be distinguished from the other participants in >>>>>>>> some way. The way that it is distinguished is by saying it bears a >>>>>>>> function. >>>>>>>> >>>>>>>> The heuristic that a molecular function is a "single step" process >>>>>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>>>>> that in such descriptions there is a single molecule that has a >>>>>>>> function. The step is the process that surrounds execution of that >>>>>>>> function. >>>>>>>> >>>>>>>> As I said, getting this distinction clear is essential for future >>>>>>>> detailing of GO processes such as I and others do. Can we start to >>>>>>>> get this right here and now? >>>>>>>> >>>>>>>> -Alan >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>> Hi, >>>>>>>> >>>>>>>> We have two SF items in which merging function and process terms >>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>> specific, single-step "process" that is essentially equivalent to >>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>> that is defined such that it doesn't describe a single activity, >>>>>>>> and is essentially the same as a process term. >>>>>>>> >>>>>>>> SF 2864212 - merge MF term >>>>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>>>> 0031202 >>>>>>>> >>>>>>>> into BP term >>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>> >>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>>>>> think that's all we would change. >>>>>>>> >>>>>>>> >>>>>>> >>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>> >>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>> >>>>>>>> >>>>>>>> SF 2864271 - merge BP term >>>>>>>> >>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>> >>>>>>>> into MF term >>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>> >>>>>>>> >>>>>>> >>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>> >>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>> >>>>>>>> In this case, the proposal is to merge the process term into the >>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>> only in the MF ontology,and has a part_of link to a process term. >>>>>>>> Specifically, the merged term would retain the name and ID from >>>>>>>> GO: >>>>>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>>>>> >>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>> >>>>>>>> thanks, >>>>>>>> m >>>>>>>> _______________________________________________ >>>>>>>> Annotation mailing list >>>>>>>> Annotation at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Annotation mailing list >>>>>>>> Annotation at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> _______________________________________________ >>>>>>> Ontology-editors mailing list >>>>>>> Ontology-editors at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>> >>>>>> -- >>>>>> Dr Jane Lomax >>>>>> GO Editorial Office >>>>>> EMBL-EBI >>>>>> Wellcome Trust Genome Campus >>>>>> Hinxton >>>>>> Cambridgeshire, UK >>>>>> CB10 1SD >>>>>> >>>>>> p: +44 1223 492516 >>>>>> f: +44 1223 494468 >>>>>> >>>>>> >>>>>> >>>>> >>>>> >>>>> >>>>> >>>>> ------------------------------------------------------------------------ >>>>> >>>>> _______________________________________________ >>>>> Annotation mailing list >>>>> Annotation at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>> >>>> >>>> >>> >> >> > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From aji at ebi.ac.uk Wed Oct 7 13:53:47 2009 From: aji at ebi.ac.uk (Amelia Ireland) Date: Wed, 7 Oct 2009 21:53:47 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACCF8E3.4080708@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> Message-ID: <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> On Oct 7, 2009, at 9:24 PM, David Hill wrote: > When it executes its role as a signal, it is signaling. Perhaps it could be recast as something like 'X binding involved in signal transduction' or, to make it more explicit, 'signal transmission' or 'signalling' ('signaling' in the US ;) )? > Jane Lomax wrote: >> Right - but it's not an occurent >> >> On Wed, 7 Oct 2009, David Hill wrote: >> >>> I think a protein tag has a function as a signal. >>> >>> D >>> >>> Jane Lomax wrote: >>>> Hi David - got a bit confused for a minute there because I think >>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>> this one: >>>> >>>> gene product MFing >>>> biological process >>>> >>>> The downside of this option is what we do with the terms in MF >>>> that aren't occurents - protein tag and their ilk. Are you up for >>>> the two nodes option? As an experiment I tried splitting MF into >>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs v/ >>>> s 7/8 MFings. I can send the file if anyone's interested. >>>> >>>> Jane >>>> >>>> On Wed, 7 Oct 2009, David Hill wrote: >>>> >>>>> Hi Everyone, >>>>> >>>>> I think the issue with the problematic function terms is that we >>>>> described them as roles because we didn't take enough time to >>>>> carefully think about what the gene products were actually >>>>> doing. When I look at them, they are almost always cases where >>>>> we wanted to assign a function because we knew the product >>>>> functioned in a process. I think we can just define them as >>>>> functions involved in processes, sometimes the functions are >>>>> generic and sometimes they are not. Transcription factor >>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>> activity' involved in 'regulation of transcription'. >>>>> >>>>> Here is what a biologist wants to know: >>>>> >>>>> Gene product X has the ability to carry out some kind of action >>>>> (function), either passively (like actin) or actively (like >>>>> histidine kinase). In the proper biological context, the gene >>>>> product will carry out that action and that will be part of a >>>>> biological process. >>>>> >>>>> >>>>> In some cases, there is a requirement for more than one gene >>>>> product for this elemental action to be carried out. The >>>>> biological process describes the overall objective of the >>>>> action. I cannot think of any case where a single gene product >>>>> acts and has a biological effect, that is succeeds in achieving >>>>> a biological objective. If that were the case, I think biology >>>>> wouldn't work, because there is always some background activity >>>>> as long as a molecule and substrates are present. This is the >>>>> distinction between a function and a process in the >>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>> biological objective (that has some effect on a biological >>>>> system) is the result of the action of the activity of many >>>>> kinase molecules, or the repeated activity of individual kinase >>>>> molecules. >>>>> >>>>> So there are three ways we can represent this: >>>>> >>>>> 1) gene product MFing >>>>> biological process >>>>> >>>>> In this case the gene product is a continuant. The >>>> potential to carry out> relationship describes the relationship >>>>> between the continuant and the MF which is an occurent. The >>>>> biological process is an occurent. >>>>> >>>>> 2) gene product MF MFing >>>>> biological process >>>>> >>>>> In this case the gene product is a continuant. The MF is a >>>>> continuant that is an enduring quality of the gene product. The >>>>> MFing is an occurent that is the realization of the MF. The >>>>> biological process is an occurent. >>>>> >>>>> 3) gene product MF BP >>>>> >>>>> In this case the gene product is a continuant. The MF is a >>>>> continuant that is an enduring quality of the gene product. The >>>>> BP is the occurent that results from the realization of the MF. >>>>> >>>>> I think that the most rigorous way to represent this would be >>>>> #2. But, I think the most practical way would be #1. When I talk >>>>> to biologists, both annotators and non-annotators, their eyes >>>>> glaze over as soon as I mention functions and functionings. The >>>>> bottom line is they don't care about it, they don't care to >>>>> think about it, and they are confused by it. They are confused >>>>> because in the absence of the gene product <>MF relationship, >>>>> all they care about is a function being executed as part of a >>>>> biological process. They could care less about the >>>>> neurotransmitters that never bind a receptor and have a >>>>> potential to execute a function, but never execute it. >>>>> >>>>> However, if we get rid of the MF<>BP side, they care about what >>>>> the potential of the gene products are. So #1 works for this >>>>> view. The MF being an occurent is the intuitive way that we >>>>> think about this. MFs being parts of processes fits with the >>>>> view I expressed above. >>>>> >>>>> #3 is not practical because it would require putting every MF >>>>> into the BP ontology as an occurent. This would essentially >>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>> us to use #1. The key is in the relationship >>>> carry out> which combines the ideas of 'has potential' and >>>>> 'carries out'. >>>>> >>>>> My 2 cents. >>>>> >>>>> David >>>>> >>>>> Judith Blake wrote: >>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>> have. >>>>>> I don't have an issue with retaining the word 'activity'. >>>>>> I find utility in describing the MFs, annotating the gp as >>>>>> having the potential of realizing a MF. >>>>>> >>>>>> I see the utility of providing explicit relation between some >>>>>> MF terms and biological processes. This will be a long journey. >>>>>> I see continued utility of providing 'function' and 'process' >>>>>> sets for the many applications and users currently use MF and >>>>>> BP in that way. >>>>>> >>>>>> I don't think we will ever fully provide the relations between >>>>>> MF and BP if only because so much isn't known about how >>>>>> molecular function(ings) contribute to many biological processes. >>>>>> >>>>>> Judy >>>>>> >>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>> >>>>>> >>>>>> >>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>> >>>>>> >>>>>>> I think the problem here is with trying to characterise MF as >>>>>>> either >>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>> >>>>>> >>>>>> >>>>>>> when in fact MF contains a combination of both. >>>>>>> >>>>>>> We've been back and forth over this many times and this seems >>>>>>> to be >>>>>>> the crux of it. >>>>>>> >>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>> (functionings or >>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>> 'Protein >>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>> think >>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>> saying everything in MF is a BFO process isn't really >>>>>>> satisfactory >>>>>>> in the long term. >>>>>>> >>>>>>> As I see it we have the following options to fix this: >>>>>>> >>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>> think >>>>>>> carefully about how annotations would be propagated over >>>>>>> realised_in. >>>>>>> >>>>>> >>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>> me, but >>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>> having >>>>>> yet another relationship to be added). That is why I am now >>>>>> convinced >>>>>> that #3 the best approach. >>>>>> >>>>>> >>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>> annotations this way. We could possibly move them to a different >>>>>>> ontology but would mean lots of re-annotation. >>>>>>> >>>>>> >>>>>> Strongly against this one. >>>>>> >>>>>> >>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>> under >>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>> functions. The >>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>> >>>>>> >>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>> and keep >>>>>> both under MF. Least amount of change for users, while still >>>>>> solving >>>>>> the problem. >>>>>> >>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>> and put >>>>>> the structural terms in an entirely separate new 4th branch, >>>>>> but I >>>>>> think a simple divide at the top of MF is enough. >>>>>> >>>>>> In any case can we please, please please, get rid of the string >>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>> >>>>>> >>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>> >>>>>> >>>>>> Nope, we've got to fix this. >>>>>> >>>>>> >>>>>>> I don't really like 2 or 4. >>>>>>> >>>>>>> Jane >>>>>>> >>>>>>> >>>>>>> Barry Smith wrote: >>>>>>> >>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>> >>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>> Ontology- >>>>>>>>> speak. >>>>>>>>> >>>>>>>> Great. >>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>> Ontology' >>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>> people who are annotating with both 'function' and >>>>>>>> 'functioning' >>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>> to >>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>> processes in which they participate. The former exist (as >>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>> e.g. the >>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>> function >>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>> forth. >>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>> functioning it will not work work well with these activities >>>>>>>> in the >>>>>>>> future. >>>>>>>> Barry >>>>>>>> >>>>>>>> >>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>> and >>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>> after. >>>>>>>>> judy >>>>>>>>> >>>>>>>>> >>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>> >>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>> Function*ings* are. >>>>>>>>> >>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>> have >>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>> repeating >>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>> single >>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>> artifact >>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>> called >>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>> relationship >>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>> intended >>>>>>>>> usage. >>>>>>>>> >>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>> And I >>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>> annotation) >>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>> >>>>>>>>> -S >>>>>>>>> >>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>> >>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>> biochemical, structural, whatever. >>>>>>>>> >>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>> describe things. >>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>> are >>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>> at >>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>> level >>>>>>>>> ontology. >>>>>>>>> >>>>>>>>> >>>>>>>>> Functions are part of processes. >>>>>>>>> >>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>> processes. >>>>>>>>> >>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>> potential >>>>>>>>> to realize a function. >>>>>>>>> >>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>> process => "potential to realize a process". >>>>>>>>> >>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>> >>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>> >>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>> >>>>>>>>> What is the reference of "That"? >>>>>>>>> >>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>> several >>>>>>>>> functions. >>>>>>>>> >>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>> >>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>> say this. >>>>>>>>> >>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>> the gene >>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>> and >>>>>>>>> the biological process. >>>>>>>>> >>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>> relation >>>>>>>>> between a molecular function and a gene product is >>>>>>>>> "inheres". The >>>>>>>>> relation between molecular function and a biological process >>>>>>>>> is >>>>>>>>> "realized_by". >>>>>>>>> >>>>>>>>> -Alan >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> Judy >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>> >>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>> into >>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>> process >>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>> the >>>>>>>>> function in a particular process. >>>>>>>>> >>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>> >>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>> functions are. >>>>>>>>> >>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>> bearers. >>>>>>>>> >>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>> have a function but never actually realize it. >>>>>>>>> >>>>>>>>> These distinctions have important consequences as people >>>>>>>>> such as >>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>> processes, when >>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>> particular >>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>> participants in >>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>> bears a >>>>>>>>> function. >>>>>>>>> >>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>> process >>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>> fact >>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>> has a >>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>> of that >>>>>>>>> function. >>>>>>>>> >>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>> future >>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>> start to >>>>>>>>> get this right here and now? >>>>>>>>> >>>>>>>>> -Alan >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>> Hi, >>>>>>>>> >>>>>>>>> We have two SF items in which merging function and process >>>>>>>>> terms >>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>> equivalent to >>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>> activity, >>>>>>>>> and is essentially the same as a process term. >>>>>>>>> >>>>>>>>> SF 2864212 - merge MF term >>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>> GO: >>>>>>>>> 0031202 >>>>>>>>> >>>>>>>>> into BP term >>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>> >>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>> graph. I >>>>>>>>> think that's all we would change. >>>>>>>>> >>>>>>>>> >>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>> >>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>> >>>>>>>>> >>>>>>>>> SF 2864271 - merge BP term >>>>>>>>> >>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>> >>>>>>>>> into MF term >>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>> >>>>>>>>> >>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>> >>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>> >>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>> the >>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>> term. >>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>> from GO: >>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>> 0006413. >>>>>>>>> >>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>> >>>>>>>>> thanks, >>>>>>>>> m >>>>>>>>> _______________________________________________ >>>>>>>>> Annotation mailing list >>>>>>>>> Annotation at geneontology.org >>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Annotation mailing list >>>>>>>>> Annotation at geneontology.org >>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Ontology-editors mailing list >>>>>>>> Ontology-editors at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>> >>>>>>> -- >>>>>>> Dr Jane Lomax >>>>>>> GO Editorial Office >>>>>>> EMBL-EBI >>>>>>> Wellcome Trust Genome Campus >>>>>>> Hinxton >>>>>>> Cambridgeshire, UK >>>>>>> CB10 1SD >>>>>>> >>>>>>> p: +44 1223 492516 >>>>>>> f: +44 1223 494468 >>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------ >>>>>> _______________________________________________ >>>>>> Annotation mailing list >>>>>> Annotation at geneontology.org >>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>> >>>>> >>>>> >>>> >>> >>> >> > > -- > David P. Hill, Ph.D. > Bioinformatics Scientist: Ontology Development > Gene Ontology Consortium > The Jackson Laboratory > www.geneontology.org > www.informatics.jax.org > tel:207-288-6430 > > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors -- Amelia Ireland GO Editorial Office http://www.berkeleybop.org || http://www.ebi.ac.uk Boycott Trader Joe's Red List seafood: http://traitorjoe.com From aji at ebi.ac.uk Wed Oct 7 13:58:49 2009 From: aji at ebi.ac.uk (Amelia Ireland) Date: Wed, 7 Oct 2009 21:58:49 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> Message-ID: <10829031-FD46-4357-99DB-C142DA0CD692@ebi.ac.uk> On Oct 7, 2009, at 9:12 PM, Jane Lomax wrote: > Hi David - got a bit confused for a minute there because I think > your #1 is equivalent to my #3. So to be clear, I'm talking about > this one: > > gene product MFing biological > process I definitely support this proposal; I think it's the most practical and most sensible, and I think that having the MFings as part of biological processes will be intuitive for biologists and those who are used to (e.g.) pathways being made up of reactions (e.g. in metabolic pathways) or binding events (e.g. in signal[l]ing). > The downside of this option is what we do with the terms in MF that > aren't occurents - protein tag and their ilk. Are you up for the two > nodes option? As an experiment I tried splitting MF into MFings and > true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. > I can send the file if anyone's interested. Could you summarize the 1/8th non-MFings? It may be that quite a number can be recast in the way that chaperone and transcription factor have been, and it would be interesting to see what is left over. -- Amelia Ireland GO Editorial Office http://www.berkeleybop.org || http://www.ebi.ac.uk Boycott Trader Joe's Red List seafood: http://traitorjoe.com From mgiglio at som.umaryland.edu Wed Oct 7 14:05:59 2009 From: mgiglio at som.umaryland.edu (Gwinn Giglio, Michelle) Date: Wed, 07 Oct 2009 17:05:59 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <10829031-FD46-4357-99DB-C142DA0CD692@ebi.ac.uk> Message-ID: Hi, I think one of the leftovers will be "structural molecule activity" and its children, used for proteins like flagellin. Michelle On 10/7/09 4:58 PM, "Amelia Ireland" wrote: > > On Oct 7, 2009, at 9:12 PM, Jane Lomax wrote: > >> Hi David - got a bit confused for a minute there because I think >> your #1 is equivalent to my #3. So to be clear, I'm talking about >> this one: >> >> gene product MFing biological >> process > > I definitely support this proposal; I think it's the most practical > and most sensible, and I think that having the MFings as part of > biological processes will be intuitive for biologists and those who > are used to (e.g.) pathways being made up of reactions (e.g. in > metabolic pathways) or binding events (e.g. in signal[l]ing). > >> The downside of this option is what we do with the terms in MF that >> aren't occurents - protein tag and their ilk. Are you up for the two >> nodes option? As an experiment I tried splitting MF into MFings and >> true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. >> I can send the file if anyone's interested. > > Could you summarize the 1/8th non-MFings? It may be that quite a > number can be recast in the way that chaperone and transcription > factor have been, and it would be interesting to see what is left over. > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation From dph at informatics.jax.org Wed Oct 7 14:36:15 2009 From: dph at informatics.jax.org (David Hill) Date: Wed, 07 Oct 2009 17:36:15 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <10829031-FD46-4357-99DB-C142DA0CD692@ebi.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <10829031-FD46-4357-99DB-C142DA0CD692@ebi.ac.uk> Message-ID: <4ACD09CF.3030309@informatics.jax.org> > > Could you summarize the 1/8th non-MFings? It may be that quite a > number can be recast in the way that chaperone and transcription > factor have been, and it would be interesting to see what is left over. I'm interested in these too. It just doesn't sit well with me that they don't describe doing something, even if they are named by their roles. D > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From dph at informatics.jax.org Wed Oct 7 14:39:32 2009 From: dph at informatics.jax.org (David Hill) Date: Wed, 07 Oct 2009 17:39:32 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: Message-ID: <4ACD0A94.2000603@informatics.jax.org> The structural molecules are tricky, but they too function. The bricks in my chimney function. I am leaning towards something like 'molecular function involved in flagellum organization' or something along these lines. D Gwinn Giglio, Michelle wrote: > Hi, > > I think one of the leftovers will be "structural molecule activity" and its > children, used for proteins like flagellin. > > Michelle > > > On 10/7/09 4:58 PM, "Amelia Ireland" wrote: > > >> On Oct 7, 2009, at 9:12 PM, Jane Lomax wrote: >> >> >>> Hi David - got a bit confused for a minute there because I think >>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>> this one: >>> >>> gene product MFing biological >>> process >>> >> I definitely support this proposal; I think it's the most practical >> and most sensible, and I think that having the MFings as part of >> biological processes will be intuitive for biologists and those who >> are used to (e.g.) pathways being made up of reactions (e.g. in >> metabolic pathways) or binding events (e.g. in signal[l]ing). >> >> >>> The downside of this option is what we do with the terms in MF that >>> aren't occurents - protein tag and their ilk. Are you up for the two >>> nodes option? As an experiment I tried splitting MF into MFings and >>> true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. >>> I can send the file if anyone's interested. >>> >> Could you summarize the 1/8th non-MFings? It may be that quite a >> number can be recast in the way that chaperone and transcription >> factor have been, and it would be interesting to see what is left over. >> >> -- >> Amelia Ireland >> GO Editorial Office >> http://www.berkeleybop.org || http://www.ebi.ac.uk >> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >> >> >> >> >> >> >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From kchris at genome.stanford.edu Wed Oct 7 14:41:37 2009 From: kchris at genome.stanford.edu (Karen Christie) Date: Wed, 7 Oct 2009 14:41:37 -0700 (PDT) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: I'm with Jane, I got a bit confused too. Though I would like to understand it better, I haven't been to enough GO meetings lately for the "ontology speak" to do much except confuse me. And as David acknowledged, as one of those people who is primarily an annotator anyway, I am really more concerned with the practical aspects of using the ontology to annotate than the nitty gritty of ontological correctness. For me, as a practical matter, I like to think of the Function, Process, and Component ontologies as representing How, What, and Where, respectively. So, I've had issues with F terms that don't address the How at all, but are essentially duplicates of either P or C terms. The F term for "splicing factor activity" that Val brought up to initiate this discussion was one of those because, as we've agreed, it duplicated the process term without providing information about the How. I also have a problem with the "structural constituent of ribosome" term because I think it basically duplicates the component term "ribosome". However, I do think that there is an appropriate function to represent for the ribosome because it has a catalytic activity. But the trickier issue seems to be with things where we understand pretty well what they "do", but what they "do" is passive. We understand actin and flagellin reasonably well and it's clear that they are essentially brick-like building blocks of structural elements: filaments, flagella. We also understand protein tag molecules (ubiquitin, SUMO) reasonably well such that it is clear that they are a protein modification, like acetylation or methylation but bigger (and encoded by a gene product instead of just being a small molecule). But although what they "do" is passive, that knowledge does address the "How" of what these gene products do in cells, so personally I'm in favor of representing these properties as functions in the function ontology. I'm less clear on how one would connect these structural functions to processes; they seem more tied to structures and thus to the component ontology. -Karen P.S. On the specific issue of protein tags as signals, it may be problematic to use the word 'signal' because I don't think that protein tags are only used in signal transduction as we currently represent that process. It may be better to think of something more neutral, perhaps just protein modification. On Wed, 7 Oct 2009, Amelia Ireland wrote: > > On Oct 7, 2009, at 9:24 PM, David Hill wrote: > >> When it executes its role as a signal, it is signaling. > > Perhaps it could be recast as something like 'X binding involved in signal > transduction' or, to make it more explicit, 'signal transmission' or > 'signalling' ('signaling' in the US ;) )? > >> Jane Lomax wrote: >>> Right - but it's not an occurent >>> >>> On Wed, 7 Oct 2009, David Hill wrote: >>> >>>> I think a protein tag has a function as a signal. >>>> >>>> D >>>> >>>> Jane Lomax wrote: >>>>> Hi David - got a bit confused for a minute there because I think your #1 >>>>> is equivalent to my #3. So to be clear, I'm talking about this one: >>>>> >>>>> gene product MFing biological >>>>> process >>>>> >>>>> The downside of this option is what we do with the terms in MF that >>>>> aren't occurents - protein tag and their ilk. Are you up for the two >>>>> nodes option? As an experiment I tried splitting MF into MFings and true >>>>> MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. I can >>>>> send the file if anyone's interested. >>>>> >>>>> Jane >>>>> >>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>> >>>>>> Hi Everyone, >>>>>> >>>>>> I think the issue with the problematic function terms is that we >>>>>> described them as roles because we didn't take enough time to carefully >>>>>> think about what the gene products were actually doing. When I look at >>>>>> them, they are almost always cases where we wanted to assign a function >>>>>> because we knew the product functioned in a process. I think we can >>>>>> just define them as functions involved in processes, sometimes the >>>>>> functions are generic and sometimes they are not. Transcription factor >>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>> activity' involved in 'regulation of transcription'. >>>>>> >>>>>> Here is what a biologist wants to know: >>>>>> >>>>>> Gene product X has the ability to carry out some kind of action >>>>>> (function), either passively (like actin) or actively (like histidine >>>>>> kinase). In the proper biological context, the gene product will carry >>>>>> out that action and that will be part of a biological process. >>>>>> >>>>>> >>>>>> In some cases, there is a requirement for more than one gene product >>>>>> for this elemental action to be carried out. The biological process >>>>>> describes the overall objective of the action. I cannot think of any >>>>>> case where a single gene product acts and has a biological effect, that >>>>>> is succeeds in achieving a biological objective. If that were the case, >>>>>> I think biology wouldn't work, because there is always some background >>>>>> activity as long as a molecule and substrates are present. This is the >>>>>> distinction between a function and a process in the misunderstood >>>>>> 'single step processes'. Even phosphorylation as a biological objective >>>>>> (that has some effect on a biological system) is the result of the >>>>>> action of the activity of many kinase molecules, or the repeated >>>>>> activity of individual kinase molecules. >>>>>> >>>>>> So there are three ways we can represent this: >>>>>> >>>>>> 1) gene product MFing biological >>>>>> process >>>>>> >>>>>> In this case the gene product is a continuant. The >>>>> carry out> relationship describes the relationship between the >>>>>> continuant and the MF which is an occurent. The biological process is >>>>>> an occurent. >>>>>> >>>>>> 2) gene product MF MFing >>>>>> biological process >>>>>> >>>>>> In this case the gene product is a continuant. The MF is a continuant >>>>>> that is an enduring quality of the gene product. The MFing is an >>>>>> occurent that is the realization of the MF. The biological process is >>>>>> an occurent. >>>>>> >>>>>> 3) gene product MF BP >>>>>> >>>>>> In this case the gene product is a continuant. The MF is a continuant >>>>>> that is an enduring quality of the gene product. The BP is the occurent >>>>>> that results from the realization of the MF. >>>>>> >>>>>> I think that the most rigorous way to represent this would be #2. But, >>>>>> I think the most practical way would be #1. When I talk to biologists, >>>>>> both annotators and non-annotators, their eyes glaze over as soon as I >>>>>> mention functions and functionings. The bottom line is they don't care >>>>>> about it, they don't care to think about it, and they are confused by >>>>>> it. They are confused because in the absence of the gene product <>MF >>>>>> relationship, all they care about is a function being executed as part >>>>>> of a biological process. They could care less about the >>>>>> neurotransmitters that never bind a receptor and have a potential to >>>>>> execute a function, but never execute it. >>>>>> >>>>>> However, if we get rid of the MF<>BP side, they care about what the >>>>>> potential of the gene products are. So #1 works for this view. The MF >>>>>> being an occurent is the intuitive way that we think about this. MFs >>>>>> being parts of processes fits with the view I expressed above. >>>>>> >>>>>> #3 is not practical because it would require putting every MF into the >>>>>> BP ontology as an occurent. This would essentially recapitulate the >>>>>> entire MF ontology in BP. It is much easier for us to use #1. The key >>>>>> is in the relationship which combines the >>>>>> ideas of 'has potential' and 'carries out'. >>>>>> >>>>>> My 2 cents. >>>>>> >>>>>> David >>>>>> >>>>>> Judith Blake wrote: >>>>>>> I agree with Suzi, I support the continuation of MF, always have. >>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>> I find utility in describing the MFs, annotating the gp as having the >>>>>>> potential of realizing a MF. >>>>>>> >>>>>>> I see the utility of providing explicit relation between some MF terms >>>>>>> and biological processes. This will be a long journey. >>>>>>> I see continued utility of providing 'function' and 'process' sets for >>>>>>> the many applications and users currently use MF and BP in that way. >>>>>>> >>>>>>> I don't think we will ever fully provide the relations between MF and >>>>>>> BP if only because so much isn't known about how molecular >>>>>>> function(ings) contribute to many biological processes. >>>>>>> >>>>>>> Judy >>>>>>> >>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>> >>>>>>> >>>>>>>> I think the problem here is with trying to characterise MF as either >>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>> >>>>>>> >>>>>>> >>>>>>>> when in fact MF contains a combination of both. >>>>>>>> >>>>>>>> We've been back and forth over this many times and this seems to be >>>>>>>> the crux of it. >>>>>>>> >>>>>>>> Some MF terms fit happily as *parts* of processes (functionings or >>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>>>>>>> tag' does not unfold over time whichever way you look at it. I think >>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>> saying everything in MF is a BFO process isn't really satisfactory >>>>>>>> in the long term. >>>>>>>> >>>>>>>> As I see it we have the following options to fix this: >>>>>>>> >>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>> participates_in?) relations between MF and BP. We'd need to think >>>>>>>> carefully about how annotations would be propagated over realised_in. >>>>>>>> >>>>>>> >>>>>>> This is what was proposed in St. Croix and it makes sense to me, but >>>>>>> it does lead us into some unnecessarily confusing areas (i.e. having >>>>>>> yet another relationship to be added). That is why I am now convinced >>>>>>> that #3 the best approach. >>>>>>> >>>>>>> >>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>> >>>>>>> >>>>>>> Strongly against this one. >>>>>>> >>>>>>> >>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both under >>>>>>>> MF or move the BFO processes into BP and keep MF for functions. The >>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>> >>>>>>> >>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and keep >>>>>>> both under MF. Least amount of change for users, while still solving >>>>>>> the problem. >>>>>>> >>>>>>> There is also 3 (c) which Amelia proposed which is to divide and put >>>>>>> the structural terms in an entirely separate new 4th branch, but I >>>>>>> think a simple divide at the top of MF is enough. >>>>>>> >>>>>>> In any case can we please, please please, get rid of the string >>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>> >>>>>>> >>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>> >>>>>>> >>>>>>> Nope, we've got to fix this. >>>>>>> >>>>>>> >>>>>>>> I don't really like 2 or 4. >>>>>>>> >>>>>>>> Jane >>>>>>>> >>>>>>>> >>>>>>>> Barry Smith wrote: >>>>>>>> >>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>> >>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>> actions' represented in the molecular function ontology. Ontology- >>>>>>>>>> speak. >>>>>>>>>> >>>>>>>>> Great. >>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>> people who are annotating with both 'function' and 'functioning' >>>>>>>>> e.g. using OBI to describe experiments, where it is important to >>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>> There is also parallel work on physiology, distinguishing e.g. the >>>>>>>>> function of the heart (to pump), from realizations of this function >>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>> functional processes such as thumpings, fibrillatings, and so forth. >>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>> functioning it will not work work well with these activities in the >>>>>>>>> future. >>>>>>>>> Barry >>>>>>>>> >>>>>>>>> >>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>> earlier. Clarifying that the gene product has the potential and >>>>>>>>>> the MF terms describes the specific action is what I was after. >>>>>>>>>> judy >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>> >>>>>>>>>> Alan has it right. Functions are not occurents, Function*ings* are. >>>>>>>>>> >>>>>>>>>> But I think that the solution that David, Karen, and Midori have >>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>>>>>>> things) the "Function*ings* are part of processes" into a single >>>>>>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>>>>>> for Function*ings* itself, it is implicit) that is for now called >>>>>>>>>> "part_of" (if I remember correctly). The name of this relationship >>>>>>>>>> may not be ideal, but we are trying to be clear about its intended >>>>>>>>>> usage. >>>>>>>>>> >>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>> between an instance and a type when anyone makes an annotation) >>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>> >>>>>>>>>> -S >>>>>>>>>> >>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>> >>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>> >>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>> describe things. >>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>> If you look at the papers on BFO you will see that functions are >>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>> functions are not occurrents. If the GO is using terminology at >>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>> foundry meeting that we were all going to use a common upper level >>>>>>>>>> ontology. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Functions are part of processes. >>>>>>>>>> >>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>> processes. >>>>>>>>>> >>>>>>>>>> A variety of structures, encoded in a molecule, have the potential >>>>>>>>>> to realize a function. >>>>>>>>>> >>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>> process => "potential to realize a process". >>>>>>>>>> >>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>> >>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>> >>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>> >>>>>>>>>> What is the reference of "That"? >>>>>>>>>> >>>>>>>>>> A molecule can have encoded the potential of engaging in several >>>>>>>>>> functions. >>>>>>>>>> >>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>> >>>>>>>>>> A molecule can bear multiple functions is the way we would say >>>>>>>>>> this. >>>>>>>>>> >>>>>>>>>> The confusion I think is that of the relationship between the gene >>>>>>>>>> product and a molecular function, and the molecular function and >>>>>>>>>> the biological process. >>>>>>>>>> >>>>>>>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>>>>>>> between a molecular function and a gene product is "inheres". The >>>>>>>>>> relation between molecular function and a biological process is >>>>>>>>>> "realized_by". >>>>>>>>>> >>>>>>>>>> -Alan >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Judy >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>> >>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>> broadly) have that make the processes happen. If one goes into >>>>>>>>>> more detail, there are a variety of participants in the process >>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>>>>>> function in a particular process. >>>>>>>>>> >>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>> >>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>> functions are. >>>>>>>>>> >>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>> bearers. >>>>>>>>>> >>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>> have a function but never actually realize it. >>>>>>>>>> >>>>>>>>>> These distinctions have important consequences as people such as >>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>> more detail by adding *participants*. Within those processes, when >>>>>>>>>> there is a molecule like an enzyme that functions in a particular >>>>>>>>>> way, it needs to be distinguished from the other participants in >>>>>>>>>> some way. The way that it is distinguished is by saying it bears a >>>>>>>>>> function. >>>>>>>>>> >>>>>>>>>> The heuristic that a molecular function is a "single step" process >>>>>>>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>>>>>>> that in such descriptions there is a single molecule that has a >>>>>>>>>> function. The step is the process that surrounds execution of that >>>>>>>>>> function. >>>>>>>>>> >>>>>>>>>> As I said, getting this distinction clear is essential for future >>>>>>>>>> detailing of GO processes such as I and others do. Can we start to >>>>>>>>>> get this right here and now? >>>>>>>>>> >>>>>>>>>> -Alan >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> We have two SF items in which merging function and process terms >>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>> specific, single-step "process" that is essentially equivalent to >>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>> that is defined such that it doesn't describe a single activity, >>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>> >>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>>>>>> 0031202 >>>>>>>>>> >>>>>>>>>> into BP term >>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>> >>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>>>>>>> think that's all we would change. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>> >>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>> >>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>> >>>>>>>>>> into MF term >>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>> >>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>> >>>>>>>>>> In this case, the proposal is to merge the process term into the >>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>> only in the MF ontology,and has a part_of link to a process term. >>>>>>>>>> Specifically, the merged term would retain the name and ID from GO: >>>>>>>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>>>>>>> >>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>> >>>>>>>>>> thanks, >>>>>>>>>> m >>>>>>>>>> _______________________________________________ >>>>>>>>>> Annotation mailing list >>>>>>>>>> Annotation at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Annotation mailing list >>>>>>>>>> Annotation at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Ontology-editors mailing list >>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>> >>>>>>>> -- >>>>>>>> Dr Jane Lomax >>>>>>>> GO Editorial Office >>>>>>>> EMBL-EBI >>>>>>>> Wellcome Trust Genome Campus >>>>>>>> Hinxton >>>>>>>> Cambridgeshire, UK >>>>>>>> CB10 1SD >>>>>>>> >>>>>>>> p: +44 1223 492516 >>>>>>>> f: +44 1223 494468 >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ------------------------------------------------------------------------ >>>>>>> _______________________________________________ >>>>>>> Annotation mailing list >>>>>>> Annotation at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>> >>>>>> >>>>>> >>>>> >>>> >>>> >>> >> >> -- >> David P. Hill, Ph.D. >> Bioinformatics Scientist: Ontology Development >> Gene Ontology Consortium >> The Jackson Laboratory >> www.geneontology.org >> www.informatics.jax.org >> tel:207-288-6430 >> >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > > _______________________________________________ > Ontology-editors mailing list > Ontology-editors at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/ontology-editors From hjd at informatics.jax.org Wed Oct 7 15:11:56 2009 From: hjd at informatics.jax.org (Harold Drabkin) Date: Wed, 07 Oct 2009 18:11:56 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACD0A94.2000603@informatics.jax.org> References: <4ACD0A94.2000603@informatics.jax.org> Message-ID: <4ACD122C.9050307@informatics.jax.org> I may regret this... another way to look at this... Can something being a component be it's function? OR: can "structure" itself be a function? brick is_a component/structural part/ of a wall; That's what it does. It does it only if it's a part of the wall. H David Hill wrote: > The structural molecules are tricky, but they too function. The bricks > in my chimney function. I am leaning towards something like 'molecular > function involved in flagellum organization' or something along these > lines. > > D > > Gwinn Giglio, Michelle wrote: >> Hi, >> >> I think one of the leftovers will be "structural molecule activity" >> and its >> children, used for proteins like flagellin. >> >> Michelle >> >> >> On 10/7/09 4:58 PM, "Amelia Ireland" wrote: >> >> >>> On Oct 7, 2009, at 9:12 PM, Jane Lomax wrote: >>> >>> >>>> Hi David - got a bit confused for a minute there because I think >>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>> this one: >>>> >>>> gene product MFing biological >>>> process >>>> >>> I definitely support this proposal; I think it's the most practical >>> and most sensible, and I think that having the MFings as part of >>> biological processes will be intuitive for biologists and those who >>> are used to (e.g.) pathways being made up of reactions (e.g. in >>> metabolic pathways) or binding events (e.g. in signal[l]ing). >>> >>> >>>> The downside of this option is what we do with the terms in MF that >>>> aren't occurents - protein tag and their ilk. Are you up for the two >>>> nodes option? As an experiment I tried splitting MF into MFings and >>>> true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. >>>> I can send the file if anyone's interested. >>>> >>> Could you summarize the 1/8th non-MFings? It may be that quite a >>> number can be recast in the way that chaperone and transcription >>> factor have been, and it would be interesting to see what is left over. >>> >>> -- >>> Amelia Ireland >>> GO Editorial Office >>> http://www.berkeleybop.org || http://www.ebi.ac.uk >>> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Annotation mailing list >>> Annotation at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/annotation >>> >> >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> > From aji at ebi.ac.uk Thu Oct 8 02:12:15 2009 From: aji at ebi.ac.uk (Amelia Ireland) Date: Thu, 8 Oct 2009 10:12:15 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: On Oct 7, 2009, at 10:41 PM, Karen Christie wrote: [snip] > But the trickier issue seems to be with things where we understand > pretty well what they "do", but what they "do" is passive. We > understand actin and flagellin reasonably well and it's clear that > they are essentially brick-like building blocks of structural > elements: filaments, flagella. We also understand protein tag > molecules (ubiquitin, SUMO) reasonably well such that it is clear > that they are a protein modification, like acetylation or > methylation but bigger (and encoded by a gene product instead of > just being a small molecule). But although what they "do" is > passive, that knowledge does address the "How" of what these gene > products do in cells, so personally I'm in favor of representing > these properties as functions in the function ontology. I suppose that in quite a number of cases, the gene product binds to something (e.g. a tag that binds to a protein to direct it to some fate, or binding to other similar proteins to form a structural feature), so if we were to identify its part in a process, it would be (e.g.) 'protein binding involved in ribosome biogenesis'. The difficulty would be in differentiating these instances of binding from those more transitory interactions, e.g. if a chaperone were used during the assembly. Perhaps some qualifiers to distinguish the type of protein or the nature of the binding? > P.S. On the specific issue of protein tags as signals, it may be > problematic to use the word 'signal' because I don't think that > protein tags are only used in signal transduction as we currently > represent that process. It may be better to think of something more > neutral, perhaps just protein modification. This may be a case like that of chaperones, where the term to describe the role of gene products in a number of different processes (protein folding, transport, etc.). It would be best to create specific terms for all the different meanings of 'protein tag'. -- Amelia Ireland GO Editorial Office http://www.berkeleybop.org || http://www.ebi.ac.uk Boycott Trader Joe's Red List seafood: http://traitorjoe.com From val at sanger.ac.uk Thu Oct 8 04:27:55 2009 From: val at sanger.ac.uk (val at sanger.ac.uk) Date: Thu, 8 Oct 2009 12:27:55 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: two things 1. "Protein tag" is not a binding term, it is covalently attached which is why it is placed at the root node of the function ontology. I know its an awful term name, but the definition does describe its "function" Acting as an indicator or marker to facilitate recognition by other molecules in the cell. Recognition of the tag, which can be covalently attached to the target molecule, may result in modification, sequestration, transport or degradation of the molecule in question. 2. By making this a "signal transduction" we are proposing some rather larger changes within the process ontology as this implies that "protein modification by small protein conjugation or removal", and perhaps other types of protein modification, belongs under signal transduction. Val > > On Oct 7, 2009, at 9:24 PM, David Hill wrote: > >> When it executes its role as a signal, it is signaling. > > Perhaps it could be recast as something like 'X binding involved in > signal transduction' or, to make it more explicit, 'signal > transmission' or 'signalling' ('signaling' in the US ;) )? > >> Jane Lomax wrote: >>> Right - but it's not an occurent >>> >>> On Wed, 7 Oct 2009, David Hill wrote: >>> >>>> I think a protein tag has a function as a signal. >>>> >>>> D >>>> >>>> Jane Lomax wrote: >>>>> Hi David - got a bit confused for a minute there because I think >>>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>>> this one: >>>>> >>>>> gene product MFing >>>>> biological process >>>>> >>>>> The downside of this option is what we do with the terms in MF >>>>> that aren't occurents - protein tag and their ilk. Are you up for >>>>> the two nodes option? As an experiment I tried splitting MF into >>>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs v/ >>>>> s 7/8 MFings. I can send the file if anyone's interested. >>>>> >>>>> Jane >>>>> >>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>> >>>>>> Hi Everyone, >>>>>> >>>>>> I think the issue with the problematic function terms is that we >>>>>> described them as roles because we didn't take enough time to >>>>>> carefully think about what the gene products were actually >>>>>> doing. When I look at them, they are almost always cases where >>>>>> we wanted to assign a function because we knew the product >>>>>> functioned in a process. I think we can just define them as >>>>>> functions involved in processes, sometimes the functions are >>>>>> generic and sometimes they are not. Transcription factor >>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>> activity' involved in 'regulation of transcription'. >>>>>> >>>>>> Here is what a biologist wants to know: >>>>>> >>>>>> Gene product X has the ability to carry out some kind of action >>>>>> (function), either passively (like actin) or actively (like >>>>>> histidine kinase). In the proper biological context, the gene >>>>>> product will carry out that action and that will be part of a >>>>>> biological process. >>>>>> >>>>>> >>>>>> In some cases, there is a requirement for more than one gene >>>>>> product for this elemental action to be carried out. The >>>>>> biological process describes the overall objective of the >>>>>> action. I cannot think of any case where a single gene product >>>>>> acts and has a biological effect, that is succeeds in achieving >>>>>> a biological objective. If that were the case, I think biology >>>>>> wouldn't work, because there is always some background activity >>>>>> as long as a molecule and substrates are present. This is the >>>>>> distinction between a function and a process in the >>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>> biological objective (that has some effect on a biological >>>>>> system) is the result of the action of the activity of many >>>>>> kinase molecules, or the repeated activity of individual kinase >>>>>> molecules. >>>>>> >>>>>> So there are three ways we can represent this: >>>>>> >>>>>> 1) gene product MFing >>>>>> biological process >>>>>> >>>>>> In this case the gene product is a continuant. The >>>>> potential to carry out> relationship describes the relationship >>>>>> between the continuant and the MF which is an occurent. The >>>>>> biological process is an occurent. >>>>>> >>>>>> 2) gene product MF MFing >>>>>> biological process >>>>>> >>>>>> In this case the gene product is a continuant. The MF is a >>>>>> continuant that is an enduring quality of the gene product. The >>>>>> MFing is an occurent that is the realization of the MF. The >>>>>> biological process is an occurent. >>>>>> >>>>>> 3) gene product MF BP >>>>>> >>>>>> In this case the gene product is a continuant. The MF is a >>>>>> continuant that is an enduring quality of the gene product. The >>>>>> BP is the occurent that results from the realization of the MF. >>>>>> >>>>>> I think that the most rigorous way to represent this would be >>>>>> #2. But, I think the most practical way would be #1. When I talk >>>>>> to biologists, both annotators and non-annotators, their eyes >>>>>> glaze over as soon as I mention functions and functionings. The >>>>>> bottom line is they don't care about it, they don't care to >>>>>> think about it, and they are confused by it. They are confused >>>>>> because in the absence of the gene product <>MF relationship, >>>>>> all they care about is a function being executed as part of a >>>>>> biological process. They could care less about the >>>>>> neurotransmitters that never bind a receptor and have a >>>>>> potential to execute a function, but never execute it. >>>>>> >>>>>> However, if we get rid of the MF<>BP side, they care about what >>>>>> the potential of the gene products are. So #1 works for this >>>>>> view. The MF being an occurent is the intuitive way that we >>>>>> think about this. MFs being parts of processes fits with the >>>>>> view I expressed above. >>>>>> >>>>>> #3 is not practical because it would require putting every MF >>>>>> into the BP ontology as an occurent. This would essentially >>>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>>> us to use #1. The key is in the relationship >>>>> carry out> which combines the ideas of 'has potential' and >>>>>> 'carries out'. >>>>>> >>>>>> My 2 cents. >>>>>> >>>>>> David >>>>>> >>>>>> Judith Blake wrote: >>>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>>> have. >>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>> I find utility in describing the MFs, annotating the gp as >>>>>>> having the potential of realizing a MF. >>>>>>> >>>>>>> I see the utility of providing explicit relation between some >>>>>>> MF terms and biological processes. This will be a long journey. >>>>>>> I see continued utility of providing 'function' and 'process' >>>>>>> sets for the many applications and users currently use MF and >>>>>>> BP in that way. >>>>>>> >>>>>>> I don't think we will ever fully provide the relations between >>>>>>> MF and BP if only because so much isn't known about how >>>>>>> molecular function(ings) contribute to many biological processes. >>>>>>> >>>>>>> Judy >>>>>>> >>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>> >>>>>>> >>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>> either >>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>> >>>>>>> >>>>>>> >>>>>>>> when in fact MF contains a combination of both. >>>>>>>> >>>>>>>> We've been back and forth over this many times and this seems >>>>>>>> to be >>>>>>>> the crux of it. >>>>>>>> >>>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>>> (functionings or >>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>> 'Protein >>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>> think >>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>> satisfactory >>>>>>>> in the long term. >>>>>>>> >>>>>>>> As I see it we have the following options to fix this: >>>>>>>> >>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>>> think >>>>>>>> carefully about how annotations would be propagated over >>>>>>>> realised_in. >>>>>>>> >>>>>>> >>>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>>> me, but >>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>> having >>>>>>> yet another relationship to be added). That is why I am now >>>>>>> convinced >>>>>>> that #3 the best approach. >>>>>>> >>>>>>> >>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>> >>>>>>> >>>>>>> Strongly against this one. >>>>>>> >>>>>>> >>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>> under >>>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>>> functions. The >>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>> >>>>>>> >>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>>> and keep >>>>>>> both under MF. Least amount of change for users, while still >>>>>>> solving >>>>>>> the problem. >>>>>>> >>>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>>> and put >>>>>>> the structural terms in an entirely separate new 4th branch, >>>>>>> but I >>>>>>> think a simple divide at the top of MF is enough. >>>>>>> >>>>>>> In any case can we please, please please, get rid of the string >>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>> >>>>>>> >>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>> >>>>>>> >>>>>>> Nope, we've got to fix this. >>>>>>> >>>>>>> >>>>>>>> I don't really like 2 or 4. >>>>>>>> >>>>>>>> Jane >>>>>>>> >>>>>>>> >>>>>>>> Barry Smith wrote: >>>>>>>> >>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>> >>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>> Ontology- >>>>>>>>>> speak. >>>>>>>>>> >>>>>>>>> Great. >>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>> Ontology' >>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>> people who are annotating with both 'function' and >>>>>>>>> 'functioning' >>>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>>> to >>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>>> e.g. the >>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>> function >>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>> forth. >>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>> functioning it will not work work well with these activities >>>>>>>>> in the >>>>>>>>> future. >>>>>>>>> Barry >>>>>>>>> >>>>>>>>> >>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>> and >>>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>>> after. >>>>>>>>>> judy >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>> >>>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>>> Function*ings* are. >>>>>>>>>> >>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>> have >>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>> repeating >>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>> single >>>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>>> artifact >>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>> called >>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>> relationship >>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>> intended >>>>>>>>>> usage. >>>>>>>>>> >>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>>> And I >>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>>> annotation) >>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>> >>>>>>>>>> -S >>>>>>>>>> >>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>> >>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>> >>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>> describe things. >>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>> are >>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>>> at >>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>> level >>>>>>>>>> ontology. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Functions are part of processes. >>>>>>>>>> >>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>> processes. >>>>>>>>>> >>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>> potential >>>>>>>>>> to realize a function. >>>>>>>>>> >>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>> process => "potential to realize a process". >>>>>>>>>> >>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>> >>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>> >>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>> >>>>>>>>>> What is the reference of "That"? >>>>>>>>>> >>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>> several >>>>>>>>>> functions. >>>>>>>>>> >>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>> >>>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>>> say this. >>>>>>>>>> >>>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>>> the gene >>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>> and >>>>>>>>>> the biological process. >>>>>>>>>> >>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>> relation >>>>>>>>>> between a molecular function and a gene product is >>>>>>>>>> "inheres". The >>>>>>>>>> relation between molecular function and a biological process >>>>>>>>>> is >>>>>>>>>> "realized_by". >>>>>>>>>> >>>>>>>>>> -Alan >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Judy >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>> >>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>>> into >>>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>>> process >>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>>> the >>>>>>>>>> function in a particular process. >>>>>>>>>> >>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>> >>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>> functions are. >>>>>>>>>> >>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>> bearers. >>>>>>>>>> >>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>> have a function but never actually realize it. >>>>>>>>>> >>>>>>>>>> These distinctions have important consequences as people >>>>>>>>>> such as >>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>>> processes, when >>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>> particular >>>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>>> participants in >>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>> bears a >>>>>>>>>> function. >>>>>>>>>> >>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>> process >>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>> fact >>>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>>> has a >>>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>>> of that >>>>>>>>>> function. >>>>>>>>>> >>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>> future >>>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>>> start to >>>>>>>>>> get this right here and now? >>>>>>>>>> >>>>>>>>>> -Alan >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>> terms >>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>>> equivalent to >>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>> activity, >>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>> >>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>>> GO: >>>>>>>>>> 0031202 >>>>>>>>>> >>>>>>>>>> into BP term >>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>> >>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>>> graph. I >>>>>>>>>> think that's all we would change. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>> >>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>> >>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>> >>>>>>>>>> into MF term >>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>> >>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>> >>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>> the >>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>> term. >>>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>>> from GO: >>>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>>> 0006413. >>>>>>>>>> >>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>> >>>>>>>>>> thanks, >>>>>>>>>> m >>>>>>>>>> _______________________________________________ >>>>>>>>>> Annotation mailing list >>>>>>>>>> Annotation at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Annotation mailing list >>>>>>>>>> Annotation at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Ontology-editors mailing list >>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>> >>>>>>>> -- >>>>>>>> Dr Jane Lomax >>>>>>>> GO Editorial Office >>>>>>>> EMBL-EBI >>>>>>>> Wellcome Trust Genome Campus >>>>>>>> Hinxton >>>>>>>> Cambridgeshire, UK >>>>>>>> CB10 1SD >>>>>>>> >>>>>>>> p: +44 1223 492516 >>>>>>>> f: +44 1223 494468 >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> ------------------------------------------------------------------------ >>>>>>> _______________________________________________ >>>>>>> Annotation mailing list >>>>>>> Annotation at geneontology.org >>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>> >>>>>> >>>>>> >>>>> >>>> >>>> >>> >> >> -- >> David P. Hill, Ph.D. >> Bioinformatics Scientist: Ontology Development >> Gene Ontology Consortium >> The Jackson Laboratory >> www.geneontology.org >> www.informatics.jax.org >> tel:207-288-6430 >> >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From val at sanger.ac.uk Thu Oct 8 04:33:42 2009 From: val at sanger.ac.uk (val at sanger.ac.uk) Date: Thu, 8 Oct 2009 12:33:42 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: > > P.S. On the specific issue of protein tags as signals, it may be > problematic to use the word 'signal' because I don't think that protein > tags are only used in signal transduction as we currently represent that > process. It may be better to think of something more neutral, perhaps just > protein modification. my second point is essentially the same as Karen has made here > > > > > On Wed, 7 Oct 2009, Amelia Ireland wrote: > >> >> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >> >>> When it executes its role as a signal, it is signaling. >> >> Perhaps it could be recast as something like 'X binding involved in >> signal >> transduction' or, to make it more explicit, 'signal transmission' or >> 'signalling' ('signaling' in the US ;) )? >> >>> Jane Lomax wrote: >>>> Right - but it's not an occurent >>>> >>>> On Wed, 7 Oct 2009, David Hill wrote: >>>> >>>>> I think a protein tag has a function as a signal. >>>>> >>>>> D >>>>> >>>>> Jane Lomax wrote: >>>>>> Hi David - got a bit confused for a minute there because I think >>>>>> your #1 >>>>>> is equivalent to my #3. So to be clear, I'm talking about this one: >>>>>> >>>>>> gene product MFing biological >>>>>> process >>>>>> >>>>>> The downside of this option is what we do with the terms in MF that >>>>>> aren't occurents - protein tag and their ilk. Are you up for the two >>>>>> nodes option? As an experiment I tried splitting MF into MFings and >>>>>> true >>>>>> MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. I >>>>>> can >>>>>> send the file if anyone's interested. >>>>>> >>>>>> Jane >>>>>> >>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>> >>>>>>> Hi Everyone, >>>>>>> >>>>>>> I think the issue with the problematic function terms is that we >>>>>>> described them as roles because we didn't take enough time to >>>>>>> carefully >>>>>>> think about what the gene products were actually doing. When I look >>>>>>> at >>>>>>> them, they are almost always cases where we wanted to assign a >>>>>>> function >>>>>>> because we knew the product functioned in a process. I think we can >>>>>>> just define them as functions involved in processes, sometimes the >>>>>>> functions are generic and sometimes they are not. Transcription >>>>>>> factor >>>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>>> activity' involved in 'regulation of transcription'. >>>>>>> >>>>>>> Here is what a biologist wants to know: >>>>>>> >>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>> (function), either passively (like actin) or actively (like >>>>>>> histidine >>>>>>> kinase). In the proper biological context, the gene product will >>>>>>> carry >>>>>>> out that action and that will be part of a biological process. >>>>>>> >>>>>>> >>>>>>> In some cases, there is a requirement for more than one gene >>>>>>> product >>>>>>> for this elemental action to be carried out. The biological process >>>>>>> describes the overall objective of the action. I cannot think of >>>>>>> any >>>>>>> case where a single gene product acts and has a biological effect, >>>>>>> that >>>>>>> is succeeds in achieving a biological objective. If that were the >>>>>>> case, >>>>>>> I think biology wouldn't work, because there is always some >>>>>>> background >>>>>>> activity as long as a molecule and substrates are present. This is >>>>>>> the >>>>>>> distinction between a function and a process in the misunderstood >>>>>>> 'single step processes'. Even phosphorylation as a biological >>>>>>> objective >>>>>>> (that has some effect on a biological system) is the result of the >>>>>>> action of the activity of many kinase molecules, or the repeated >>>>>>> activity of individual kinase molecules. >>>>>>> >>>>>>> So there are three ways we can represent this: >>>>>>> >>>>>>> 1) gene product MFing >>>>>>> biological >>>>>>> process >>>>>>> >>>>>>> In this case the gene product is a continuant. The >>>>>> to >>>>>>> carry out> relationship describes the relationship between the >>>>>>> continuant and the MF which is an occurent. The biological process >>>>>>> is >>>>>>> an occurent. >>>>>>> >>>>>>> 2) gene product MF MFing >>>>>>> biological process >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>> continuant >>>>>>> that is an enduring quality of the gene product. The MFing is an >>>>>>> occurent that is the realization of the MF. The biological process >>>>>>> is >>>>>>> an occurent. >>>>>>> >>>>>>> 3) gene product MF BP >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>> continuant >>>>>>> that is an enduring quality of the gene product. The BP is the >>>>>>> occurent >>>>>>> that results from the realization of the MF. >>>>>>> >>>>>>> I think that the most rigorous way to represent this would be #2. >>>>>>> But, >>>>>>> I think the most practical way would be #1. When I talk to >>>>>>> biologists, >>>>>>> both annotators and non-annotators, their eyes glaze over as soon >>>>>>> as I >>>>>>> mention functions and functionings. The bottom line is they don't >>>>>>> care >>>>>>> about it, they don't care to think about it, and they are confused >>>>>>> by >>>>>>> it. They are confused because in the absence of the gene product >>>>>>> <>MF >>>>>>> relationship, all they care about is a function being executed as >>>>>>> part >>>>>>> of a biological process. They could care less about the >>>>>>> neurotransmitters that never bind a receptor and have a potential >>>>>>> to >>>>>>> execute a function, but never execute it. >>>>>>> >>>>>>> However, if we get rid of the MF<>BP side, they care about what the >>>>>>> potential of the gene products are. So #1 works for this view. The >>>>>>> MF >>>>>>> being an occurent is the intuitive way that we think about this. >>>>>>> MFs >>>>>>> being parts of processes fits with the view I expressed above. >>>>>>> >>>>>>> #3 is not practical because it would require putting every MF into >>>>>>> the >>>>>>> BP ontology as an occurent. This would essentially recapitulate the >>>>>>> entire MF ontology in BP. It is much easier for us to use #1. The >>>>>>> key >>>>>>> is in the relationship which combines >>>>>>> the >>>>>>> ideas of 'has potential' and 'carries out'. >>>>>>> >>>>>>> My 2 cents. >>>>>>> >>>>>>> David >>>>>>> >>>>>>> Judith Blake wrote: >>>>>>>> I agree with Suzi, I support the continuation of MF, always have. >>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>> I find utility in describing the MFs, annotating the gp as having >>>>>>>> the >>>>>>>> potential of realizing a MF. >>>>>>>> >>>>>>>> I see the utility of providing explicit relation between some MF >>>>>>>> terms >>>>>>>> and biological processes. This will be a long journey. >>>>>>>> I see continued utility of providing 'function' and 'process' sets >>>>>>>> for >>>>>>>> the many applications and users currently use MF and BP in that >>>>>>>> way. >>>>>>>> >>>>>>>> I don't think we will ever fully provide the relations between MF >>>>>>>> and >>>>>>>> BP if only because so much isn't known about how molecular >>>>>>>> function(ings) contribute to many biological processes. >>>>>>>> >>>>>>>> Judy >>>>>>>> >>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>> >>>>>>>> >>>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>>> either >>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>> >>>>>>>>> We've been back and forth over this many times and this seems to >>>>>>>>> be >>>>>>>>> the crux of it. >>>>>>>>> >>>>>>>>> Some MF terms fit happily as *parts* of processes (functionings >>>>>>>>> or >>>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>>> 'Protein >>>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>>> think >>>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>>> satisfactory >>>>>>>>> in the long term. >>>>>>>>> >>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>> >>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>> participates_in?) relations between MF and BP. We'd need to think >>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>> realised_in. >>>>>>>>> >>>>>>>> >>>>>>>> This is what was proposed in St. Croix and it makes sense to me, >>>>>>>> but >>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>>> having >>>>>>>> yet another relationship to be added). That is why I am now >>>>>>>> convinced >>>>>>>> that #3 the best approach. >>>>>>>> >>>>>>>> >>>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>> >>>>>>>> >>>>>>>> Strongly against this one. >>>>>>>> >>>>>>>> >>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>>> under >>>>>>>>> MF or move the BFO processes into BP and keep MF for functions. >>>>>>>>> The >>>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>> >>>>>>>> >>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and >>>>>>>> keep >>>>>>>> both under MF. Least amount of change for users, while still >>>>>>>> solving >>>>>>>> the problem. >>>>>>>> >>>>>>>> There is also 3 (c) which Amelia proposed which is to divide and >>>>>>>> put >>>>>>>> the structural terms in an entirely separate new 4th branch, but I >>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>> >>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>> >>>>>>>> >>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>> >>>>>>>> >>>>>>>> Nope, we've got to fix this. >>>>>>>> >>>>>>>> >>>>>>>>> I don't really like 2 or 4. >>>>>>>>> >>>>>>>>> Jane >>>>>>>>> >>>>>>>>> >>>>>>>>> Barry Smith wrote: >>>>>>>>> >>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>> >>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>> Ontology- >>>>>>>>>>> speak. >>>>>>>>>>> >>>>>>>>>> Great. >>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>>> Ontology' >>>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>>> people who are annotating with both 'function' and 'functioning' >>>>>>>>>> e.g. using OBI to describe experiments, where it is important to >>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>>> There is also parallel work on physiology, distinguishing e.g. >>>>>>>>>> the >>>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>>> function >>>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>> forth. >>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>> functioning it will not work work well with these activities in >>>>>>>>>> the >>>>>>>>>> future. >>>>>>>>>> Barry >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>>> and >>>>>>>>>>> the MF terms describes the specific action is what I was after. >>>>>>>>>>> judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Alan has it right. Functions are not occurents, Function*ings* >>>>>>>>>>> are. >>>>>>>>>>> >>>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>>> have >>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>>> repeating >>>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>>> single >>>>>>>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>>> called >>>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>>> relationship >>>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>>> intended >>>>>>>>>>> usage. >>>>>>>>>>> >>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. And >>>>>>>>>>> I >>>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>> between an instance and a type when anyone makes an annotation) >>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>> >>>>>>>>>>> -S >>>>>>>>>>> >>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>> >>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>> describe things. >>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>>> are >>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>> functions are not occurrents. If the GO is using terminology at >>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>>> level >>>>>>>>>>> ontology. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Functions are part of processes. >>>>>>>>>>> >>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>> processes. >>>>>>>>>>> >>>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>>> potential >>>>>>>>>>> to realize a function. >>>>>>>>>>> >>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>> >>>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>> >>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>> >>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>> >>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>> >>>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>>> several >>>>>>>>>>> functions. >>>>>>>>>>> >>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>> >>>>>>>>>>> A molecule can bear multiple functions is the way we would say >>>>>>>>>>> this. >>>>>>>>>>> >>>>>>>>>>> The confusion I think is that of the relationship between the >>>>>>>>>>> gene >>>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>>> and >>>>>>>>>>> the biological process. >>>>>>>>>>> >>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>>> relation >>>>>>>>>>> between a molecular function and a gene product is "inheres". >>>>>>>>>>> The >>>>>>>>>>> relation between molecular function and a biological process is >>>>>>>>>>> "realized_by". >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>> broadly) have that make the processes happen. If one goes into >>>>>>>>>>> more detail, there are a variety of participants in the process >>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>>>>>>> function in a particular process. >>>>>>>>>>> >>>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>>> >>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>> functions are. >>>>>>>>>>> >>>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>>> bearers. >>>>>>>>>>> >>>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>> >>>>>>>>>>> These distinctions have important consequences as people such >>>>>>>>>>> as >>>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>>> more detail by adding *participants*. Within those processes, >>>>>>>>>>> when >>>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>>> particular >>>>>>>>>>> way, it needs to be distinguished from the other participants >>>>>>>>>>> in >>>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>>> bears a >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>>> process >>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>>> fact >>>>>>>>>>> that in such descriptions there is a single molecule that has a >>>>>>>>>>> function. The step is the process that surrounds execution of >>>>>>>>>>> that >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>>> future >>>>>>>>>>> detailing of GO processes such as I and others do. Can we start >>>>>>>>>>> to >>>>>>>>>>> get this right here and now? >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>>> terms >>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>> specific, single-step "process" that is essentially equivalent >>>>>>>>>>> to >>>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>>> activity, >>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>> >>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>>>>>>> 0031202 >>>>>>>>>>> >>>>>>>>>>> into BP term >>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>> >>>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP graph. >>>>>>>>>>> I >>>>>>>>>>> think that's all we would change. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>> >>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>> >>>>>>>>>>> into MF term >>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>>> the >>>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>>> term. >>>>>>>>>>> Specifically, the merged term would retain the name and ID from >>>>>>>>>>> GO: >>>>>>>>>>> 0004479, and would be part_of translational initiation >>>>>>>>>>> GO:0006413. >>>>>>>>>>> >>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>> >>>>>>>>>>> thanks, >>>>>>>>>>> m >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Ontology-editors mailing list >>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>> >>>>>>>>> -- >>>>>>>>> Dr Jane Lomax >>>>>>>>> GO Editorial Office >>>>>>>>> EMBL-EBI >>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>> Hinxton >>>>>>>>> Cambridgeshire, UK >>>>>>>>> CB10 1SD >>>>>>>>> >>>>>>>>> p: +44 1223 492516 >>>>>>>>> f: +44 1223 494468 >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------------------------------------ >>>>>>>> _______________________________________________ >>>>>>>> Annotation mailing list >>>>>>>> Annotation at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>> >>>>> >>>> >>> >>> -- >>> David P. Hill, Ph.D. >>> Bioinformatics Scientist: Ontology Development >>> Gene Ontology Consortium >>> The Jackson Laboratory >>> www.geneontology.org >>> www.informatics.jax.org >>> tel:207-288-6430 >>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >> >> -- >> Amelia Ireland >> GO Editorial Office >> http://www.berkeleybop.org || http://www.ebi.ac.uk >> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >> >> >> >> >> >> >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From dph at informatics.jax.org Thu Oct 8 05:32:28 2009 From: dph at informatics.jax.org (David Hill) Date: Thu, 08 Oct 2009 08:32:28 -0400 Subject: [Ontology-editors] [Annotation] What is the role of Cav3 In-Reply-To: References: Message-ID: <4ACDDBDC.5010602@informatics.jax.org> or: Asking 'what is the finction of Caveolin 3 (Cav3) in muscle? ' Answer "Caveolin 3 contributes to the structural organization of a muscle cell" David Judith Blake wrote: > >From talk I'm listening to this moment by Larry Hunter at BioCreative meeting... > > Putting together reasoning network of ontologies for explanation...(causal reasoning) > Including NLP information... > Looking for where experimental nodes are not co-incident with knowledge nodes.. > > Example (you may have heard this before from Larry...) > What is the role of Cav3... > Hooking up cav3 with muscle cells...creating cross-product terms, relationships, etc. > > [Using knowledge capture; intermediate level of knowledge between BFO and GO > Have to build by hand, level of knowledge in textbooks captured in automated system > Involves, at the moment, human encoding the knowledge] > > This is the language used... > > Asking > 'what is the role of Caveolin 3 (Cav3) in muscle? ' > > Answer > "Caveolin 3 is a structural component of a muscle cell" > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From dph at informatics.jax.org Thu Oct 8 05:34:57 2009 From: dph at informatics.jax.org (David Hill) Date: Thu, 08 Oct 2009 08:34:57 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: <4ACDDC71.1040105@informatics.jax.org> > > [snip] > >> But the trickier issue seems to be with things where we understand >> pretty well what they "do", but what they "do" is passive. We >> understand actin and flagellin reasonably well and it's clear that >> they are essentially brick-like building blocks of structural >> elements: filaments, flagella. We also understand protein tag >> molecules (ubiquitin, SUMO) reasonably well such that it is clear >> that they are a protein modification, like acetylation or methylation >> but bigger (and encoded by a gene product instead of just being a >> small molecule). But although what they "do" is passive, that >> knowledge does address the "How" of what these gene products do in >> cells, so personally I'm in favor of representing these properties as >> functions in the function ontology. > > I suppose that in quite a number of cases, the gene product binds to > something (e.g. a tag that binds to a protein to direct it to some > fate, or binding to other similar proteins to form a structural > feature), so if we were to identify its part in a process, it would be > (e.g.) 'protein binding involved in ribosome biogenesis'. The > difficulty would be in differentiating these instances of binding from > those more transitory interactions, e.g. if a chaperone were used > during the assembly. Perhaps some qualifiers to distinguish the type > of protein or the nature of the binding? exactly, like 'DNA binding involved in transcription regulation' > >> P.S. On the specific issue of protein tags as signals, it may be >> problematic to use the word 'signal' because I don't think that >> protein tags are only used in signal transduction as we currently >> represent that process. It may be better to think of something more >> neutral, perhaps just protein modification. > > > This may be a case like that of chaperones, where the term to describe > the role of gene products in a number of different processes (protein > folding, transport, etc.). It would be best to create specific terms > for all the different meanings of 'protein tag'. I agree here too. David > > -- > Amelia Ireland > GO Editorial Office > http://www.berkeleybop.org || http://www.ebi.ac.uk > Boycott Trader Joe's Red List seafood: http://traitorjoe.com > > > > > > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From dph at informatics.jax.org Thu Oct 8 05:40:44 2009 From: dph at informatics.jax.org (David Hill) Date: Thu, 08 Oct 2009 08:40:44 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: <4ACDDDCC.9080105@informatics.jax.org> > 2. > By making this a "signal transduction" we are proposing some rather larger > changes within the process ontology as this implies that > "protein modification by small protein conjugation or removal", and > perhaps other types of protein modification, belongs under signal > transduction. > I certainly do not think this is a 'signal transduction'. Signal transduction is a process that does something as the result of some types of signaling. I think its function is to signal, or to convey information. When a protein is ubiquitinated, the ubiquitin conveys information that says 'hey look at me, I need to be degraded'. In many cases the purpose of this action is to regulate another molecular function, but let's not go there yet. David > > Val > > > > >> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >> >> >>> When it executes its role as a signal, it is signaling. >>> >> Perhaps it could be recast as something like 'X binding involved in >> signal transduction' or, to make it more explicit, 'signal >> transmission' or 'signalling' ('signaling' in the US ;) )? >> >> >>> Jane Lomax wrote: >>> >>>> Right - but it's not an occurent >>>> >>>> On Wed, 7 Oct 2009, David Hill wrote: >>>> >>>> >>>>> I think a protein tag has a function as a signal. >>>>> >>>>> D >>>>> >>>>> Jane Lomax wrote: >>>>> >>>>>> Hi David - got a bit confused for a minute there because I think >>>>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>>>> this one: >>>>>> >>>>>> gene product MFing >>>>>> biological process >>>>>> >>>>>> The downside of this option is what we do with the terms in MF >>>>>> that aren't occurents - protein tag and their ilk. Are you up for >>>>>> the two nodes option? As an experiment I tried splitting MF into >>>>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs v/ >>>>>> s 7/8 MFings. I can send the file if anyone's interested. >>>>>> >>>>>> Jane >>>>>> >>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>> >>>>>> >>>>>>> Hi Everyone, >>>>>>> >>>>>>> I think the issue with the problematic function terms is that we >>>>>>> described them as roles because we didn't take enough time to >>>>>>> carefully think about what the gene products were actually >>>>>>> doing. When I look at them, they are almost always cases where >>>>>>> we wanted to assign a function because we knew the product >>>>>>> functioned in a process. I think we can just define them as >>>>>>> functions involved in processes, sometimes the functions are >>>>>>> generic and sometimes they are not. Transcription factor >>>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>>> activity' involved in 'regulation of transcription'. >>>>>>> >>>>>>> Here is what a biologist wants to know: >>>>>>> >>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>> (function), either passively (like actin) or actively (like >>>>>>> histidine kinase). In the proper biological context, the gene >>>>>>> product will carry out that action and that will be part of a >>>>>>> biological process. >>>>>>> >>>>>>> >>>>>>> In some cases, there is a requirement for more than one gene >>>>>>> product for this elemental action to be carried out. The >>>>>>> biological process describes the overall objective of the >>>>>>> action. I cannot think of any case where a single gene product >>>>>>> acts and has a biological effect, that is succeeds in achieving >>>>>>> a biological objective. If that were the case, I think biology >>>>>>> wouldn't work, because there is always some background activity >>>>>>> as long as a molecule and substrates are present. This is the >>>>>>> distinction between a function and a process in the >>>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>>> biological objective (that has some effect on a biological >>>>>>> system) is the result of the action of the activity of many >>>>>>> kinase molecules, or the repeated activity of individual kinase >>>>>>> molecules. >>>>>>> >>>>>>> So there are three ways we can represent this: >>>>>>> >>>>>>> 1) gene product MFing >>>>>>> biological process >>>>>>> >>>>>>> In this case the gene product is a continuant. The >>>>>> potential to carry out> relationship describes the relationship >>>>>>> between the continuant and the MF which is an occurent. The >>>>>>> biological process is an occurent. >>>>>>> >>>>>>> 2) gene product MF MFing >>>>>>> biological process >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>> MFing is an occurent that is the realization of the MF. The >>>>>>> biological process is an occurent. >>>>>>> >>>>>>> 3) gene product MF BP >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>> BP is the occurent that results from the realization of the MF. >>>>>>> >>>>>>> I think that the most rigorous way to represent this would be >>>>>>> #2. But, I think the most practical way would be #1. When I talk >>>>>>> to biologists, both annotators and non-annotators, their eyes >>>>>>> glaze over as soon as I mention functions and functionings. The >>>>>>> bottom line is they don't care about it, they don't care to >>>>>>> think about it, and they are confused by it. They are confused >>>>>>> because in the absence of the gene product <>MF relationship, >>>>>>> all they care about is a function being executed as part of a >>>>>>> biological process. They could care less about the >>>>>>> neurotransmitters that never bind a receptor and have a >>>>>>> potential to execute a function, but never execute it. >>>>>>> >>>>>>> However, if we get rid of the MF<>BP side, they care about what >>>>>>> the potential of the gene products are. So #1 works for this >>>>>>> view. The MF being an occurent is the intuitive way that we >>>>>>> think about this. MFs being parts of processes fits with the >>>>>>> view I expressed above. >>>>>>> >>>>>>> #3 is not practical because it would require putting every MF >>>>>>> into the BP ontology as an occurent. This would essentially >>>>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>>>> us to use #1. The key is in the relationship >>>>>> carry out> which combines the ideas of 'has potential' and >>>>>>> 'carries out'. >>>>>>> >>>>>>> My 2 cents. >>>>>>> >>>>>>> David >>>>>>> >>>>>>> Judith Blake wrote: >>>>>>> >>>>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>>>> have. >>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>> I find utility in describing the MFs, annotating the gp as >>>>>>>> having the potential of realizing a MF. >>>>>>>> >>>>>>>> I see the utility of providing explicit relation between some >>>>>>>> MF terms and biological processes. This will be a long journey. >>>>>>>> I see continued utility of providing 'function' and 'process' >>>>>>>> sets for the many applications and users currently use MF and >>>>>>>> BP in that way. >>>>>>>> >>>>>>>> I don't think we will ever fully provide the relations between >>>>>>>> MF and BP if only because so much isn't known about how >>>>>>>> molecular function(ings) contribute to many biological processes. >>>>>>>> >>>>>>>> Judy >>>>>>>> >>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>>> either >>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>> >>>>>>>>> We've been back and forth over this many times and this seems >>>>>>>>> to be >>>>>>>>> the crux of it. >>>>>>>>> >>>>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>>>> (functionings or >>>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>>> 'Protein >>>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>>> think >>>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>>> satisfactory >>>>>>>>> in the long term. >>>>>>>>> >>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>> >>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>>>> think >>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>> realised_in. >>>>>>>>> >>>>>>>>> >>>>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>>>> me, but >>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>>> having >>>>>>>> yet another relationship to be added). That is why I am now >>>>>>>> convinced >>>>>>>> that #3 the best approach. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>> >>>>>>>>> >>>>>>>> Strongly against this one. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>>> under >>>>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>>>> functions. The >>>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>> >>>>>>>>> >>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>>>> and keep >>>>>>>> both under MF. Least amount of change for users, while still >>>>>>>> solving >>>>>>>> the problem. >>>>>>>> >>>>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>>>> and put >>>>>>>> the structural terms in an entirely separate new 4th branch, >>>>>>>> but I >>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>> >>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>> >>>>>>>>> >>>>>>>> Nope, we've got to fix this. >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> I don't really like 2 or 4. >>>>>>>>> >>>>>>>>> Jane >>>>>>>>> >>>>>>>>> >>>>>>>>> Barry Smith wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>> Ontology- >>>>>>>>>>> speak. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> Great. >>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>>> Ontology' >>>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>>> people who are annotating with both 'function' and >>>>>>>>>> 'functioning' >>>>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>>>> to >>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>>>> e.g. the >>>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>>> function >>>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>> forth. >>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>> functioning it will not work work well with these activities >>>>>>>>>> in the >>>>>>>>>> future. >>>>>>>>>> Barry >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>>> and >>>>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>>>> after. >>>>>>>>>>> judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>>>> Function*ings* are. >>>>>>>>>>> >>>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>>> have >>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>>> repeating >>>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>>> single >>>>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>>>> artifact >>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>>> called >>>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>>> relationship >>>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>>> intended >>>>>>>>>>> usage. >>>>>>>>>>> >>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>>>> And I >>>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>>>> annotation) >>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>> >>>>>>>>>>> -S >>>>>>>>>>> >>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>> >>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>> describe things. >>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>>> are >>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>>>> at >>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>>> level >>>>>>>>>>> ontology. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Functions are part of processes. >>>>>>>>>>> >>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>> processes. >>>>>>>>>>> >>>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>>> potential >>>>>>>>>>> to realize a function. >>>>>>>>>>> >>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>> >>>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>> >>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>> >>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>> >>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>> >>>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>>> several >>>>>>>>>>> functions. >>>>>>>>>>> >>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>> >>>>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>>>> say this. >>>>>>>>>>> >>>>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>>>> the gene >>>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>>> and >>>>>>>>>>> the biological process. >>>>>>>>>>> >>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>>> relation >>>>>>>>>>> between a molecular function and a gene product is >>>>>>>>>>> "inheres". The >>>>>>>>>>> relation between molecular function and a biological process >>>>>>>>>>> is >>>>>>>>>>> "realized_by". >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>>>> into >>>>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>>>> process >>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>>>> the >>>>>>>>>>> function in a particular process. >>>>>>>>>>> >>>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>>> >>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>> functions are. >>>>>>>>>>> >>>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>>> bearers. >>>>>>>>>>> >>>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>> >>>>>>>>>>> These distinctions have important consequences as people >>>>>>>>>>> such as >>>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>>>> processes, when >>>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>>> particular >>>>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>>>> participants in >>>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>>> bears a >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>>> process >>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>>> fact >>>>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>>>> has a >>>>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>>>> of that >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>>> future >>>>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>>>> start to >>>>>>>>>>> get this right here and now? >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>>> terms >>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>>>> equivalent to >>>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>>> activity, >>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>> >>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>>>> GO: >>>>>>>>>>> 0031202 >>>>>>>>>>> >>>>>>>>>>> into BP term >>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>> >>>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>>>> graph. I >>>>>>>>>>> think that's all we would change. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>> >>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>> >>>>>>>>>>> into MF term >>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> >>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>>> the >>>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>>> term. >>>>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>>>> from GO: >>>>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>>>> 0006413. >>>>>>>>>>> >>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>> >>>>>>>>>>> thanks, >>>>>>>>>>> m >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Ontology-editors mailing list >>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>> >>>>>>>>>> >>>>>>>>> -- >>>>>>>>> Dr Jane Lomax >>>>>>>>> GO Editorial Office >>>>>>>>> EMBL-EBI >>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>> Hinxton >>>>>>>>> Cambridgeshire, UK >>>>>>>>> CB10 1SD >>>>>>>>> >>>>>>>>> p: +44 1223 492516 >>>>>>>>> f: +44 1223 494468 >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------------------------------------ >>>>>>>> _______________________________________________ >>>>>>>> Annotation mailing list >>>>>>>> Annotation at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>> >>>>>>>> >>>>>>> >>>>> >>> -- >>> David P. Hill, Ph.D. >>> Bioinformatics Scientist: Ontology Development >>> Gene Ontology Consortium >>> The Jackson Laboratory >>> www.geneontology.org >>> www.informatics.jax.org >>> tel:207-288-6430 >>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>> >> -- >> Amelia Ireland >> GO Editorial Office >> http://www.berkeleybop.org || http://www.ebi.ac.uk >> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >> >> >> >> >> >> >> _______________________________________________ >> Annotation mailing list >> Annotation at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/annotation >> >> > > > > > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From val at sanger.ac.uk Thu Oct 8 05:51:48 2009 From: val at sanger.ac.uk (val at sanger.ac.uk) Date: Thu, 8 Oct 2009 13:51:48 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACDDDCC.9080105@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> <4ACDDDCC.9080105@informatics.jax.org> Message-ID: <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> So where in the ontology would the proposed new term be placed. I can't work out where it would go or what it would be called. X binding involved in signal transmission doesnt work because it isn't binding would the new term go in the process ontology under protein modification by small protein conjugation or removal? Would this term be reclassified as a process with no links to the function ontology? > >> 2. >> By making this a "signal transduction" we are proposing some rather >> larger >> changes within the process ontology as this implies that >> "protein modification by small protein conjugation or removal", and >> perhaps other types of protein modification, belongs under signal >> transduction. >> > I certainly do not think this is a 'signal transduction'. Signal > transduction is a process that does something as the result of some > types of signaling. I think its function is to signal, or to convey > information. When a protein is ubiquitinated, the ubiquitin conveys > information that says 'hey look at me, I need to be degraded'. In many > cases the purpose of this action is to regulate another molecular > function, but let's not go there yet. > > David >> >> Val >> >> >> >> >>> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >>> >>> >>>> When it executes its role as a signal, it is signaling. >>>> >>> Perhaps it could be recast as something like 'X binding involved in >>> signal transduction' or, to make it more explicit, 'signal >>> transmission' or 'signalling' ('signaling' in the US ;) )? >>> >>> >>>> Jane Lomax wrote: >>>> >>>>> Right - but it's not an occurent >>>>> >>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>> >>>>> >>>>>> I think a protein tag has a function as a signal. >>>>>> >>>>>> D >>>>>> >>>>>> Jane Lomax wrote: >>>>>> >>>>>>> Hi David - got a bit confused for a minute there because I think >>>>>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>>>>> this one: >>>>>>> >>>>>>> gene product MFing >>>>>>> biological process >>>>>>> >>>>>>> The downside of this option is what we do with the terms in MF >>>>>>> that aren't occurents - protein tag and their ilk. Are you up for >>>>>>> the two nodes option? As an experiment I tried splitting MF into >>>>>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs v/ >>>>>>> s 7/8 MFings. I can send the file if anyone's interested. >>>>>>> >>>>>>> Jane >>>>>>> >>>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>>> >>>>>>> >>>>>>>> Hi Everyone, >>>>>>>> >>>>>>>> I think the issue with the problematic function terms is that we >>>>>>>> described them as roles because we didn't take enough time to >>>>>>>> carefully think about what the gene products were actually >>>>>>>> doing. When I look at them, they are almost always cases where >>>>>>>> we wanted to assign a function because we knew the product >>>>>>>> functioned in a process. I think we can just define them as >>>>>>>> functions involved in processes, sometimes the functions are >>>>>>>> generic and sometimes they are not. Transcription factor >>>>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>>>> activity' involved in 'regulation of transcription'. >>>>>>>> >>>>>>>> Here is what a biologist wants to know: >>>>>>>> >>>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>>> (function), either passively (like actin) or actively (like >>>>>>>> histidine kinase). In the proper biological context, the gene >>>>>>>> product will carry out that action and that will be part of a >>>>>>>> biological process. >>>>>>>> >>>>>>>> >>>>>>>> In some cases, there is a requirement for more than one gene >>>>>>>> product for this elemental action to be carried out. The >>>>>>>> biological process describes the overall objective of the >>>>>>>> action. I cannot think of any case where a single gene product >>>>>>>> acts and has a biological effect, that is succeeds in achieving >>>>>>>> a biological objective. If that were the case, I think biology >>>>>>>> wouldn't work, because there is always some background activity >>>>>>>> as long as a molecule and substrates are present. This is the >>>>>>>> distinction between a function and a process in the >>>>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>>>> biological objective (that has some effect on a biological >>>>>>>> system) is the result of the action of the activity of many >>>>>>>> kinase molecules, or the repeated activity of individual kinase >>>>>>>> molecules. >>>>>>>> >>>>>>>> So there are three ways we can represent this: >>>>>>>> >>>>>>>> 1) gene product MFing >>>>>>>> biological process >>>>>>>> >>>>>>>> In this case the gene product is a continuant. The >>>>>>> potential to carry out> relationship describes the relationship >>>>>>>> between the continuant and the MF which is an occurent. The >>>>>>>> biological process is an occurent. >>>>>>>> >>>>>>>> 2) gene product MF MFing >>>>>>>> biological process >>>>>>>> >>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>> MFing is an occurent that is the realization of the MF. The >>>>>>>> biological process is an occurent. >>>>>>>> >>>>>>>> 3) gene product MF BP >>>>>>>> >>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>> BP is the occurent that results from the realization of the MF. >>>>>>>> >>>>>>>> I think that the most rigorous way to represent this would be >>>>>>>> #2. But, I think the most practical way would be #1. When I talk >>>>>>>> to biologists, both annotators and non-annotators, their eyes >>>>>>>> glaze over as soon as I mention functions and functionings. The >>>>>>>> bottom line is they don't care about it, they don't care to >>>>>>>> think about it, and they are confused by it. They are confused >>>>>>>> because in the absence of the gene product <>MF relationship, >>>>>>>> all they care about is a function being executed as part of a >>>>>>>> biological process. They could care less about the >>>>>>>> neurotransmitters that never bind a receptor and have a >>>>>>>> potential to execute a function, but never execute it. >>>>>>>> >>>>>>>> However, if we get rid of the MF<>BP side, they care about what >>>>>>>> the potential of the gene products are. So #1 works for this >>>>>>>> view. The MF being an occurent is the intuitive way that we >>>>>>>> think about this. MFs being parts of processes fits with the >>>>>>>> view I expressed above. >>>>>>>> >>>>>>>> #3 is not practical because it would require putting every MF >>>>>>>> into the BP ontology as an occurent. This would essentially >>>>>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>>>>> us to use #1. The key is in the relationship >>>>>>> carry out> which combines the ideas of 'has potential' and >>>>>>>> 'carries out'. >>>>>>>> >>>>>>>> My 2 cents. >>>>>>>> >>>>>>>> David >>>>>>>> >>>>>>>> Judith Blake wrote: >>>>>>>> >>>>>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>>>>> have. >>>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>>> I find utility in describing the MFs, annotating the gp as >>>>>>>>> having the potential of realizing a MF. >>>>>>>>> >>>>>>>>> I see the utility of providing explicit relation between some >>>>>>>>> MF terms and biological processes. This will be a long journey. >>>>>>>>> I see continued utility of providing 'function' and 'process' >>>>>>>>> sets for the many applications and users currently use MF and >>>>>>>>> BP in that way. >>>>>>>>> >>>>>>>>> I don't think we will ever fully provide the relations between >>>>>>>>> MF and BP if only because so much isn't known about how >>>>>>>>> molecular function(ings) contribute to many biological processes. >>>>>>>>> >>>>>>>>> Judy >>>>>>>>> >>>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>>>> either >>>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>>> >>>>>>>>>> We've been back and forth over this many times and this seems >>>>>>>>>> to be >>>>>>>>>> the crux of it. >>>>>>>>>> >>>>>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>>>>> (functionings or >>>>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>>>> 'Protein >>>>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>>>> think >>>>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>>>> satisfactory >>>>>>>>>> in the long term. >>>>>>>>>> >>>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>>> >>>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>>>>> think >>>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>>> realised_in. >>>>>>>>>> >>>>>>>>>> >>>>>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>>>>> me, but >>>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>>>> having >>>>>>>>> yet another relationship to be added). That is why I am now >>>>>>>>> convinced >>>>>>>>> that #3 the best approach. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>>> >>>>>>>>>> >>>>>>>>> Strongly against this one. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>>>> under >>>>>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>>>>> functions. The >>>>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>>> >>>>>>>>>> >>>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>>>>> and keep >>>>>>>>> both under MF. Least amount of change for users, while still >>>>>>>>> solving >>>>>>>>> the problem. >>>>>>>>> >>>>>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>>>>> and put >>>>>>>>> the structural terms in an entirely separate new 4th branch, >>>>>>>>> but I >>>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>>> >>>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>>> >>>>>>>>>> >>>>>>>>> Nope, we've got to fix this. >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> I don't really like 2 or 4. >>>>>>>>>> >>>>>>>>>> Jane >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> Barry Smith wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>>> Ontology- >>>>>>>>>>>> speak. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> Great. >>>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>>>> Ontology' >>>>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>>>> people who are annotating with both 'function' and >>>>>>>>>>> 'functioning' >>>>>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>>>>> to >>>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>>>>> e.g. the >>>>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>>>> function >>>>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>>> forth. >>>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>>> functioning it will not work work well with these activities >>>>>>>>>>> in the >>>>>>>>>>> future. >>>>>>>>>>> Barry >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>>>> and >>>>>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>>>>> after. >>>>>>>>>>>> judy >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>>> >>>>>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>>>>> Function*ings* are. >>>>>>>>>>>> >>>>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>>>> have >>>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>>>> repeating >>>>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>>>> single >>>>>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>>>>> artifact >>>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>>>> called >>>>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>>>> relationship >>>>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>>>> intended >>>>>>>>>>>> usage. >>>>>>>>>>>> >>>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>>>>> And I >>>>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>>>>> annotation) >>>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>>> >>>>>>>>>>>> -S >>>>>>>>>>>> >>>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>>> >>>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>>> >>>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>>> describe things. >>>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>>>> are >>>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>>>>> at >>>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>>>> level >>>>>>>>>>>> ontology. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Functions are part of processes. >>>>>>>>>>>> >>>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>>> processes. >>>>>>>>>>>> >>>>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>>>> potential >>>>>>>>>>>> to realize a function. >>>>>>>>>>>> >>>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>>> >>>>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>>> >>>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>>> >>>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>>> >>>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>>> >>>>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>>>> several >>>>>>>>>>>> functions. >>>>>>>>>>>> >>>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>>> >>>>>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>>>>> say this. >>>>>>>>>>>> >>>>>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>>>>> the gene >>>>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>>>> and >>>>>>>>>>>> the biological process. >>>>>>>>>>>> >>>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>>>> relation >>>>>>>>>>>> between a molecular function and a gene product is >>>>>>>>>>>> "inheres". The >>>>>>>>>>>> relation between molecular function and a biological process >>>>>>>>>>>> is >>>>>>>>>>>> "realized_by". >>>>>>>>>>>> >>>>>>>>>>>> -Alan >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Judy >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>>> >>>>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>>>>> into >>>>>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>>>>> process >>>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>>>>> the >>>>>>>>>>>> function in a particular process. >>>>>>>>>>>> >>>>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>>>> >>>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>>> functions are. >>>>>>>>>>>> >>>>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>>>> bearers. >>>>>>>>>>>> >>>>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>>> >>>>>>>>>>>> These distinctions have important consequences as people >>>>>>>>>>>> such as >>>>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>>>>> processes, when >>>>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>>>> particular >>>>>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>>>>> participants in >>>>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>>>> bears a >>>>>>>>>>>> function. >>>>>>>>>>>> >>>>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>>>> process >>>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>>>> fact >>>>>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>>>>> has a >>>>>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>>>>> of that >>>>>>>>>>>> function. >>>>>>>>>>>> >>>>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>>>> future >>>>>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>>>>> start to >>>>>>>>>>>> get this right here and now? >>>>>>>>>>>> >>>>>>>>>>>> -Alan >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>>> Hi, >>>>>>>>>>>> >>>>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>>>> terms >>>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>>>>> equivalent to >>>>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>>>> activity, >>>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>>> >>>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>>>>> GO: >>>>>>>>>>>> 0031202 >>>>>>>>>>>> >>>>>>>>>>>> into BP term >>>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>>> >>>>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>>>>> graph. I >>>>>>>>>>>> think that's all we would change. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>>> >>>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>>> >>>>>>>>>>>> into MF term >>>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>>> >>>>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>>>> the >>>>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>>>> term. >>>>>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>>>>> from GO: >>>>>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>>>>> 0006413. >>>>>>>>>>>> >>>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>>> >>>>>>>>>>>> thanks, >>>>>>>>>>>> m >>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Ontology-editors mailing list >>>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> -- >>>>>>>>>> Dr Jane Lomax >>>>>>>>>> GO Editorial Office >>>>>>>>>> EMBL-EBI >>>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>>> Hinxton >>>>>>>>>> Cambridgeshire, UK >>>>>>>>>> CB10 1SD >>>>>>>>>> >>>>>>>>>> p: +44 1223 492516 >>>>>>>>>> f: +44 1223 494468 >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> ------------------------------------------------------------------------ >>>>>>>>> _______________________________________________ >>>>>>>>> Annotation mailing list >>>>>>>>> Annotation at geneontology.org >>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>> >>>> -- >>>> David P. Hill, Ph.D. >>>> Bioinformatics Scientist: Ontology Development >>>> Gene Ontology Consortium >>>> The Jackson Laboratory >>>> www.geneontology.org >>>> www.informatics.jax.org >>>> tel:207-288-6430 >>>> >>>> _______________________________________________ >>>> Ontology-editors mailing list >>>> Ontology-editors at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>> >>> -- >>> Amelia Ireland >>> GO Editorial Office >>> http://www.berkeleybop.org || http://www.ebi.ac.uk >>> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Annotation mailing list >>> Annotation at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/annotation >>> >>> >> >> >> >> >> > > -- > David P. Hill, Ph.D. > Bioinformatics Scientist: Ontology Development > Gene Ontology Consortium > The Jackson Laboratory > www.geneontology.org > www.informatics.jax.org > tel:207-288-6430 > > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From dph at informatics.jax.org Thu Oct 8 06:10:03 2009 From: dph at informatics.jax.org (David Hill) Date: Thu, 08 Oct 2009 09:10:03 -0400 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> <4ACDDDCC.9080105@informatics.jax.org> <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> Message-ID: <4ACDE4AB.8090308@informatics.jax.org> val at sanger.ac.uk wrote: > So where in the ontology would the proposed new term be placed. I can't > work out where it would go or what it would be called. > We could ave a term called 'molecular signal activity'. It would be a molecular function that is part_of the process of 'signaling'. I could imagine that gene products that have roles as tags and receptor ligands would be included here. > X binding involved in signal transmission > > doesnt work because it isn't binding > > would the new term go in the process ontology under > protein modification by small protein conjugation or removal? > > Would this term be reclassified as a process with no links to the function > ontology? > > > > >>> 2. >>> By making this a "signal transduction" we are proposing some rather >>> larger >>> changes within the process ontology as this implies that >>> "protein modification by small protein conjugation or removal", and >>> perhaps other types of protein modification, belongs under signal >>> transduction. >>> >>> >> I certainly do not think this is a 'signal transduction'. Signal >> transduction is a process that does something as the result of some >> types of signaling. I think its function is to signal, or to convey >> information. When a protein is ubiquitinated, the ubiquitin conveys >> information that says 'hey look at me, I need to be degraded'. In many >> cases the purpose of this action is to regulate another molecular >> function, but let's not go there yet. >> >> David >> >>> Val >>> >>> >>> >>> >>> >>>> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >>>> >>>> >>>> >>>>> When it executes its role as a signal, it is signaling. >>>>> >>>>> >>>> Perhaps it could be recast as something like 'X binding involved in >>>> signal transduction' or, to make it more explicit, 'signal >>>> transmission' or 'signalling' ('signaling' in the US ;) )? >>>> >>>> >>>> >>>>> Jane Lomax wrote: >>>>> >>>>> >>>>>> Right - but it's not an occurent >>>>>> >>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>> >>>>>> >>>>>> >>>>>>> I think a protein tag has a function as a signal. >>>>>>> >>>>>>> D >>>>>>> >>>>>>> Jane Lomax wrote: >>>>>>> >>>>>>> >>>>>>>> Hi David - got a bit confused for a minute there because I think >>>>>>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>>>>>> this one: >>>>>>>> >>>>>>>> gene product MFing >>>>>>>> biological process >>>>>>>> >>>>>>>> The downside of this option is what we do with the terms in MF >>>>>>>> that aren't occurents - protein tag and their ilk. Are you up for >>>>>>>> the two nodes option? As an experiment I tried splitting MF into >>>>>>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs v/ >>>>>>>> s 7/8 MFings. I can send the file if anyone's interested. >>>>>>>> >>>>>>>> Jane >>>>>>>> >>>>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> Hi Everyone, >>>>>>>>> >>>>>>>>> I think the issue with the problematic function terms is that we >>>>>>>>> described them as roles because we didn't take enough time to >>>>>>>>> carefully think about what the gene products were actually >>>>>>>>> doing. When I look at them, they are almost always cases where >>>>>>>>> we wanted to assign a function because we knew the product >>>>>>>>> functioned in a process. I think we can just define them as >>>>>>>>> functions involved in processes, sometimes the functions are >>>>>>>>> generic and sometimes they are not. Transcription factor >>>>>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>>>>> activity' involved in 'regulation of transcription'. >>>>>>>>> >>>>>>>>> Here is what a biologist wants to know: >>>>>>>>> >>>>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>>>> (function), either passively (like actin) or actively (like >>>>>>>>> histidine kinase). In the proper biological context, the gene >>>>>>>>> product will carry out that action and that will be part of a >>>>>>>>> biological process. >>>>>>>>> >>>>>>>>> >>>>>>>>> In some cases, there is a requirement for more than one gene >>>>>>>>> product for this elemental action to be carried out. The >>>>>>>>> biological process describes the overall objective of the >>>>>>>>> action. I cannot think of any case where a single gene product >>>>>>>>> acts and has a biological effect, that is succeeds in achieving >>>>>>>>> a biological objective. If that were the case, I think biology >>>>>>>>> wouldn't work, because there is always some background activity >>>>>>>>> as long as a molecule and substrates are present. This is the >>>>>>>>> distinction between a function and a process in the >>>>>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>>>>> biological objective (that has some effect on a biological >>>>>>>>> system) is the result of the action of the activity of many >>>>>>>>> kinase molecules, or the repeated activity of individual kinase >>>>>>>>> molecules. >>>>>>>>> >>>>>>>>> So there are three ways we can represent this: >>>>>>>>> >>>>>>>>> 1) gene product MFing >>>>>>>>> biological process >>>>>>>>> >>>>>>>>> In this case the gene product is a continuant. The >>>>>>>> potential to carry out> relationship describes the relationship >>>>>>>>> between the continuant and the MF which is an occurent. The >>>>>>>>> biological process is an occurent. >>>>>>>>> >>>>>>>>> 2) gene product MF MFing >>>>>>>>> biological process >>>>>>>>> >>>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>>> MFing is an occurent that is the realization of the MF. The >>>>>>>>> biological process is an occurent. >>>>>>>>> >>>>>>>>> 3) gene product MF BP >>>>>>>>> >>>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>>> BP is the occurent that results from the realization of the MF. >>>>>>>>> >>>>>>>>> I think that the most rigorous way to represent this would be >>>>>>>>> #2. But, I think the most practical way would be #1. When I talk >>>>>>>>> to biologists, both annotators and non-annotators, their eyes >>>>>>>>> glaze over as soon as I mention functions and functionings. The >>>>>>>>> bottom line is they don't care about it, they don't care to >>>>>>>>> think about it, and they are confused by it. They are confused >>>>>>>>> because in the absence of the gene product <>MF relationship, >>>>>>>>> all they care about is a function being executed as part of a >>>>>>>>> biological process. They could care less about the >>>>>>>>> neurotransmitters that never bind a receptor and have a >>>>>>>>> potential to execute a function, but never execute it. >>>>>>>>> >>>>>>>>> However, if we get rid of the MF<>BP side, they care about what >>>>>>>>> the potential of the gene products are. So #1 works for this >>>>>>>>> view. The MF being an occurent is the intuitive way that we >>>>>>>>> think about this. MFs being parts of processes fits with the >>>>>>>>> view I expressed above. >>>>>>>>> >>>>>>>>> #3 is not practical because it would require putting every MF >>>>>>>>> into the BP ontology as an occurent. This would essentially >>>>>>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>>>>>> us to use #1. The key is in the relationship >>>>>>>> carry out> which combines the ideas of 'has potential' and >>>>>>>>> 'carries out'. >>>>>>>>> >>>>>>>>> My 2 cents. >>>>>>>>> >>>>>>>>> David >>>>>>>>> >>>>>>>>> Judith Blake wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>>>>>> have. >>>>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>>>> I find utility in describing the MFs, annotating the gp as >>>>>>>>>> having the potential of realizing a MF. >>>>>>>>>> >>>>>>>>>> I see the utility of providing explicit relation between some >>>>>>>>>> MF terms and biological processes. This will be a long journey. >>>>>>>>>> I see continued utility of providing 'function' and 'process' >>>>>>>>>> sets for the many applications and users currently use MF and >>>>>>>>>> BP in that way. >>>>>>>>>> >>>>>>>>>> I don't think we will ever fully provide the relations between >>>>>>>>>> MF and BP if only because so much isn't known about how >>>>>>>>>> molecular function(ings) contribute to many biological processes. >>>>>>>>>> >>>>>>>>>> Judy >>>>>>>>>> >>>>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>>>>> either >>>>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>>>> >>>>>>>>>>> We've been back and forth over this many times and this seems >>>>>>>>>>> to be >>>>>>>>>>> the crux of it. >>>>>>>>>>> >>>>>>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>>>>>> (functionings or >>>>>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>>>>> 'Protein >>>>>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>>>>> think >>>>>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>>>>> satisfactory >>>>>>>>>>> in the long term. >>>>>>>>>>> >>>>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>>>> >>>>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>>>>>> think >>>>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>>>> realised_in. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>>>>>> me, but >>>>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>>>>> having >>>>>>>>>> yet another relationship to be added). That is why I am now >>>>>>>>>> convinced >>>>>>>>>> that #3 the best approach. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> Strongly against this one. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>>>>> under >>>>>>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>>>>>> functions. The >>>>>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>>>>>> and keep >>>>>>>>>> both under MF. Least amount of change for users, while still >>>>>>>>>> solving >>>>>>>>>> the problem. >>>>>>>>>> >>>>>>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>>>>>> and put >>>>>>>>>> the structural terms in an entirely separate new 4th branch, >>>>>>>>>> but I >>>>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>>>> >>>>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> Nope, we've got to fix this. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> I don't really like 2 or 4. >>>>>>>>>>> >>>>>>>>>>> Jane >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Barry Smith wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>>>> Ontology- >>>>>>>>>>>>> speak. >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> Great. >>>>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>>>>> Ontology' >>>>>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>>>>> people who are annotating with both 'function' and >>>>>>>>>>>> 'functioning' >>>>>>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>>>>>> to >>>>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>>>>>> e.g. the >>>>>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>>>>> function >>>>>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>>>> forth. >>>>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>>>> functioning it will not work work well with these activities >>>>>>>>>>>> in the >>>>>>>>>>>> future. >>>>>>>>>>>> Barry >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>>>>> and >>>>>>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>>>>>> after. >>>>>>>>>>>>> judy >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>>>>>> Function*ings* are. >>>>>>>>>>>>> >>>>>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>>>>> have >>>>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>>>>> repeating >>>>>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>>>>> single >>>>>>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>>>>>> artifact >>>>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>>>>> called >>>>>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>>>>> relationship >>>>>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>>>>> intended >>>>>>>>>>>>> usage. >>>>>>>>>>>>> >>>>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>>>>>> And I >>>>>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>>>>>> annotation) >>>>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>>>> >>>>>>>>>>>>> -S >>>>>>>>>>>>> >>>>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>>>> >>>>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>>>> describe things. >>>>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>>>>> are >>>>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>>>>>> at >>>>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>>>>> level >>>>>>>>>>>>> ontology. >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> Functions are part of processes. >>>>>>>>>>>>> >>>>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>>>> processes. >>>>>>>>>>>>> >>>>>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>>>>> potential >>>>>>>>>>>>> to realize a function. >>>>>>>>>>>>> >>>>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>>>> >>>>>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>>>> >>>>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>>>> >>>>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>>>> >>>>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>>>> >>>>>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>>>>> several >>>>>>>>>>>>> functions. >>>>>>>>>>>>> >>>>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>>>> >>>>>>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>>>>>> say this. >>>>>>>>>>>>> >>>>>>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>>>>>> the gene >>>>>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>>>>> and >>>>>>>>>>>>> the biological process. >>>>>>>>>>>>> >>>>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>>>>> relation >>>>>>>>>>>>> between a molecular function and a gene product is >>>>>>>>>>>>> "inheres". The >>>>>>>>>>>>> relation between molecular function and a biological process >>>>>>>>>>>>> is >>>>>>>>>>>>> "realized_by". >>>>>>>>>>>>> >>>>>>>>>>>>> -Alan >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> Judy >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>>>> >>>>>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>>>>>> into >>>>>>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>>>>>> process >>>>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>>>>>> the >>>>>>>>>>>>> function in a particular process. >>>>>>>>>>>>> >>>>>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>>>>> >>>>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>>>> functions are. >>>>>>>>>>>>> >>>>>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>>>>> bearers. >>>>>>>>>>>>> >>>>>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>>>> >>>>>>>>>>>>> These distinctions have important consequences as people >>>>>>>>>>>>> such as >>>>>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>>>>>> processes, when >>>>>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>>>>> particular >>>>>>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>>>>>> participants in >>>>>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>>>>> bears a >>>>>>>>>>>>> function. >>>>>>>>>>>>> >>>>>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>>>>> process >>>>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>>>>> fact >>>>>>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>>>>>> has a >>>>>>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>>>>>> of that >>>>>>>>>>>>> function. >>>>>>>>>>>>> >>>>>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>>>>> future >>>>>>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>>>>>> start to >>>>>>>>>>>>> get this right here and now? >>>>>>>>>>>>> >>>>>>>>>>>>> -Alan >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>>>> Hi, >>>>>>>>>>>>> >>>>>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>>>>> terms >>>>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>>>>>> equivalent to >>>>>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>>>>> activity, >>>>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>>>> >>>>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>>>>>> GO: >>>>>>>>>>>>> 0031202 >>>>>>>>>>>>> >>>>>>>>>>>>> into BP term >>>>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>>>> >>>>>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>>>>>> graph. I >>>>>>>>>>>>> think that's all we would change. >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>>>> >>>>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>>>> >>>>>>>>>>>>> into MF term >>>>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>>>> >>>>>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>>>>> the >>>>>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>>>>> term. >>>>>>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>>>>>> from GO: >>>>>>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>>>>>> 0006413. >>>>>>>>>>>>> >>>>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>>>> >>>>>>>>>>>>> thanks, >>>>>>>>>>>>> m >>>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>> Ontology-editors mailing list >>>>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> -- >>>>>>>>>>> Dr Jane Lomax >>>>>>>>>>> GO Editorial Office >>>>>>>>>>> EMBL-EBI >>>>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>>>> Hinxton >>>>>>>>>>> Cambridgeshire, UK >>>>>>>>>>> CB10 1SD >>>>>>>>>>> >>>>>>>>>>> p: +44 1223 492516 >>>>>>>>>>> f: +44 1223 494468 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> ------------------------------------------------------------------------ >>>>>>>>>> _______________________________________________ >>>>>>>>>> Annotation mailing list >>>>>>>>>> Annotation at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> >>>>> -- >>>>> David P. Hill, Ph.D. >>>>> Bioinformatics Scientist: Ontology Development >>>>> Gene Ontology Consortium >>>>> The Jackson Laboratory >>>>> www.geneontology.org >>>>> www.informatics.jax.org >>>>> tel:207-288-6430 >>>>> >>>>> _______________________________________________ >>>>> Ontology-editors mailing list >>>>> Ontology-editors at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>> >>>>> >>>> -- >>>> Amelia Ireland >>>> GO Editorial Office >>>> http://www.berkeleybop.org || http://www.ebi.ac.uk >>>> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >>>> >>>> >>>> >>>> >>>> >>>> >>>> _______________________________________________ >>>> Annotation mailing list >>>> Annotation at geneontology.org >>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>> >>>> >>>> >>> >>> >>> >>> >> -- >> David P. Hill, Ph.D. >> Bioinformatics Scientist: Ontology Development >> Gene Ontology Consortium >> The Jackson Laboratory >> www.geneontology.org >> www.informatics.jax.org >> tel:207-288-6430 >> >> >> > > > > > -- David P. Hill, Ph.D. Bioinformatics Scientist: Ontology Development Gene Ontology Consortium The Jackson Laboratory www.geneontology.org www.informatics.jax.org tel:207-288-6430 From val at sanger.ac.uk Thu Oct 8 06:23:52 2009 From: val at sanger.ac.uk (val at sanger.ac.uk) Date: Thu, 8 Oct 2009 14:23:52 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACDE4AB.8090308@informatics.jax.org> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> <4ACDDDCC.9080105@informatics.jax.org> <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> <4ACDE4AB.8090308@informatics.jax.org> Message-ID: <360f86f4a71ee96ad65e87ab9c76bdd0.squirrel@squirrelmail.internal.sanger.ac.uk> > > > val at sanger.ac.uk wrote: >> So where in the ontology would the proposed new term be placed. I can't >> work out where it would go or what it would be called. >> > We could ave a term called 'molecular signal activity'. It would be a > molecular function that is part_of the process of 'signaling'. I could > imagine that gene products that have roles as tags and receptor ligands > would be included here. That would work. If people agree we could rename protein tag to this now, the definition would need to be broadened but it would not affect existing annotations. The process link couldn't be added as present as we have no term "signaling". Is this part of Jen's overhaul? The only disadvantage I see, is that if you wanted to use the term to retreive all of the molecules which function specfically as "protein labels" in ubiquitination, neddylation, urmylation etc, you could no longer use this term to do this as it would now include receptor ligands.... val >> X binding involved in signal transmission >> >> doesnt work because it isn't binding >> >> would the new term go in the process ontology under >> protein modification by small protein conjugation or removal? >> >> Would this term be reclassified as a process with no links to the >> function >> ontology? >> >> >> >> >>>> 2. >>>> By making this a "signal transduction" we are proposing some rather >>>> larger >>>> changes within the process ontology as this implies that >>>> "protein modification by small protein conjugation or removal", and >>>> perhaps other types of protein modification, belongs under signal >>>> transduction. >>>> >>>> >>> I certainly do not think this is a 'signal transduction'. Signal >>> transduction is a process that does something as the result of some >>> types of signaling. I think its function is to signal, or to convey >>> information. When a protein is ubiquitinated, the ubiquitin conveys >>> information that says 'hey look at me, I need to be degraded'. In many >>> cases the purpose of this action is to regulate another molecular >>> function, but let's not go there yet. >>> >>> David >>> >>>> Val >>>> >>>> >>>> >>>> >>>> >>>>> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >>>>> >>>>> >>>>> >>>>>> When it executes its role as a signal, it is signaling. >>>>>> >>>>>> >>>>> Perhaps it could be recast as something like 'X binding involved in >>>>> signal transduction' or, to make it more explicit, 'signal >>>>> transmission' or 'signalling' ('signaling' in the US ;) )? >>>>> >>>>> >>>>> >>>>>> Jane Lomax wrote: >>>>>> >>>>>> >>>>>>> Right - but it's not an occurent >>>>>>> >>>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>>> I think a protein tag has a function as a signal. >>>>>>>> >>>>>>>> D >>>>>>>> >>>>>>>> Jane Lomax wrote: >>>>>>>> >>>>>>>> >>>>>>>>> Hi David - got a bit confused for a minute there because I think >>>>>>>>> your #1 is equivalent to my #3. So to be clear, I'm talking about >>>>>>>>> this one: >>>>>>>>> >>>>>>>>> gene product MFing >>>>>>>>> biological process >>>>>>>>> >>>>>>>>> The downside of this option is what we do with the terms in MF >>>>>>>>> that aren't occurents - protein tag and their ilk. Are you up for >>>>>>>>> the two nodes option? As an experiment I tried splitting MF into >>>>>>>>> MFings and true MFs. It comes out roughly as about 1/8 true MFs >>>>>>>>> v/ >>>>>>>>> s 7/8 MFings. I can send the file if anyone's interested. >>>>>>>>> >>>>>>>>> Jane >>>>>>>>> >>>>>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>>> Hi Everyone, >>>>>>>>>> >>>>>>>>>> I think the issue with the problematic function terms is that we >>>>>>>>>> described them as roles because we didn't take enough time to >>>>>>>>>> carefully think about what the gene products were actually >>>>>>>>>> doing. When I look at them, they are almost always cases where >>>>>>>>>> we wanted to assign a function because we knew the product >>>>>>>>>> functioned in a process. I think we can just define them as >>>>>>>>>> functions involved in processes, sometimes the functions are >>>>>>>>>> generic and sometimes they are not. Transcription factor >>>>>>>>>> activity is and example of one that is not. It is_a 'DNA binding >>>>>>>>>> activity' involved in 'regulation of transcription'. >>>>>>>>>> >>>>>>>>>> Here is what a biologist wants to know: >>>>>>>>>> >>>>>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>>>>> (function), either passively (like actin) or actively (like >>>>>>>>>> histidine kinase). In the proper biological context, the gene >>>>>>>>>> product will carry out that action and that will be part of a >>>>>>>>>> biological process. >>>>>>>>>> >>>>>>>>>> >>>>>>>>>> In some cases, there is a requirement for more than one gene >>>>>>>>>> product for this elemental action to be carried out. The >>>>>>>>>> biological process describes the overall objective of the >>>>>>>>>> action. I cannot think of any case where a single gene product >>>>>>>>>> acts and has a biological effect, that is succeeds in achieving >>>>>>>>>> a biological objective. If that were the case, I think biology >>>>>>>>>> wouldn't work, because there is always some background activity >>>>>>>>>> as long as a molecule and substrates are present. This is the >>>>>>>>>> distinction between a function and a process in the >>>>>>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>>>>>> biological objective (that has some effect on a biological >>>>>>>>>> system) is the result of the action of the activity of many >>>>>>>>>> kinase molecules, or the repeated activity of individual kinase >>>>>>>>>> molecules. >>>>>>>>>> >>>>>>>>>> So there are three ways we can represent this: >>>>>>>>>> >>>>>>>>>> 1) gene product MFing >>>>>>>>>> biological process >>>>>>>>>> >>>>>>>>>> In this case the gene product is a continuant. The >>>>>>>>> potential to carry out> relationship describes the relationship >>>>>>>>>> between the continuant and the MF which is an occurent. The >>>>>>>>>> biological process is an occurent. >>>>>>>>>> >>>>>>>>>> 2) gene product MF MFing >>>>>>>>>> biological process >>>>>>>>>> >>>>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>>>> MFing is an occurent that is the realization of the MF. The >>>>>>>>>> biological process is an occurent. >>>>>>>>>> >>>>>>>>>> 3) gene product MF BP >>>>>>>>>> >>>>>>>>>> In this case the gene product is a continuant. The MF is a >>>>>>>>>> continuant that is an enduring quality of the gene product. The >>>>>>>>>> BP is the occurent that results from the realization of the MF. >>>>>>>>>> >>>>>>>>>> I think that the most rigorous way to represent this would be >>>>>>>>>> #2. But, I think the most practical way would be #1. When I talk >>>>>>>>>> to biologists, both annotators and non-annotators, their eyes >>>>>>>>>> glaze over as soon as I mention functions and functionings. The >>>>>>>>>> bottom line is they don't care about it, they don't care to >>>>>>>>>> think about it, and they are confused by it. They are confused >>>>>>>>>> because in the absence of the gene product <>MF relationship, >>>>>>>>>> all they care about is a function being executed as part of a >>>>>>>>>> biological process. They could care less about the >>>>>>>>>> neurotransmitters that never bind a receptor and have a >>>>>>>>>> potential to execute a function, but never execute it. >>>>>>>>>> >>>>>>>>>> However, if we get rid of the MF<>BP side, they care about what >>>>>>>>>> the potential of the gene products are. So #1 works for this >>>>>>>>>> view. The MF being an occurent is the intuitive way that we >>>>>>>>>> think about this. MFs being parts of processes fits with the >>>>>>>>>> view I expressed above. >>>>>>>>>> >>>>>>>>>> #3 is not practical because it would require putting every MF >>>>>>>>>> into the BP ontology as an occurent. This would essentially >>>>>>>>>> recapitulate the entire MF ontology in BP. It is much easier for >>>>>>>>>> us to use #1. The key is in the relationship >>>>>>>>> carry out> which combines the ideas of 'has potential' and >>>>>>>>>> 'carries out'. >>>>>>>>>> >>>>>>>>>> My 2 cents. >>>>>>>>>> >>>>>>>>>> David >>>>>>>>>> >>>>>>>>>> Judith Blake wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> I agree with Suzi, I support the continuation of MF, always >>>>>>>>>>> have. >>>>>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>>>>> I find utility in describing the MFs, annotating the gp as >>>>>>>>>>> having the potential of realizing a MF. >>>>>>>>>>> >>>>>>>>>>> I see the utility of providing explicit relation between some >>>>>>>>>>> MF terms and biological processes. This will be a long >>>>>>>>>>> journey. >>>>>>>>>>> I see continued utility of providing 'function' and 'process' >>>>>>>>>>> sets for the many applications and users currently use MF and >>>>>>>>>>> BP in that way. >>>>>>>>>>> >>>>>>>>>>> I don't think we will ever fully provide the relations between >>>>>>>>>>> MF and BP if only because so much isn't known about how >>>>>>>>>>> molecular function(ings) contribute to many biological >>>>>>>>>>> processes. >>>>>>>>>>> >>>>>>>>>>> Judy >>>>>>>>>>> >>>>>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> I think the problem here is with trying to characterise MF as >>>>>>>>>>>> either >>>>>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>>>>> >>>>>>>>>>>> We've been back and forth over this many times and this seems >>>>>>>>>>>> to be >>>>>>>>>>>> the crux of it. >>>>>>>>>>>> >>>>>>>>>>>> Some MF terms fit happily as *parts* of processes >>>>>>>>>>>> (functionings or >>>>>>>>>>>> little processes as I prefer to call them ;-)) e.g >>>>>>>>>>>> 'allene-oxide >>>>>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. >>>>>>>>>>>> 'Protein >>>>>>>>>>>> tag' does not unfold over time whichever way you look at it. I >>>>>>>>>>>> think >>>>>>>>>>>> the strategy of ignoring the existence of this latter class >>>>>>>>>>>> and >>>>>>>>>>>> saying everything in MF is a BFO process isn't really >>>>>>>>>>>> satisfactory >>>>>>>>>>>> in the long term. >>>>>>>>>>>> >>>>>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>>>>> >>>>>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>>>>> participates_in?) relations between MF and BP. We'd need to >>>>>>>>>>>> think >>>>>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>>>>> realised_in. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> This is what was proposed in St. Croix and it makes sense to >>>>>>>>>>> me, but >>>>>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. >>>>>>>>>>> having >>>>>>>>>>> yet another relationship to be added). That is why I am now >>>>>>>>>>> convinced >>>>>>>>>>> that #3 the best approach. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> 2. Call everything in MF a BFO process and remove any terms >>>>>>>>>>>> that >>>>>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>>>>> annotations this way. We could possibly move them to a >>>>>>>>>>>> different >>>>>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> Strongly against this one. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both >>>>>>>>>>>> under >>>>>>>>>>>> MF or move the BFO processes into BP and keep MF for >>>>>>>>>>>> functions. The >>>>>>>>>>>> disadvantage here is that we wouldn't have the function for >>>>>>>>>>>> e.g. >>>>>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, >>>>>>>>>>> and keep >>>>>>>>>>> both under MF. Least amount of change for users, while still >>>>>>>>>>> solving >>>>>>>>>>> the problem. >>>>>>>>>>> >>>>>>>>>>> There is also 3 (c) which Amelia proposed which is to divide >>>>>>>>>>> and put >>>>>>>>>>> the structural terms in an entirely separate new 4th branch, >>>>>>>>>>> but I >>>>>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>>>>> >>>>>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> Nope, we've got to fix this. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> I don't really like 2 or 4. >>>>>>>>>>>> >>>>>>>>>>>> Jane >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Barry Smith wrote: >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>>>>> Ontology- >>>>>>>>>>>>>> speak. >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> Great. >>>>>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning >>>>>>>>>>>>> Ontology' >>>>>>>>>>>>> This is not just ontological persnicketiness. There are >>>>>>>>>>>>> already >>>>>>>>>>>>> people who are annotating with both 'function' and >>>>>>>>>>>>> 'functioning' >>>>>>>>>>>>> e.g. using OBI to describe experiments, where it is important >>>>>>>>>>>>> to >>>>>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>>>>> potentials) even when the relevant instrument is switched >>>>>>>>>>>>> off. >>>>>>>>>>>>> There is also parallel work on physiology, distinguishing >>>>>>>>>>>>> e.g. the >>>>>>>>>>>>> function of the heart (to pump), from realizations of this >>>>>>>>>>>>> function >>>>>>>>>>>>> in functionings (pumpings), and both of these in turn from >>>>>>>>>>>>> non- >>>>>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>>>>> forth. >>>>>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>>>>> functioning it will not work work well with these activities >>>>>>>>>>>>> in the >>>>>>>>>>>>> future. >>>>>>>>>>>>> Barry >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> Yes what is being implemented is what was originally >>>>>>>>>>>>>> proposed >>>>>>>>>>>>>> earlier. Clarifying that the gene product has the potential >>>>>>>>>>>>>> and >>>>>>>>>>>>>> the MF terms describes the specific action is what I was >>>>>>>>>>>>>> after. >>>>>>>>>>>>>> judy >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>> Alan has it right. Functions are not occurents, >>>>>>>>>>>>>> Function*ings* are. >>>>>>>>>>>>>> >>>>>>>>>>>>>> But I think that the solution that David, Karen, and Midori >>>>>>>>>>>>>> have >>>>>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid >>>>>>>>>>>>>> repeating >>>>>>>>>>>>>> things) the "Function*ings* are part of processes" into a >>>>>>>>>>>>>> single >>>>>>>>>>>>>> artifact, a relationship (i.e. their is no information >>>>>>>>>>>>>> artifact >>>>>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now >>>>>>>>>>>>>> called >>>>>>>>>>>>>> "part_of" (if I remember correctly). The name of this >>>>>>>>>>>>>> relationship >>>>>>>>>>>>>> may not be ideal, but we are trying to be clear about its >>>>>>>>>>>>>> intended >>>>>>>>>>>>>> usage. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. >>>>>>>>>>>>>> And I >>>>>>>>>>>>>> seem to remember that David and I talked about this (and >>>>>>>>>>>>>> many >>>>>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>>>>> between an instance and a type when anyone makes an >>>>>>>>>>>>>> annotation) >>>>>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>>>>> >>>>>>>>>>>>>> -S >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>>>>> >>>>>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>>>>> describe things. >>>>>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>>>>> If you look at the papers on BFO you will see that functions >>>>>>>>>>>>>> are >>>>>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>>>>> functions are not occurrents. If the GO is using terminology >>>>>>>>>>>>>> at >>>>>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>>>>> foundry meeting that we were all going to use a common upper >>>>>>>>>>>>>> level >>>>>>>>>>>>>> ontology. >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> Functions are part of processes. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>>>>> processes. >>>>>>>>>>>>>> >>>>>>>>>>>>>> A variety of structures, encoded in a molecule, have the >>>>>>>>>>>>>> potential >>>>>>>>>>>>>> to realize a function. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>>>>> substitution "potential to realize a function" + function >>>>>>>>>>>>>> is_a >>>>>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>>>>> >>>>>>>>>>>>>> The error is that it is the functions that are the >>>>>>>>>>>>>> potentials. >>>>>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>>>>> >>>>>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>>>>> >>>>>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>>>>> >>>>>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>>>>> >>>>>>>>>>>>>> A molecule can have encoded the potential of engaging in >>>>>>>>>>>>>> several >>>>>>>>>>>>>> functions. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>>>>> >>>>>>>>>>>>>> A molecule can bear multiple functions is the way we would >>>>>>>>>>>>>> say this. >>>>>>>>>>>>>> >>>>>>>>>>>>>> The confusion I think is that of the relationship between >>>>>>>>>>>>>> the gene >>>>>>>>>>>>>> product and a molecular function, and the molecular function >>>>>>>>>>>>>> and >>>>>>>>>>>>>> the biological process. >>>>>>>>>>>>>> >>>>>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The >>>>>>>>>>>>>> relation >>>>>>>>>>>>>> between a molecular function and a gene product is >>>>>>>>>>>>>> "inheres". The >>>>>>>>>>>>>> relation between molecular function and a biological process >>>>>>>>>>>>>> is >>>>>>>>>>>>>> "realized_by". >>>>>>>>>>>>>> >>>>>>>>>>>>>> -Alan >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> Judy >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>>>>> >>>>>>>>>>>>>> The problem is that this still confuses function with >>>>>>>>>>>>>> process. >>>>>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>>>>> broadly) have that make the processes happen. If one goes >>>>>>>>>>>>>> into >>>>>>>>>>>>>> more detail, there are a variety of participants in the >>>>>>>>>>>>>> process >>>>>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of >>>>>>>>>>>>>> the >>>>>>>>>>>>>> function in a particular process. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Every function has a corresponding process, its >>>>>>>>>>>>>> "realization". >>>>>>>>>>>>>> >>>>>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>>>>> functions are. >>>>>>>>>>>>>> >>>>>>>>>>>>>> The functions have a physical basis in the structure of >>>>>>>>>>>>>> their >>>>>>>>>>>>>> bearers. >>>>>>>>>>>>>> >>>>>>>>>>>>>> The functions exist before any process happens. A molecule >>>>>>>>>>>>>> may >>>>>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>>>>> >>>>>>>>>>>>>> These distinctions have important consequences as people >>>>>>>>>>>>>> such as >>>>>>>>>>>>>> myself work on adding more detail to the GO. The processes >>>>>>>>>>>>>> get >>>>>>>>>>>>>> more detail by adding *participants*. Within those >>>>>>>>>>>>>> processes, when >>>>>>>>>>>>>> there is a molecule like an enzyme that functions in a >>>>>>>>>>>>>> particular >>>>>>>>>>>>>> way, it needs to be distinguished from the other >>>>>>>>>>>>>> participants in >>>>>>>>>>>>>> some way. The way that it is distinguished is by saying it >>>>>>>>>>>>>> bears a >>>>>>>>>>>>>> function. >>>>>>>>>>>>>> >>>>>>>>>>>>>> The heuristic that a molecular function is a "single step" >>>>>>>>>>>>>> process >>>>>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the >>>>>>>>>>>>>> fact >>>>>>>>>>>>>> that in such descriptions there is a single molecule that >>>>>>>>>>>>>> has a >>>>>>>>>>>>>> function. The step is the process that surrounds execution >>>>>>>>>>>>>> of that >>>>>>>>>>>>>> function. >>>>>>>>>>>>>> >>>>>>>>>>>>>> As I said, getting this distinction clear is essential for >>>>>>>>>>>>>> future >>>>>>>>>>>>>> detailing of GO processes such as I and others do. Can we >>>>>>>>>>>>>> start to >>>>>>>>>>>>>> get this right here and now? >>>>>>>>>>>>>> >>>>>>>>>>>>>> -Alan >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>>>>> Hi, >>>>>>>>>>>>>> >>>>>>>>>>>>>> We have two SF items in which merging function and process >>>>>>>>>>>>>> terms >>>>>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>>>>> specific, single-step "process" that is essentially >>>>>>>>>>>>>> equivalent to >>>>>>>>>>>>>> a "function" (in the GO sense); the other is a "function" >>>>>>>>>>>>>> term >>>>>>>>>>>>>> that is defined such that it doesn't describe a single >>>>>>>>>>>>>> activity, >>>>>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>>>>> >>>>>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>>>>> RNA splicing factor activity, transesterification mechanism >>>>>>>>>>>>>> GO: >>>>>>>>>>>>>> 0031202 >>>>>>>>>>>>>> >>>>>>>>>>>>>> into BP term >>>>>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>>>>> >>>>>>>>>>>>>> For this one, we would merge the MF term into the BP term, >>>>>>>>>>>>>> and >>>>>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP >>>>>>>>>>>>>> graph. I >>>>>>>>>>>>>> think that's all we would change. >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>>>>> >>>>>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>>>>> >>>>>>>>>>>>>> into MF term >>>>>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>>> >>>>>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>>>>> >>>>>>>>>>>>>> In this case, the proposal is to merge the process term into >>>>>>>>>>>>>> the >>>>>>>>>>>>>> function term, ensuring that the merged term has is_a >>>>>>>>>>>>>> ancestry >>>>>>>>>>>>>> only in the MF ontology,and has a part_of link to a process >>>>>>>>>>>>>> term. >>>>>>>>>>>>>> Specifically, the merged term would retain the name and ID >>>>>>>>>>>>>> from GO: >>>>>>>>>>>>>> 0004479, and would be part_of translational initiation GO: >>>>>>>>>>>>>> 0006413. >>>>>>>>>>>>>> >>>>>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>>>>> >>>>>>>>>>>>>> thanks, >>>>>>>>>>>>>> m >>>>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>>>> Annotation mailing list >>>>>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>>> >>>>>>>>>>>>> _______________________________________________ >>>>>>>>>>>>> Ontology-editors mailing list >>>>>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>>> >>>>>>>>>>>> -- >>>>>>>>>>>> Dr Jane Lomax >>>>>>>>>>>> GO Editorial Office >>>>>>>>>>>> EMBL-EBI >>>>>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>>>>> Hinxton >>>>>>>>>>>> Cambridgeshire, UK >>>>>>>>>>>> CB10 1SD >>>>>>>>>>>> >>>>>>>>>>>> p: +44 1223 492516 >>>>>>>>>>>> f: +44 1223 494468 >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> ------------------------------------------------------------------------ >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>> -- >>>>>> David P. Hill, Ph.D. >>>>>> Bioinformatics Scientist: Ontology Development >>>>>> Gene Ontology Consortium >>>>>> The Jackson Laboratory >>>>>> www.geneontology.org >>>>>> www.informatics.jax.org >>>>>> tel:207-288-6430 >>>>>> >>>>>> _______________________________________________ >>>>>> Ontology-editors mailing list >>>>>> Ontology-editors at geneontology.org >>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>> >>>>>> >>>>> -- >>>>> Amelia Ireland >>>>> GO Editorial Office >>>>> http://www.berkeleybop.org || http://www.ebi.ac.uk >>>>> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> Annotation mailing list >>>>> Annotation at geneontology.org >>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>> >>>>> >>>>> >>>> >>>> >>>> >>>> >>> -- >>> David P. Hill, Ph.D. >>> Bioinformatics Scientist: Ontology Development >>> Gene Ontology Consortium >>> The Jackson Laboratory >>> www.geneontology.org >>> www.informatics.jax.org >>> tel:207-288-6430 >>> >>> >>> >> >> >> >> >> > > -- > David P. Hill, Ph.D. > Bioinformatics Scientist: Ontology Development > Gene Ontology Consortium > The Jackson Laboratory > www.geneontology.org > www.informatics.jax.org > tel:207-288-6430 > > -- The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE. From jane at ebi.ac.uk Thu Oct 8 06:27:27 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Thu, 8 Oct 2009 14:27:27 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> Message-ID: Karen - it's probably worth having you having a look at the BFO, especially the definitions of process and function: http://bioportal.bioontology.org/visualize/40358 Jane On Wed, 7 Oct 2009, Karen Christie wrote: > I'm with Jane, I got a bit confused too. Though I would like to understand it > better, I haven't been to enough GO meetings lately for the "ontology speak" > to do much except confuse me. And as David acknowledged, as one of those > people who is primarily an annotator anyway, I am really more concerned with > the practical aspects of using the ontology to annotate than the nitty gritty > of ontological correctness. > > For me, as a practical matter, I like to think of the Function, Process, and > Component ontologies as representing How, What, and Where, respectively. So, > I've had issues with F terms that don't address the How at all, but are > essentially duplicates of either P or C terms. The F term for "splicing > factor activity" that Val brought up to initiate this discussion was one of > those because, as we've agreed, it duplicated the process term without > providing information about the How. I also have a problem with the > "structural constituent of ribosome" term because I think it basically > duplicates the component term "ribosome". However, I do think that there is > an appropriate function to represent for the ribosome because it has a > catalytic activity. > > But the trickier issue seems to be with things where we understand pretty > well what they "do", but what they "do" is passive. We understand actin and > flagellin reasonably well and it's clear that they are essentially brick-like > building blocks of structural elements: filaments, flagella. We also > understand protein tag molecules (ubiquitin, SUMO) reasonably well such that > it is clear that they are a protein modification, like acetylation or > methylation but bigger (and encoded by a gene product instead of just being a > small molecule). But although what they "do" is passive, that knowledge does > address the "How" of what these gene products do in cells, so personally I'm > in favor of representing these properties as functions in the function > ontology. > > I'm less clear on how one would connect these structural functions to > processes; they seem more tied to structures and thus to the component > ontology. > > -Karen > > P.S. On the specific issue of protein tags as signals, it may be problematic > to use the word 'signal' because I don't think that protein tags are only > used in signal transduction as we currently represent that process. It may be > better to think of something more neutral, perhaps just protein modification. > > > > > On Wed, 7 Oct 2009, Amelia Ireland wrote: > >> >> On Oct 7, 2009, at 9:24 PM, David Hill wrote: >> >>> When it executes its role as a signal, it is signaling. >> >> Perhaps it could be recast as something like 'X binding involved in signal >> transduction' or, to make it more explicit, 'signal transmission' or >> 'signalling' ('signaling' in the US ;) )? >> >>> Jane Lomax wrote: >>>> Right - but it's not an occurent >>>> >>>> On Wed, 7 Oct 2009, David Hill wrote: >>>> >>>>> I think a protein tag has a function as a signal. >>>>> >>>>> D >>>>> >>>>> Jane Lomax wrote: >>>>>> Hi David - got a bit confused for a minute there because I think your >>>>>> #1 is equivalent to my #3. So to be clear, I'm talking about this one: >>>>>> >>>>>> gene product MFing biological >>>>>> process >>>>>> >>>>>> The downside of this option is what we do with the terms in MF that >>>>>> aren't occurents - protein tag and their ilk. Are you up for the two >>>>>> nodes option? As an experiment I tried splitting MF into MFings and >>>>>> true MFs. It comes out roughly as about 1/8 true MFs v/s 7/8 MFings. I >>>>>> can send the file if anyone's interested. >>>>>> >>>>>> Jane >>>>>> >>>>>> On Wed, 7 Oct 2009, David Hill wrote: >>>>>> >>>>>>> Hi Everyone, >>>>>>> >>>>>>> I think the issue with the problematic function terms is that we >>>>>>> described them as roles because we didn't take enough time to >>>>>>> carefully think about what the gene products were actually doing. When >>>>>>> I look at them, they are almost always cases where we wanted to assign >>>>>>> a function because we knew the product functioned in a process. I >>>>>>> think we can just define them as functions involved in processes, >>>>>>> sometimes the functions are generic and sometimes they are not. >>>>>>> Transcription factor activity is and example of one that is not. It >>>>>>> is_a 'DNA binding activity' involved in 'regulation of transcription'. >>>>>>> >>>>>>> Here is what a biologist wants to know: >>>>>>> >>>>>>> Gene product X has the ability to carry out some kind of action >>>>>>> (function), either passively (like actin) or actively (like histidine >>>>>>> kinase). In the proper biological context, the gene product will carry >>>>>>> out that action and that will be part of a biological process. >>>>>>> >>>>>>> >>>>>>> In some cases, there is a requirement for more than one gene product >>>>>>> for this elemental action to be carried out. The biological process >>>>>>> describes the overall objective of the action. I cannot think of any >>>>>>> case where a single gene product acts and has a biological effect, >>>>>>> that is succeeds in achieving a biological objective. If that were the >>>>>>> case, I think biology wouldn't work, because there is always some >>>>>>> background activity as long as a molecule and substrates are present. >>>>>>> This is the distinction between a function and a process in the >>>>>>> misunderstood 'single step processes'. Even phosphorylation as a >>>>>>> biological objective (that has some effect on a biological system) is >>>>>>> the result of the action of the activity of many kinase molecules, or >>>>>>> the repeated activity of individual kinase molecules. >>>>>>> >>>>>>> So there are three ways we can represent this: >>>>>>> >>>>>>> 1) gene product MFing >>>>>>> biological process >>>>>>> >>>>>>> In this case the gene product is a continuant. The >>>>>> carry out> relationship describes the relationship between the >>>>>>> continuant and the MF which is an occurent. The biological process is >>>>>>> an occurent. >>>>>>> >>>>>>> 2) gene product MF MFing >>>>>>> biological process >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a continuant >>>>>>> that is an enduring quality of the gene product. The MFing is an >>>>>>> occurent that is the realization of the MF. The biological process is >>>>>>> an occurent. >>>>>>> >>>>>>> 3) gene product MF BP >>>>>>> >>>>>>> In this case the gene product is a continuant. The MF is a continuant >>>>>>> that is an enduring quality of the gene product. The BP is the >>>>>>> occurent that results from the realization of the MF. >>>>>>> >>>>>>> I think that the most rigorous way to represent this would be #2. But, >>>>>>> I think the most practical way would be #1. When I talk to biologists, >>>>>>> both annotators and non-annotators, their eyes glaze over as soon as I >>>>>>> mention functions and functionings. The bottom line is they don't care >>>>>>> about it, they don't care to think about it, and they are confused by >>>>>>> it. They are confused because in the absence of the gene product <>MF >>>>>>> relationship, all they care about is a function being executed as part >>>>>>> of a biological process. They could care less about the >>>>>>> neurotransmitters that never bind a receptor and have a potential to >>>>>>> execute a function, but never execute it. >>>>>>> >>>>>>> However, if we get rid of the MF<>BP side, they care about what the >>>>>>> potential of the gene products are. So #1 works for this view. The MF >>>>>>> being an occurent is the intuitive way that we think about this. MFs >>>>>>> being parts of processes fits with the view I expressed above. >>>>>>> >>>>>>> #3 is not practical because it would require putting every MF into the >>>>>>> BP ontology as an occurent. This would essentially recapitulate the >>>>>>> entire MF ontology in BP. It is much easier for us to use #1. The key >>>>>>> is in the relationship which combines the >>>>>>> ideas of 'has potential' and 'carries out'. >>>>>>> >>>>>>> My 2 cents. >>>>>>> >>>>>>> David >>>>>>> >>>>>>> Judith Blake wrote: >>>>>>>> I agree with Suzi, I support the continuation of MF, always have. >>>>>>>> I don't have an issue with retaining the word 'activity'. >>>>>>>> I find utility in describing the MFs, annotating the gp as having the >>>>>>>> potential of realizing a MF. >>>>>>>> >>>>>>>> I see the utility of providing explicit relation between some MF >>>>>>>> terms and biological processes. This will be a long journey. >>>>>>>> I see continued utility of providing 'function' and 'process' sets >>>>>>>> for the many applications and users currently use MF and BP in that >>>>>>>> way. >>>>>>>> >>>>>>>> I don't think we will ever fully provide the relations between MF and >>>>>>>> BP if only because so much isn't known about how molecular >>>>>>>> function(ings) contribute to many biological processes. >>>>>>>> >>>>>>>> Judy >>>>>>>> >>>>>>>> On 10/6/09 8:45 PM, "Suzanna Lewis" wrote: >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> On Oct 6, 2009, at 6:24 AM, Jane Lomax wrote: >>>>>>>> >>>>>>>> >>>>>>>>> I think the problem here is with trying to characterise MF as either >>>>>>>>> *all* functions or *all* processes (i.e functionings) >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> when in fact MF contains a combination of both. >>>>>>>>> >>>>>>>>> We've been back and forth over this many times and this seems to be >>>>>>>>> the crux of it. >>>>>>>>> >>>>>>>>> Some MF terms fit happily as *parts* of processes (functionings or >>>>>>>>> little processes as I prefer to call them ;-)) e.g 'allene-oxide >>>>>>>>> cyclase activity'. Some clearly do not, e.g. 'protein tag'. 'Protein >>>>>>>>> tag' does not unfold over time whichever way you look at it. I think >>>>>>>>> the strategy of ignoring the existence of this latter class and >>>>>>>>> saying everything in MF is a BFO process isn't really satisfactory >>>>>>>>> in the long term. >>>>>>>>> >>>>>>>>> As I see it we have the following options to fix this: >>>>>>>>> >>>>>>>>> 1. Take away the word 'activity' from the term names and call >>>>>>>>> everything in MF a BFO function. Make realised_in (or >>>>>>>>> participates_in?) relations between MF and BP. We'd need to think >>>>>>>>> carefully about how annotations would be propagated over >>>>>>>>> realised_in. >>>>>>>>> >>>>>>>> >>>>>>>> This is what was proposed in St. Croix and it makes sense to me, but >>>>>>>> it does lead us into some unnecessarily confusing areas (i.e. having >>>>>>>> yet another relationship to be added). That is why I am now convinced >>>>>>>> that #3 the best approach. >>>>>>>> >>>>>>>> >>>>>>>>> 2. Call everything in MF a BFO process and remove any terms that >>>>>>>>> don't fit as processes e.g. protein tag. We lose a lot of >>>>>>>>> annotations this way. We could possibly move them to a different >>>>>>>>> ontology but would mean lots of re-annotation. >>>>>>>>> >>>>>>>> >>>>>>>> Strongly against this one. >>>>>>>> >>>>>>>> >>>>>>>>> 3. Divide MF into BFO processes and BFO functions, either both under >>>>>>>>> MF or move the BFO processes into BP and keep MF for functions. The >>>>>>>>> disadvantage here is that we wouldn't have the function for e.g. >>>>>>>>> allene-oxide cyclase - we'd only have the process. You could I >>>>>>>>> suppose make 'allene-oxide cyclase function' and 'allene-oxide >>>>>>>>> cyclase activity' but that gets pretty confusing. >>>>>>>>> >>>>>>>> >>>>>>>> I'm for 3 (a) Divide MF into BFO processes and BFO functions, and >>>>>>>> keep >>>>>>>> both under MF. Least amount of change for users, while still solving >>>>>>>> the problem. >>>>>>>> >>>>>>>> There is also 3 (c) which Amelia proposed which is to divide and put >>>>>>>> the structural terms in an entirely separate new 4th branch, but I >>>>>>>> think a simple divide at the top of MF is enough. >>>>>>>> >>>>>>>> In any case can we please, please please, get rid of the string >>>>>>>> "activity". It adds nothing and is quite gratuitous. >>>>>>>> >>>>>>>> >>>>>>>>> 4. Ignore the problem and hope it goes away (it won't). >>>>>>>>> >>>>>>>> >>>>>>>> Nope, we've got to fix this. >>>>>>>> >>>>>>>> >>>>>>>>> I don't really like 2 or 4. >>>>>>>>> >>>>>>>>> Jane >>>>>>>>> >>>>>>>>> >>>>>>>>> Barry Smith wrote: >>>>>>>>> >>>>>>>>>> At 08:33 AM 10/1/2009, Judith Blake wrote: >>>>>>>>>> >>>>>>>>>>> Ok Ok I'll use the word 'functioning's' for the 'occurrants >>>>>>>>>>> actions' represented in the molecular function ontology. >>>>>>>>>>> Ontology- >>>>>>>>>>> speak. >>>>>>>>>>> >>>>>>>>>> Great. >>>>>>>>>> Ideally, MF should be renamed the 'Molecular Functioning Ontology' >>>>>>>>>> This is not just ontological persnicketiness. There are already >>>>>>>>>> people who are annotating with both 'function' and 'functioning' >>>>>>>>>> e.g. using OBI to describe experiments, where it is important to >>>>>>>>>> record both the functions e.g. of lab instruments, and the >>>>>>>>>> processes in which they participate. The former exist (as >>>>>>>>>> potentials) even when the relevant instrument is switched off. >>>>>>>>>> There is also parallel work on physiology, distinguishing e.g. the >>>>>>>>>> function of the heart (to pump), from realizations of this function >>>>>>>>>> in functionings (pumpings), and both of these in turn from non- >>>>>>>>>> functional processes such as thumpings, fibrillatings, and so >>>>>>>>>> forth. >>>>>>>>>> If the GO were to entrench, now, a confusion of function and >>>>>>>>>> functioning it will not work work well with these activities in the >>>>>>>>>> future. >>>>>>>>>> Barry >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Yes what is being implemented is what was originally proposed >>>>>>>>>>> earlier. Clarifying that the gene product has the potential and >>>>>>>>>>> the MF terms describes the specific action is what I was after. >>>>>>>>>>> judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/30/09 8:30 PM, "Suzanna Lewis" >>>>>>>>>>> <suzi at fruitfly.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Alan has it right. Functions are not occurents, Function*ings* >>>>>>>>>>> are. >>>>>>>>>>> >>>>>>>>>>> But I think that the solution that David, Karen, and Midori have >>>>>>>>>>> proposed is fine. We've consciously collapsed (to avoid repeating >>>>>>>>>>> things) the "Function*ings* are part of processes" into a single >>>>>>>>>>> artifact, a relationship (i.e. their is no information artifact >>>>>>>>>>> for Function*ings* itself, it is implicit) that is for now called >>>>>>>>>>> "part_of" (if I remember correctly). The name of this relationship >>>>>>>>>>> may not be ideal, but we are trying to be clear about its intended >>>>>>>>>>> usage. >>>>>>>>>>> >>>>>>>>>>> Chris & David first proposed this at the St. Croix meeting. And I >>>>>>>>>>> seem to remember that David and I talked about this (and many >>>>>>>>>>> other things, such as the nature of the implied relationship >>>>>>>>>>> between an instance and a type when anyone makes an annotation) >>>>>>>>>>> again at the relation ontology meeting in Denver. >>>>>>>>>>> >>>>>>>>>>> -S >>>>>>>>>>> >>>>>>>>>>> On Sep 29, 2009, at 11:45 AM, Alan Ruttenberg wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 2:38 PM, Judith Blake >>>>>>>>>>> <Judith.Blake at jax.org> wrote: >>>>>>>>>>> >>>>>>>>>>> Functions are occurrants. A function describes an action, >>>>>>>>>>> biochemical, structural, whatever. >>>>>>>>>>> >>>>>>>>>>> A function doesn't describe anything. Information artifact >>>>>>>>>>> describe things. >>>>>>>>>>> Sorry to be picky about words. It's my job. >>>>>>>>>>> If you look at the papers on BFO you will see that functions are >>>>>>>>>>> continuants and occurrents are disjoint from continuants. So >>>>>>>>>>> functions are not occurrents. If the GO is using terminology at >>>>>>>>>>> odds with BFO that needs to be fixed - we agreed at the last >>>>>>>>>>> foundry meeting that we were all going to use a common upper level >>>>>>>>>>> ontology. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Functions are part of processes. >>>>>>>>>>> >>>>>>>>>>> Function*ings* (i.e. realizations of functions) are part of >>>>>>>>>>> processes. >>>>>>>>>>> >>>>>>>>>>> A variety of structures, encoded in a molecule, have the potential >>>>>>>>>>> to realize a function. >>>>>>>>>>> >>>>>>>>>>> Type error: processes don't have potential. (concluded by >>>>>>>>>>> substitution "potential to realize a function" + function is_a >>>>>>>>>>> process => "potential to realize a process". >>>>>>>>>>> >>>>>>>>>>> The error is that it is the functions that are the potentials. >>>>>>>>>>> Potentials are realized. The realizations are processes. >>>>>>>>>>> >>>>>>>>>>> We discussed this very recently with Michael and he agreed. >>>>>>>>>>> >>>>>>>>>>> That is a property of the molecule, not of the function. >>>>>>>>>>> >>>>>>>>>>> What is the reference of "That"? >>>>>>>>>>> >>>>>>>>>>> A molecule can have encoded the potential of engaging in several >>>>>>>>>>> functions. >>>>>>>>>>> >>>>>>>>>>> Absolutely. (as long as you rewrite "functions" -> >>>>>>>>>>> "functionings" (i.e. processes).) >>>>>>>>>>> >>>>>>>>>>> A molecule can bear multiple functions is the way we would say >>>>>>>>>>> this. >>>>>>>>>>> >>>>>>>>>>> The confusion I think is that of the relationship between the gene >>>>>>>>>>> product and a molecular function, and the molecular function and >>>>>>>>>>> the biological process. >>>>>>>>>>> >>>>>>>>>>> And what do you say these are. I (and BFO) are clear: The relation >>>>>>>>>>> between a molecular function and a gene product is "inheres". The >>>>>>>>>>> relation between molecular function and a biological process is >>>>>>>>>>> "realized_by". >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Judy >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On 9/29/09 1:52 PM, "Alan Ruttenberg" >>>>>>>>>>> <alanruttenberg at gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>> The problem is that this still confuses function with process. >>>>>>>>>>> Functions are the things that the enzymes (using the term >>>>>>>>>>> broadly) have that make the processes happen. If one goes into >>>>>>>>>>> more detail, there are a variety of participants in the process >>>>>>>>>>> (substrates, products, "cofactors"), but only one bearer of the >>>>>>>>>>> function in a particular process. >>>>>>>>>>> >>>>>>>>>>> Every function has a corresponding process, its "realization". >>>>>>>>>>> >>>>>>>>>>> Functions are not parts of processes. The realizations of >>>>>>>>>>> functions are. >>>>>>>>>>> >>>>>>>>>>> The functions have a physical basis in the structure of their >>>>>>>>>>> bearers. >>>>>>>>>>> >>>>>>>>>>> The functions exist before any process happens. A molecule may >>>>>>>>>>> have a function but never actually realize it. >>>>>>>>>>> >>>>>>>>>>> These distinctions have important consequences as people such as >>>>>>>>>>> myself work on adding more detail to the GO. The processes get >>>>>>>>>>> more detail by adding *participants*. Within those processes, when >>>>>>>>>>> there is a molecule like an enzyme that functions in a particular >>>>>>>>>>> way, it needs to be distinguished from the other participants in >>>>>>>>>>> some way. The way that it is distinguished is by saying it bears a >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> The heuristic that a molecular function is a "single step" process >>>>>>>>>>> is misused, IMO. That it is a "single step" derives from the fact >>>>>>>>>>> that in such descriptions there is a single molecule that has a >>>>>>>>>>> function. The step is the process that surrounds execution of that >>>>>>>>>>> function. >>>>>>>>>>> >>>>>>>>>>> As I said, getting this distinction clear is essential for future >>>>>>>>>>> detailing of GO processes such as I and others do. Can we start to >>>>>>>>>>> get this right here and now? >>>>>>>>>>> >>>>>>>>>>> -Alan >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> On Tue, Sep 29, 2009 at 10:30 AM, Midori Harris >>>>>>>>>>> <midori at ebi.ac.uk> wrote: >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> We have two SF items in which merging function and process terms >>>>>>>>>>> looks like a viable solution to a problem. One is an overly- >>>>>>>>>>> specific, single-step "process" that is essentially equivalent to >>>>>>>>>>> a "function" (in the GO sense); the other is a "function" term >>>>>>>>>>> that is defined such that it doesn't describe a single activity, >>>>>>>>>>> and is essentially the same as a process term. >>>>>>>>>>> >>>>>>>>>>> SF 2864212 - merge MF term >>>>>>>>>>> RNA splicing factor activity, transesterification mechanism GO: >>>>>>>>>>> 0031202 >>>>>>>>>>> >>>>>>>>>>> into BP term >>>>>>>>>>> RNA splicing, via transesterification reactions GO:0000375 >>>>>>>>>>> >>>>>>>>>>> For this one, we would merge the MF term into the BP term, and >>>>>>>>>>> ensure that the merged term has is_a ancestry in the BP graph. I >>>>>>>>>>> think that's all we would change. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864212&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> SF 2864271 - merge BP term >>>>>>>>>>> >>>>>>>>>>> conversion of met-tRNAf to fmet-tRNA GO:0001718 >>>>>>>>>>> >>>>>>>>>>> into MF term >>>>>>>>>>> methionyl-tRNA formyltransferase activity GO:0004479 >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> https://sourceforge.net/tracker/? >>>>>>>>>>>> >>>>>>>>>>> func=detail&aid=2864271&group_id=36855&atid=440764 >>>>>>>>>>> >>>>>>>>>>> In this case, the proposal is to merge the process term into the >>>>>>>>>>> function term, ensuring that the merged term has is_a ancestry >>>>>>>>>>> only in the MF ontology,and has a part_of link to a process term. >>>>>>>>>>> Specifically, the merged term would retain the name and ID from >>>>>>>>>>> GO: >>>>>>>>>>> 0004479, and would be part_of translational initiation GO:0006413. >>>>>>>>>>> >>>>>>>>>>> Does anyone have any objections or other comments on these? >>>>>>>>>>> >>>>>>>>>>> thanks, >>>>>>>>>>> m >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> _______________________________________________ >>>>>>>>>>> Annotation mailing list >>>>>>>>>>> Annotation at geneontology.org >>>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> _______________________________________________ >>>>>>>>>> Ontology-editors mailing list >>>>>>>>>> Ontology-editors at geneontology.org >>>>>>>>>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >>>>>>>>>> >>>>>>>>> -- >>>>>>>>> Dr Jane Lomax >>>>>>>>> GO Editorial Office >>>>>>>>> EMBL-EBI >>>>>>>>> Wellcome Trust Genome Campus >>>>>>>>> Hinxton >>>>>>>>> Cambridgeshire, UK >>>>>>>>> CB10 1SD >>>>>>>>> >>>>>>>>> p: +44 1223 492516 >>>>>>>>> f: +44 1223 494468 >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> ------------------------------------------------------------------------ >>>>>>>> _______________________________________________ >>>>>>>> Annotation mailing list >>>>>>>> Annotation at geneontology.org >>>>>>>> http://fafner.stanford.edu/mailman/listinfo/annotation >>>>>>>> >>>>>>> >>>>>>> >>>>>> >>>>> >>>>> >>>> >>> >>> -- >>> David P. Hill, Ph.D. >>> Bioinformatics Scientist: Ontology Development >>> Gene Ontology Consortium >>> The Jackson Laboratory >>> www.geneontology.org >>> www.informatics.jax.org >>> tel:207-288-6430 >>> >>> _______________________________________________ >>> Ontology-editors mailing list >>> Ontology-editors at geneontology.org >>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors >> >> -- >> Amelia Ireland >> GO Editorial Office >> http://www.berkeleybop.org || http://www.ebi.ac.uk >> Boycott Trader Joe's Red List seafood: http://traitorjoe.com >> >> >> >> >> >> >> _______________________________________________ >> Ontology-editors mailing list >> Ontology-editors at geneontology.org >> http://fafner.stanford.edu/mailman/listinfo/ontology-editors > _______________________________________________ > Annotation mailing list > Annotation at geneontology.org > http://fafner.stanford.edu/mailman/listinfo/annotation > -- Dr Jane Lomax GO Editorial Office EMBL-EBI Wellcome Trust Genome Campus Hinxton Cambridgeshire, UK CB10 1SD p: +44 1223 492516 f: +44 1223 494468 From jdeegan at ebi.ac.uk Thu Oct 8 06:35:58 2009 From: jdeegan at ebi.ac.uk (Jennifer Deegan (nee Clark)) Date: Thu, 08 Oct 2009 14:35:58 +0100 Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <360f86f4a71ee96ad65e87ab9c76bdd0.squirrel@squirrelmail.internal.sanger.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> <4ACDDDCC.9080105@informatics.jax.org> <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> <4ACDE4AB.8090308@informatics.jax.org> <360f86f4a71ee96ad65e87ab9c76bdd0.squirrel@squirrelmail.internal.sanger.ac.uk> Message-ID: <4ACDEABE.10203@ebi.ac.uk> Hi Val, > > That would work. If people agree we could rename protein tag to this now, > the definition would need to be broadened but it would not affect existing > annotations. > The process link couldn't be added as present as we have no term > "signaling". Is this part of Jen's overhaul? Yes I can do this in my branch file if it's agreed. I'd really appreciate if the group could agree a good definition though, and something really clear to put in the comments field. I can see how this term might be confusing to people who haven't witnessed the original discussion. Thanks, Jennifer From jane at ebi.ac.uk Thu Oct 8 07:32:54 2009 From: jane at ebi.ac.uk (Jane Lomax) Date: Thu, 8 Oct 2009 15:32:54 +0100 (BST) Subject: [Ontology-editors] [Annotation] merging MF and BP terms In-Reply-To: <4ACDEABE.10203@ebi.ac.uk> References: <4ACC8F79.2060808@informatics.jax.org> <4ACCF696.2040408@informatics.jax.org> <4ACCF8E3.4080708@informatics.jax.org> <0F7D5752-3EF0-4437-9A33-EEF50A60CB99@ebi.ac.uk> <4ACDDDCC.9080105@informatics.jax.org> <1008d7666d31443317725196da0e5485.squirrel@squirrelmail.internal.sanger.ac.uk> <4ACDE4AB.8090308@informatics.jax.org> <360f86f4a71ee96ad65e87ab9c76bdd0.squirrel@squirrelmail.internal.sanger.ac.uk> <4ACDEABE.10203@ebi.ac.uk> Message-ID: Hi - okay, here is the list of GO MF terms that I don't think are processes. There are about 1500, so 17% of MF: GO:0046935 1-phosphatidylinositol-3-kinase regulator activity GO:0033224 2-aminoethylphosphonate transmembrane transporter activity GO:0015649 2-keto-3-deoxygluconate:hydrogen symporter activity GO:0046964 3'-phosphoadenosine 5'-phosphosulfate transmembrane transporter activity GO:0015551 3-hydroxyphenyl propanoate transmembrane transporter activity GO:0015540 3-hydroxyphenyl propionate:hydrogen ion symporter activity GO:0042926 3-hydroxyphenylpropionic acid transporter activity GO:0042890 3-phenylpropionic acid transmembrane transporter activity GO:0015231 5-formyltetrahydrofolate transporter activity GO:0001586 5-HT1 receptor activity GO:0001587 5-HT2 receptor activity GO:0019909 [pyruvate dehydrogenase (lipoamide)] phosphatase regulator activity GO:0005250 A-type (transient outward) potassium channel activity GO:0015123 acetate transmembrane transporter activity GO:0043893 acetate:cation symporter activity GO:0015360 acetate:hydrogen symporter activity GO:0008521 acetyl-CoA transporter activity GO:0015325 acetyl-CoA:CoA antiporter activity GO:0030549 acetylcholine receptor activator activity GO:0015464 acetylcholine receptor activity GO:0030550 acetylcholine receptor inhibitor activity GO:0030548 acetylcholine receptor regulator activity GO:0005277 acetylcholine transmembrane transporter activity GO:0022848 acetylcholine-gated cation channel activity GO:0005278 acetylcholine:hydrogen antiporter activity GO:0010307 acetylglutamate kinase regulator activity GO:0010698 acetyltransferase activator activity GO:0042934 achromobactin transporter activity GO:0015172 acidic amino acid transmembrane transporter activity GO:0042911 acridine transporter activity GO:0042962 acridine:hydrogen antiporter activity GO:0015555 acriflavine resistant pump activity GO:0015566 acriflavine transporter activity GO:0032397 activating MHC class I receptor activity GO:0022853 active ion transmembrane transporter activity GO:0022854 active large uncharged polar molecule transmembrane transporter activity GO:0022804 active transmembrane transporter activity GO:0017002 activin receptor activity GO:0016361 activin receptor activity, type I GO:0016362 activin receptor activity, type II GO:0015227 acyl carnitine transporter activity GO:0000036 acyl carrier activity GO:0000295 adenine nucleotide transmembrane transporter activity GO:0015207 adenine transmembrane transporter activity GO:0001609 adenosine receptor activity, G-protein coupled GO:0000774 adenyl-nucleotide exchange factor activity GO:0010856 adenylate cyclase activator activity GO:0001640 adenylate cyclase inhibiting metabotropic glutamate receptor activity GO:0010855 adenylate cyclase inhibitor activity GO:0010854 adenylate cyclase regulator activity GO:0015475 adhesin autotransporter activity GO:0015217 ADP transmembrane transporter activity GO:0045032 ADP-activated nucleotide receptor activity GO:0004935 adrenoceptor activity GO:0004978 adrenocorticotropin receptor activity GO:0001605 adrenomedullin receptor activity GO:0050785 advanced glycation end-product receptor activity GO:0043861 agmatine:putrescine antiporter activity GO:0022858 alanine transmembrane transporter activity GO:0015655 alanine:sodium symporter activity GO:0015665 alcohol transmembrane transporter activity GO:0042876 aldarate transmembrane transporter activity GO:0042879 aldonate transmembrane transporter activity GO:0015567 alkane transporter activity GO:0042959 alkanesulfonate transporter activity GO:0042917 alkylphosphonate transmembrane transporter activity GO:0015124 allantoate transmembrane transporter activity GO:0005274 allantoin uptake transmembrane transporter activity GO:0008261 allatostatin receptor activity GO:0015593 allose transmembrane transporter activity GO:0016964 alpha-2 macroglobulin receptor activity GO:0004936 alpha-adrenergic receptor activity GO:0004971 alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionate selective glutamate receptor activity GO:0015066 alpha-amylase inhibitor activity GO:0015151 alpha-glucoside transmembrane transporter activity GO:0005352 alpha-glucoside:hydrogen symporter activity GO:0015139 alpha-ketoglutarate transmembrane transporter activity GO:0015532 alpha-ketoglutarate:hydrogen symporter activity GO:0004937 alpha1-adrenergic receptor activity GO:0004938 alpha2-adrenergic receptor activity GO:0015083 aluminum ion transmembrane transporter activity GO:0042887 amide transporter activity GO:0015240 amiloride transporter activity GO:0015280 amiloride-sensitive sodium channel activity GO:0015308 amiloride:hydrogen antiporter activity GO:0008227 amine receptor activity GO:0005275 amine transmembrane transporter activity GO:0015171 amino acid transmembrane transporter activity GO:0015247 aminophospholipid transporter activity GO:0015241 aminotriazole transporter activity GO:0015314 aminotriazole:hydrogen antiporter activity GO:0051739 ammonia transporter activity GO:0015251 ammonium channel activity GO:0008519 ammonium transmembrane transporter activity GO:0080122 AMP transmembrane transporter activity GO:0004942 anaphylatoxin receptor activity GO:0004882 androgen receptor activity GO:0001595 angiotensin receptor activity GO:0001596 angiotensin type I receptor activity GO:0004945 angiotensin type II receptor activity GO:0005253 anion channel activity GO:0015380 anion exchanger activity GO:0008509 anion transmembrane transporter activity GO:0015301 anion:anion antiporter activity GO:0015296 anion:cation symporter activity GO:0043856 anti-sigma factor antagonist activity GO:0042895 antibiotic transporter activity GO:0042960 antimonite secondary active transmembrane transporter activity GO:0015104 antimonite transmembrane transporter activity GO:0015297 antiporter activity GO:0060182 apelin receptor activity GO:0034188 apolipoprotein A-I receptor activity GO:0030226 apolipoprotein receptor activity GO:0016505 apoptotic protease activator activity GO:0015523 arabinose efflux transmembrane transporter activity GO:0042901 arabinose polymer transmembrane transporter activity GO:0042900 arabinose transmembrane transporter activity GO:0015518 arabinose:hydrogen symporter activity GO:0015167 arabitol transmembrane transporter activity GO:0042951 arbutin transmembrane transporter activity GO:0008060 ARF GTPase activator activity GO:0005086 ARF guanyl-nucleotide exchange factor activity GO:0015557 arginine targeting transmembrane transporter activity GO:0015181 arginine transmembrane transporter activity GO:0043862 arginine:agmatine antiporter activity GO:0043858 arginine:ornithine antiporter activity GO:0015173 aromatic amino acid transmembrane transporter activity GO:0015494 aromatic amino acid:hydrogen symporter activity GO:0008490 arsenite secondary active transmembrane transporter activity GO:0015105 arsenite transmembrane transporter activity GO:0004874 aryl hydrocarbon receptor activity GO:0004873 asialoglycoprotein receptor activity GO:0070906 aspartate:alanine antiporter activity GO:0019828 aspartic-type endopeptidase inhibitor activity GO:0005347 ATP transmembrane transporter activity GO:0015272 ATP-activated inward rectifier potassium channel activity GO:0045031 ATP-activated nucleotide receptor activity GO:0005224 ATP-binding and phosphorylation-dependent chloride channel activity GO:0004931 ATP-gated cation channel activity GO:0005471 ATP:ADP antiporter activity GO:0001671 ATPase activator activity GO:0042030 ATPase inhibitor activity GO:0060590 ATPase regulator activity GO:0015474 autotransporter activity GO:0015457 auxiliary transport protein activity GO:0010329 auxin efflux transmembrane transporter activity GO:0010328 auxin influx transmembrane transporter activity GO:0009672 auxin:hydrogen symporter activity GO:0008046 axon guidance receptor activity GO:0045118 azole transporter activity GO:0045119 azole:hydrogen antiporter activity GO:0022885 bacteriocin transmembrane transporter activity GO:0017163 basal transcription repressor activity GO:0015174 basic amino acid transmembrane transporter activity GO:0015242 benomyl transporter activity GO:0015310 benomyl:hydrogen antiporter activity GO:0042925 benzoate transporter activity GO:0008503 benzodiazepine receptor activity GO:0004939 beta-adrenergic receptor activity GO:0001761 beta-alanine transmembrane transporter activity GO:0004979 beta-endorphin receptor activity GO:0015160 beta-glucan transmembrane transporter activity GO:0015573 beta-glucoside transmembrane transporter activity GO:0004940 beta1-adrenergic receptor activity GO:0004941 beta2-adrenergic receptor activity GO:0015052 beta3-adrenergic receptor activity GO:0015199 betaine transmembrane transporter activity GO:0015106 bicarbonate transmembrane transporter activity GO:0015545 bicyclomycin transporter activity GO:0015125 bile acid transmembrane transporter activity GO:0008508 bile acid:sodium symporter activity GO:0015351 bilirubin secondary active transmembrane transporter activity GO:0015127 bilirubin transmembrane transporter activity GO:0045125 bioactive lipid receptor activity GO:0015224 biopterin transporter activity GO:0015225 biotin transporter activity GO:0033038 bitter taste receptor activity GO:0009882 blue light photoreceptor activity GO:0004946 bombesin receptor activity GO:0080139 boron efflux transmembrane transporter activity GO:0046715 boron transporter activity GO:0080138 boron uptake transmembrane transporter activity GO:0008288 boss receptor activity GO:0004947 bradykinin receptor activity GO:0060175 brain-derived neurotrophic factor receptor activity GO:0016527 brain-specific angiogenesis inhibitor activity GO:0015658 branched-chain aliphatic amino acid transmembrane transporter activity GO:0015657 branched-chain amino acid:sodium symporter activity GO:0016493 C-C chemokine receptor activity GO:0016494 C-X-C chemokine receptor activity GO:0016495 C-X3-C chemokine receptor activity GO:0004943 C3a anaphylatoxin receptor activity GO:0015556 C4-dicarboxylate transmembrane transporter activity GO:0004944 C5a anaphylatoxin receptor activity GO:0015490 cadaverine transmembrane transporter activity GO:0015086 cadmium ion transmembrane transporter activity GO:0001635 calcitonin gene-related polypeptide receptor activity GO:0004948 calcitonin receptor activity GO:0005227 calcium activated cation channel activity GO:0005262 calcium channel activity GO:0019855 calcium channel inhibitor activity GO:0005246 calcium channel regulator activity GO:0015085 calcium ion transmembrane transporter activity GO:0008048 calcium sensitive guanylate cyclase activator activity GO:0008273 calcium, potassium:sodium antiporter activity GO:0015269 calcium-activated potassium channel activity GO:0010859 calcium-dependent cysteine-type endopeptidase inhibitor activity GO:0008427 calcium-dependent protein kinase inhibitor activity GO:0010858 calcium-dependent protein kinase regulator activity GO:0008597 calcium-dependent protein serine/threonine phosphatase regulator activity GO:0048763 calcium-induced calcium release activity GO:0015278 calcium-release channel activity GO:0015368 calcium:cation antiporter activity GO:0015369 calcium:hydrogen antiporter activity GO:0005432 calcium:sodium antiporter activity GO:0001646 cAMP receptor activity GO:0008140 cAMP response element binding protein binding GO:0004862 cAMP-dependent protein kinase inhibitor activity GO:0008603 cAMP-dependent protein kinase regulator activity GO:0015126 canalicular bile acid transmembrane transporter activity GO:0004949 cannabinoid receptor activity GO:0034291 canonical holin activity GO:0015161 capsular polysaccharide transmembrane transporter activity GO:0015144 carbohydrate transmembrane transporter activity GO:0015549 carbonyl cyanide m-chlorophenylhydrazone transporter activity GO:0046943 carboxylic acid transmembrane transporter activity GO:0015226 carnitine transporter activity GO:0005476 carnitine:acyl carnitine antiporter activity GO:0008656 caspase activator activity GO:0043027 caspase inhibitor activity GO:0043028 caspase regulator activity GO:0005261 cation channel activity GO:0046583 cation efflux transmembrane transporter activity GO:0008324 cation transmembrane transporter activity GO:0005416 cation:amino acid symporter activity GO:0015491 cation:cation antiporter activity GO:0015377 cation:chloride symporter activity GO:0005402 cation:sugar symporter activity GO:0015326 cationic amino acid transmembrane transporter activity GO:0019191 cellobiose transmembrane transporter activity GO:0015267 channel activity GO:0016248 channel inhibitor activity GO:0016249 channel localizer activity GO:0016247 channel regulator activity GO:0022886 channel-forming ionophore activity GO:0042056 chemoattractant activity GO:0004950 chemokine receptor activity GO:0046817 chemokine receptor antagonist activity GO:0045499 chemorepellent activity GO:0042896 chloramphenicol transporter activity GO:0015107 chlorate transmembrane transporter activity GO:0005254 chloride channel activity GO:0019869 chloride channel inhibitor activity GO:0017081 chloride channel regulator activity GO:0015108 chloride transmembrane transporter activity GO:0070089 chloride-activated potassium channel activity GO:0010290 chlorophyll catabolite transporter activity GO:0004951 cholecystokinin receptor activity GO:0017127 cholesterol transporter activity GO:0015220 choline transmembrane transporter activity GO:0005307 choline:sodium symporter activity GO:0005429 chromaffin granule amine transmembrane transporter activity GO:0015109 chromate transmembrane transporter activity GO:0070835 chromium ion transmembrane transporter activity GO:0042933 chrysobactin transporter activity GO:0004897 ciliary neurotrophic factor receptor activity GO:0015137 citrate transmembrane transporter activity GO:0015531 citrate:hydrogen symporter activity GO:0015515 citrate:succinate antiporter activity GO:0005456 CMP-sialic acid transmembrane transporter activity GO:0015235 cobalamin transporter activity GO:0015087 cobalt ion transmembrane transporter activity GO:0015228 coenzyme A transmembrane transporter activity GO:0051185 coenzyme transporter activity GO:0051184 cofactor transporter activity GO:0042912 colicin transmembrane transporter activity GO:0001863 collectin receptor activity GO:0001857 complement component C1q receptor activity GO:0004876 complement component C3a receptor activity GO:0004877 complement component C3b receptor activity GO:0001860 complement component C3d receptor activity GO:0001859 complement component C3dg receptor activity GO:0001861 complement component C4b receptor activity GO:0004878 complement component C5a receptor activity GO:0001858 complement component iC3b receptor activity GO:0004875 complement receptor activity GO:0005357 constitutive hydrogen:glucose symporter activity GO:0016531 copper chaperone activity GO:0051981 copper chelate transmembrane transporter activity GO:0005375 copper ion transmembrane transporter activity GO:0015088 copper uptake transmembrane transporter activity GO:0051982 copper-nicotianamine transmembrane transporter activity GO:0035237 corazonin receptor activity GO:0015026 coreceptor activity GO:0060724 coreceptor activity involved in epidermal growth factor receptor signaling pathway GO:0015028 coreceptor, insoluble ligand activity GO:0015027 coreceptor, soluble ligand activity GO:0015056 corticotrophin-releasing factor receptor activity GO:0043404 corticotropin-releasing hormone receptor activity GO:0031963 cortisol receptor activity GO:0005308 creatine transporter activity GO:0005309 creatine:sodium symporter activity GO:0015110 cyanate transmembrane transporter activity GO:0010853 cyclase activator activity GO:0010852 cyclase inhibitor activity GO:0010851 cyclase regulator activity GO:0017132 cyclic nucleotide-dependent guanyl-nucleotide exchange factor activity GO:0043855 cyclic nucleotide-gated ion channel activity GO:0043854 cyclic nucleotide-gated mechanosensitive ion channel activity GO:0016534 cyclin-dependent protein kinase 5 activator activity GO:0016536 cyclin-dependent protein kinase 5 activator regulator activity GO:0019914 cyclin-dependent protein kinase activating kinase regulator activity GO:0004861 cyclin-dependent protein kinase inhibitor activity GO:0016538 cyclin-dependent protein kinase regulator activity GO:0015243 cycloheximide transporter activity GO:0015309 cycloheximide:hydrogen antiporter activity GO:0033229 cysteine transmembrane transporter activity GO:0004869 cysteine-type endopeptidase inhibitor activity GO:0001631 cysteinyl leukotriene receptor activity GO:0015328 cystine secondary active transmembrane transporter activity GO:0015327 cystine:glutamate antiporter activity GO:0015212 cytidine transmembrane transporter activity GO:0004896 cytokine receptor activity GO:0009884 cytokinin receptor activity GO:0015209 cytosine transmembrane transporter activity GO:0015504 cytosine:hydrogen ion symporter activity GO:0042944 D-alanine transmembrane transporter activity GO:0042943 D-amino acid transmembrane transporter activity GO:0042877 D-galactarate transmembrane transporter activity GO:0042881 D-galactonate transmembrane transporter activity GO:0042878 D-glucarate transmembrane transporter activity GO:0055056 D-glucose transmembrane transporter activity GO:0055055 D-glucose:hydrogen symporter activity GO:0042880 D-glucuronate transmembrane transporter activity GO:0048474 D-methionine transmembrane transporter activity O:0015591 D-ribose transmembrane transporter activity GO:0042945 D-serine transmembrane transporter activity GO:0015148 D-xylose transmembrane transporter activity GO:0015519 D-xylose:hydrogen symporter activity GO:0005035 death receptor activity GO:0015451 decarboxylation-driven active transmembrane transporter activity GO:0005040 decoy death receptor activity GO:0033300 dehydroascorbic acid transporter activity GO:0005251 delayed rectifier potassium channel activity GO:0004986 delta-opioid receptor activity GO:0030233 deoxynucleotide transmembrane transporter activity GO:0060703 deoxyribonuclease inhibitor activity GO:0022859 dephosphorylation-gated channel activity GO:0042957 dextrin transmembrane transporter activity GO:0015082 di-, tri-valent inorganic cation transmembrane transporter activity GO:0015364 dicarboxylate:inorganic phosphate antiporter activity GO:0005310 dicarboxylic acid transmembrane transporter activity GO:0015270 dihydropyridine-sensitive calcium channel activity GO:0042936 dipeptide transporter activity GO:0015154 disaccharide transmembrane transporter activity GO:0008036 diuretic hormone receptor activity GO:0030337 DNA polymerase processivity factor activity GO:0008657 DNA topoisomerase (ATP-hydrolyzing) inhibitor activity GO:0051035 DNA transmembrane transporter activity GO:0001590 dopamine D1 receptor activity GO:0001670 dopamine D2 receptor activity GO:0001592 dopamine D3 receptor activity GO:0001593 dopamine D4 receptor activity GO:0001589 dopamine D5 receptor activity GO:0004952 dopamine receptor activity GO:0001591 dopamine receptor activity, coupled via Gi/Go GO:0001588 dopamine receptor activity, coupled via Gs GO:0005329 dopamine transmembrane transporter activity GO:0005330 dopamine:sodium symporter activity GO:0015238 drug transporter activity GO:0015307 drug:hydrogen antiporter activity GO:0004858 dUTP pyrophosphatase inhibitor activity GO:0042654 ecdysis-triggering hormone receptor activity GO:0004884 ecdysteroid hormone receptor activity GO:0051471 ectoine transmembrane transporter activity GO:0015562 efflux transmembrane transporter activity GO:0009055 electron carrier activity GO:0045155 electron transporter, transferring electrons from CoQH2-cytochrome c reductase complex and cytochrome c oxidase complex activity GO:0046028 electron transporter, transferring electrons from cytochrome b6/f complex of photosystem II activity GO:0045153 electron transporter, transferring electrons within CoQH2-cytochrome c reductase complex activity GO:0045158 electron transporter, transferring electrons within cytochrome b6/f complex of photosystem II activity GO:0045154 electron transporter, transferring electrons within cytochrome c oxidase complex activity GO:0045156 electron transporter, transferring electrons within the cyclic electron transport pathway of photosynthesis activity GO:0045157 electron transporter, transferring electrons within the noncyclic electron transport pathway of photosynthesis activity GO:0004866 endopeptidase inhibitor activity GO:0060698 endoribonuclease inhibitor activity GO:0004962 endothelin receptor activity GO:0001599 endothelin-A receptor activity GO:0001600 endothelin-B receptor activity GO:0015570 energizer of outer membrane receptor-mediated transport activity GO:0031992 energy transducer activity GO:0042931 enterobactin transporter activity GO:0008047 enzyme activator activity GO:0004857 enzyme inhibitor activity GO:0030234 enzyme regulator activity GO:0005003 ephrin receptor activity GO:0005006 epidermal growth factor receptor activity GO:0001625 Epstein-Barr Virus-induced receptor activity GO:0015397 equilibrative nucleoside transmembrane transporter, nitrobenzyl-thioinosine-insensitive activity GO:0015396 equilibrative nucleoside transmembrane transporter, nitrobenzyl-thioinosine-sensitive activity GO:0004900 erythropoietin receptor activity GO:0030284 estrogen receptor activity GO:0034228 ethanolamine transmembrane transporter activity GO:0042557 eukaryotic elongation factor-2 kinase activator activity GO:0042556 eukaryotic elongation factor-2 kinase regulator activity GO:0005231 excitatory extracellular ligand-gated ion channel activity GO:0005230 extracellular ligand-gated ion channel activity GO:0030023 extracellular matrix constituent conferring elasticity GO:0030197 extracellular matrix constituent, lubricant activity GO:0005201 extracellular matrix structural constituent GO:0030021 extracellular matrix structural constituent conferring compression resistance GO:0030020 extracellular matrix structural constituent conferring tensile strength GO:0008068 extracellular-glutamate-gated chloride channel activity GO:0005234 extracellular-glutamate-gated ion channel activity GO:0016934 extracellular-glycine-gated chloride channel activity GO:0016933 extracellular-glycine-gated ion channel activity GO:0005395 eye pigment precursor transporter activity GO:0022814 facilitated diffusion GO:0005472 FAD carrier activity GO:0015230 FAD transmembrane transporter activity GO:0031516 far-red light photoreceptor activity GO:0015245 fatty acid transporter activity GO:0015607 fatty acyl CoA transporter activity GO:0015246 fatty acyl transporter activity GO:0015345 ferric enterobactin:hydrogen symporter activity GO:0015091 ferric iron transmembrane transporter activity GO:0015621 ferric triacetylfusarinine C transmembrane transporter activity GO:0015346 ferric triacetylfusarinine C:hydrogen symporter activity GO:0015620 ferric-enterobactin transmembrane transporter activity GO:0015622 ferric-hydroxamate transmembrane transporter activity GO:0019535 ferric-vibriobactin transmembrane transporter activity GO:0042929 ferrichrome transporter activity GO:0070287 ferritin receptor activity GO:0015093 ferrous iron transmembrane transporter activity GO:0015639 ferrous iron uptake transmembrane transporter activity GO:0005007 fibroblast growth factor receptor activity GO:0030353 fibroblast growth factor receptor antagonist activity GO:0015473 fimbrial usher porin activity GO:0015244 fluconazole transporter activity GO:0015313 fluconazole:hydrogen antiporter activity GO:0008517 folic acid transporter activity GO:0004963 follicle-stimulating hormone receptor activity GO:0015660 formate efflux transmembrane transporter activity GO:0015499 formate transmembrane transporter activity GO:0015659 formate uptake transmembrane transporter activity GO:0042898 fosmidomycin transporter activity GO:0005353 fructose transmembrane transporter activity GO:0015284 fructose uniporter activity GO:0055054 fructose:hydrogen symporter activity GO:0015150 fucose transmembrane transporter activity GO:0015535 fucose:hydrogen symporter activity GO:0015138 fumarate transmembrane transporter activity GO:0015467 G-protein activated inward rectifier potassium channel activity GO:0001637 G-protein chemoattractant receptor activity GO:0016907 G-protein coupled acetylcholine receptor activity GO:0001647 G-protein coupled cytokinin receptor activity GO:0008020 G-protein coupled photoreceptor activity GO:0004930 G-protein coupled receptor activity GO:0016609 G-protein coupled serotonin receptor activity GO:0016917 GABA receptor activity GO:0004890 GABA-A receptor activity GO:0004965 GABA-B receptor activity GO:0022851 GABA-gated chloride ion channel activity GO:0015577 galactitol transmembrane transporter activity GO:0019196 galactosamine transmembrane transporter activity GO:0005354 galactose transmembrane transporter activity GO:0050782 galactose uniporter activity GO:0015517 galactose:hydrogen symporter activity GO:0015371 galactose:sodium symporter activity GO:0015550 galacturonate transmembrane transporter activity GO:0004966 galanin receptor activity GO:0015495 gamma-aminobutyric acid:hydrogen symporter activity GO:0005332 gamma-aminobutyric acid:sodium symporter activity GO:0005243 gap junction channel activity GO:0055077 gap junction hemi-channel activity GO:0016519 gastric inhibitory peptide receptor activity GO:0015054 gastrin receptor activity GO:0001628 gastropyloric receptor activity GO:0022836 gated channel activity GO:0005092 GDP-dissociation inhibitor activity GO:0016219 GDP-dissociation stimulator activity GO:0005457 GDP-fucose transmembrane transporter activity GO:0005458 GDP-mannose transmembrane transporter activity GO:0016251 general RNA polymerase II transcription factor activity GO:0016565 general transcriptional repressor activity GO:0030160 GKAP/Homer scaffold activity GO:0016167 glial cell line-derived neurotrophic factor receptor activity GO:0004967 glucagon receptor activity GO:0004883 glucocorticoid receptor activity GO:0015128 gluconate transmembrane transporter activity GO:0008524 glucose 6-phosphate:phosphate antiporter activity GO:0005355 glucose transmembrane transporter activity GO:0015304 glucose uniporter activity GO:0015152 glucose-6-phosphate transmembrane transporter activity GO:0005412 glucose:sodium symporter activity GO:0042947 glucoside transmembrane transporter activity GO:0051474 glucosylglycerol transmembrane transporter activity GO:0015135 glucuronate transmembrane transporter activity GO:0015488 glucuronide:monovalent cation symporter activity GO:0015164 glucuronoside transmembrane transporter activity GO:0008066 glutamate receptor activity GO:0022849 glutamate-gated calcium ion channel activity GO:0070909 glutamate:gamma-aminobutyric acid antiporter activity GO:0015501 glutamate:sodium symporter activity GO:0034634 glutathione transmembrane transporter activity GO:0015503 glutathione-regulated potassium exporter activity GO:0051408 glyceraldehyde 3-phosphate:inorganic phosphate antiporter activity GO:0015254 glycerol channel activity GO:0015168 glycerol transmembrane transporter activity GO:0015169 glycerol-3-phosphate transmembrane transporter activity GO:0015527 glycerol-phosphate:inorganic phosphate antiporter activity GO:0051407 glycerone phosphate:inorganic phosphate antiporter activity GO:0001406 glycerophosphodiester transmembrane transporter activity GO:0015653 glycine betaine:hydrogen symporter activity GO:0015187 glycine transmembrane transporter activity GO:0022852 glycine-gated chloride ion channel activity GO:0004891 glycine-inhibited chloride channel activity GO:0015375 glycine:sodium symporter activity GO:0043879 glycolate transmembrane transporter activity GO:0017089 glycolipid transporter activity GO:0034202 glycolipid-translocating activity GO:0034437 glycoprotein transporter activity GO:0015486 glycoside-pentoside-hexuronide:cation symporter activity GO:0004968 gonadotropin-releasing hormone receptor activity GO:0005004 GPI-linked ephrin receptor activity GO:0004902 granulocyte colony-stimulating factor receptor activity GO:0004901 granulocyte macrophage colony-stimulating factor receptor activity GO:0042913 group A colicin transmembrane transporter activity GO:0001641 group II metabotropic glutamate receptor activity GO:0001642 group III metabotropic glutamate receptor activity GO:0004903 growth hormone receptor activity GO:0001616 growth hormone secretagogue receptor activity GO:0016520 growth hormone-releasing hormone receptor activity GO:0060308 GTP cyclohydrolase I regulator activity GO:0010292 GTP:GDP antiporter activity GO:0005096 GTPase activator activity GO:0005095 GTPase inhibitor activity GO:0030695 GTPase regulator activity GO:0001409 guanine nucleotide transmembrane transporter activity GO:0015208 guanine transmembrane transporter activity GO:0005085 guanyl-nucleotide exchange factor activity GO:0005091 guanyl-nucleotide exchange factor adaptor activity GO:0030250 guanylate cyclase activator activity GO:0030251 guanylate cyclase inhibitor activity GO:0030249 guanylate cyclase regulator activity GO:0042349 guiding stereospecific synthesis activity GO:0008313 gurken receptor activity GO:0008158 hedgehog receptor activity GO:0015476 hemaglutinin autotransporter activity GO:0005062 hematopoietin/interferon-class (D200-domain) cytokine receptor signal transducer activity GO:0015232 heme transporter activity GO:0005008 hepatocyte growth factor receptor activity GO:0015119 hexose phosphate transmembrane transporter activity GO:0015149 hexose transmembrane transporter activity GO:0008516 hexose uniporter activity GO:0015526 hexose-phosphate:inorganic phosphate antiporter activity GO:0009679 hexose:hydrogen symporter activity GO:0015134 hexuronate transmembrane transporter activity GO:0015539 hexuronate:cation symporter activity GO:0015163 hexuronide transmembrane transporter activity GO:0005289 high affinity arginine transmembrane transporter activity GO:0005287 high affinity basic amino acid transmembrane transporter activity GO:0015089 high affinity copper ion transmembrane transporter activity GO:0015092 high affinity ferric uptake transmembrane transporter activity GO:0015330 high affinity glutamine transmembrane transporter activity GO:0019768 high affinity IgE receptor activity GO:0019771 high affinity IgG receptor activity GO:0005316 high affinity inorganic phosphate:sodium symporter activity GO:0005291 high affinity L-histidine transmembrane transporter activity GO:0005292 high affinity lysine transmembrane transporter activity GO:0015334 high affinity oligopeptide transporter activity GO:0048249 high affinity phosphate transmembrane transporter activity GO:0015398 high affinity secondary active ammonium transmembrane transporter activity GO:0015362 high affinity sodium:dicarboxylate symporter activity GO:0015381 high affinity sulfate transmembrane transporter activity GO:0009675 high affinity sulfate:hydrogen symporter activity GO:0000006 high affinity zinc uptake transmembrane transporter activity GO:0070380 high mobility group box 1 receptor activity GO:0030373 high molecular weight B cell growth factor receptor activity GO:0008331 high voltage-gated calcium channel activity GO:0005314 high-affinity glutamate transmembrane transporter activity GO:0005358 high-affinity hydrogen:glucose symporter activity GO:0034216 high-affinity thiamin:hydrogen symporter activity GO:0005300 high-affinity tryptophan transmembrane transporter activity GO:0070506 high-density lipoprotein receptor activity GO:0004969 histamine receptor activity GO:0019182 histamine-gated chloride channel activity GO:0070907 histidine:histamine antiporter activity GO:0035034 histone acetyltransferase regulator activity GO:0046811 histone deacetylase inhibitor activity GO:0035033 histone deacetylase regulator activity GO:0030108 HLA-A specific activating MHC class I receptor activity GO:0030107 HLA-A specific inhibitory MHC class I receptor activity GO:0030109 HLA-B specific inhibitory MHC class I receptor activity GO:0030110 HLA-C specific inhibitory MHC class I receptor activity GO:0034290 holin activity GO:0042970 homoserine transmembrane transporter activity GO:0015252 hydrogen ion channel activity GO:0015078 hydrogen ion transmembrane transporter activity GO:0005280 hydrogen:amino acid symporter activity GO:0005356 hydrogen:glucose symporter activity GO:0005297 hydrogen:proline symporter activity GO:0005276 hydrogen:vesicular amine antiporter activity GO:0015522 hydrophobic uncoupler:hydrogen antiporter activity GO:0033287 hydroxyectoine transmembrane transporter activity GO:0030369 ICAM-3 receptor activity GO:0004953 icosanoid receptor activity GO:0019766 IgA receptor activity GO:0019767 IgE receptor activity GO:0019770 IgG receptor activity GO:0001793 IgM receptor activity GO:0019763 immunoglobulin receptor activity GO:0008262 importin-alpha export receptor activity GO:0000188 inactivation of MAPK activity GO:0000200 inactivation of MAPK activity during cell wall biogenesis GO:0000173 inactivation of MAPK activity involved in osmosensory signaling pathway GO:0005237 inhibitory extracellular ligand-gated ion channel activity GO:0032396 inhibitory MHC class I receptor activity GO:0005452 inorganic anion exchanger activity GO:0015103 inorganic anion transmembrane transporter activity GO:0022890 inorganic cation transmembrane transporter activity GO:0030504 inorganic diphosphate transmembrane transporter activity GO:0005315 inorganic phosphate transmembrane transporter activity GO:0015318 inorganic solute uptake transmembrane transporter activity GO:0005220 inositol 1,4,5-trisphosphate-sensitive calcium-release channel activity GO:0070917 inositol phosphoceramide synthase regulator activity GO:0008095 inositol-1,4,5-trisphosphate receptor activity GO:0005009 insulin receptor activity GO:0005284 insulin-activated sodium:amino acid symporter activity GO:0005010 insulin-like growth factor receptor activity GO:0005360 insulin-responsive hydrogen:glucose symporter activity GO:0004904 interferon receptor activity GO:0004906 interferon-gamma receptor activity GO:0004908 interleukin-1 receptor activity GO:0005152 interleukin-1 receptor antagonist activity GO:0045352 interleukin-1 Type I receptor antagonist activity GO:0045353 interleukin-1 Type II receptor antagonist activity GO:0004909 interleukin-1, Type I, activating receptor activity GO:0004910 interleukin-1, Type II, blocking receptor activity GO:0004920 interleukin-10 receptor activity GO:0004921 interleukin-11 receptor activity GO:0016517 interleukin-12 receptor activity GO:0016515 interleukin-13 receptor activity GO:0016518 interleukin-14 receptor activity GO:0042010 interleukin-15 receptor activity GO:0042012 interleukin-16 receptor activity GO:0030368 interleukin-17 receptor activity GO:0042008 interleukin-18 receptor activity GO:0042014 interleukin-19 receptor activity GO:0004911 interleukin-2 receptor activity GO:0042016 interleukin-20 receptor activity GO:0001532 interleukin-21 receptor activity GO:0042018 interleukin-22 receptor activity GO:0042020 interleukin-23 receptor activity GO:0045506 interleukin-24 receptor activity GO:0045507 interleukin-25 receptor activity GO:0045508 interleukin-26 receptor activity GO:0045509 interleukin-27 receptor activity GO:0004912 interleukin-3 receptor activity GO:0002114 interleukin-33 receptor activity GO:0070747 interleukin-35 receptor activity GO:0004913 interleukin-4 receptor activity GO:0004914 interleukin-5 receptor activity GO:0004915 interleukin-6 receptor activity GO:0004917 interleukin-7 receptor activity GO:0004918 interleukin-8 receptor activity GO:0004919 interleukin-9 receptor activity GO:0022894 Intermediate conductance calcium-activated potassium channel activity GO:0005229 intracellular calcium activated chloride channel activity GO:0005222 intracellular cAMP activated cation channel activity GO:0005223 intracellular cGMP activated cation channel activity GO:0005221 intracellular cyclic nucleotide activated cation channel activity GO:0005218 intracellular ligand-gated calcium channel activity GO:0005217 intracellular ligand-gated ion channel activity GO:0005228 intracellular sodium activated potassium channel activity GO:0005242 inward rectifier potassium channel activity GO:0070320 inward rectifier potassium channel inhibitor activity GO:0015111 iodide transmembrane transporter activity GO:0005216 ion channel activity GO:0008200 ion channel inhibitor activity GO:0022839 ion gated channel activity GO:0015075 ion transmembrane transporter activity GO:0015404 ion-gradient-driven energizer activity GO:0004970 ionotropic glutamate receptor activity GO:0034986 iron chaperone activity GO:0015603 iron chelate transmembrane transporter activity GO:0005381 iron ion transmembrane transporter activity GO:0051980 iron-nicotianamine transmembrane transporter activity GO:0019793 ISG15 carrier activity GO:0034658 isopropylmalate transmembrane transporter activity GO:0008269 JAK pathway signal transduction adaptor activity O:0004880 juvenile hormone receptor activity GO:0015277 kainate selective glutamate receptor activity GO:0004987 kappa-opioid receptor activity GO:0019209 kinase activator activity GO:0019210 kinase inhibitor activity GO:0019207 kinase regulator activity GO:0015180 L-alanine transmembrane transporter activity GO:0034639 L-amino acid efflux transmembrane transporter activity GO:0015179 L-amino acid transmembrane transporter activity GO:0015147 L-arabinose transmembrane transporter activity GO:0008520 L-ascorbate:sodium symporter activity GO:0015229 L-ascorbic acid transporter activity GO:0015182 L-asparagine transmembrane transporter activity GO:0015183 L-aspartate transmembrane transporter activity GO:0015184 L-cystine transmembrane transporter activity GO:0015626 L-diaminopimelate transmembrane transporter activity GO:0015185 L-gamma-aminobutyric acid transmembrane transporter activity GO:0005313 L-glutamate transmembrane transporter activity GO:0015186 L-glutamine transmembrane transporter activity GO:0005290 L-histidine transmembrane transporter activity GO:0034590 L-hydroxyproline transmembrane transporter activity GO:0015568 L-idonate transmembrane transporter activity GO:0015188 L-isoleucine transmembrane transporter activity GO:0015190 L-leucine transmembrane transporter activity GO:0015661 L-lysine efflux transmembrane transporter activity GO:0015189 L-lysine transmembrane transporter activity GO:0000102 L-methionine secondary active transmembrane transporter activity GO:0015191 L-methionine transmembrane transporter activity GO:0000064 L-ornithine transmembrane transporter activity GO:0015192 L-phenylalanine transmembrane transporter activity GO:0015193 L-proline transmembrane transporter activity GO:0015194 L-serine transmembrane transporter activity GO:0015195 L-threonine transmembrane transporter activity GO:0015196 L-tryptophan transmembrane transporter activity GO:0005302 L-tyrosine transmembrane transporter activity GO:0005304 L-valine transmembrane transporter activity GO:0015129 lactate transmembrane transporter activity GO:0015650 lactate:hydrogen symporter activity GO:0033568 lactoferrin receptor activity GO:0033569 lactoferrin transmembrane transporter activity GO:0042971 lactone transporter activity GO:0015155 lactose transmembrane transporter activity GO:0015528 lactose:hydrogen symporter activity GO:0005055 laminin receptor activity GO:0060072 large conductance calcium-activated potassium channel activity GO:0022815 large uncharged polar molecule transmembrane transporter activity GO:0016524 latrotoxin receptor activity GO:0015094 lead ion transmembrane transporter activity GO:0022840 leak channel activity GO:0042071 leucokinin receptor activity GO:0004923 leukemia inhibitory factor receptor activity GO:0004899 leukemia inhibitory factor receptor beta-protein activity GO:0001632 leukotriene B4 receptor activity GO:0004974 leukotriene receptor activity GO:0004879 ligand-dependent nuclear receptor activity GO:0030374 ligand-dependent nuclear receptor transcription coactivator activity GO:0022834 ligand-gated channel activity GO:0015276 ligand-gated ion channel activity GO:0003706 ligand-regulated transcription factor activity GO:0055104 ligase inhibitor activity GO:0055103 ligase regulator activity GO:0031993 light transducer activity GO:0010461 light-activated channel activity GO:0008086 light-activated voltage-gated calcium channel activity GO:0015454 light-driven active transmembrane transporter activity GO:0060229 lipase activator activity GO:0055102 lipase inhibitor activity GO:0005319 lipid transporter activity GO:0015648 lipid-linked peptidoglycan transporter activity GO:0001877 lipoarabinomannan receptor activity GO:0015634 lipopolysaccharide exporter activity GO:0001875 lipopolysaccharide receptor activity GO:0015221 lipopolysaccharide transmembrane transporter activity GO:0060230 lipoprotein lipase activator activity GO:0030228 lipoprotein receptor activity GO:0042954 lipoprotein transporter activity GO:0070892 lipoteichoic acid receptor activity GO:0010348 lithium:hydrogen antiporter activity GO:0005324 long-chain fatty acid transporter activity GO:0015483 long-chain fatty acid transporting porin activity GO:0019769 low affinity IgE receptor activity GO:0019772 low affinity IgG receptor activity GO:0015090 low affinity iron ion transmembrane transporter activity GO:0080054 low affinity nitrate transmembrane transporter activity GO:0009673 low affinity phosphate transmembrane transporter activity GO:0015400 low affinity secondary active ammonium transmembrane transporter activity GO:0015361 low affinity sodium:dicarboxylate symporter activity GO:0009676 low affinity sulfate:hydrogen symporter activity GO:0008332 low voltage-gated calcium channel activity GO:0015458 low voltage-gated potassium channel auxiliary protein activity GO:0005362 low-affinity glucose:sodium symporter activity GO:0005359 low-affinity hydrogen:glucose symporter activity GO:0022893 low-affinity tryptophan transmembrane transporter activity GO:0000007 low-affinity zinc ion transmembrane transporter activity GO:0005041 low-density lipoprotein receptor activity GO:0004964 lutropin-choriogonadotropic hormone receptor activity GO:0043872 lysine:cadaverine antiporter activity GO:0015493 lysine:hydrogen symporter activity GO:0070915 lysophosphatidic acid receptor activity GO:0051978 lysophospholipid transporter activity GO:0001619 lysosphingolipid and lysophosphatidic acid receptor activity GO:0060241 lysozyme inhibitor activity GO:0022884 macromolecule transmembrane transporter activity GO:0005011 macrophage colony stimulating factor receptor activity GO:0015095 magnesium ion transmembrane transporter activity GO:0015140 malate transmembrane transporter activity GO:0015366 malate:hydrogen symporter activity GO:0043882 malate:sodium symporter activity GO:0042958 maltodextrin transmembrane transporter activity GO:0005363 maltose transmembrane transporter activity GO:0015481 maltose transporting porin activity GO:0005364 maltose:hydrogen symporter activity GO:0005384 manganese ion transmembrane transporter activity GO:0010486 manganese:hydrogen antiporter activity GO:0015575 mannitol transmembrane transporter activity GO:0015578 mannose transmembrane transporter activity GO:0055053 mannose:hydrogen symporter activity GO:0051477 mannosylglycerate transmembrane transporter activity GO:0005078 MAP-kinase scaffold activity GO:0001623 Mas proto-oncogene receptor activity GO:0004933 mating-type a-factor pheromone receptor activity GO:0004934 mating-type alpha-factor pheromone receptor activity GO:0004932 mating-type factor pheromone receptor activity GO:0022833 mechanically gated channel activity GO:0008381 mechanically-gated ion channel activity GO:0030273 melanin-concentrating hormone receptor activity GO:0004977 melanocortin receptor activity GO:0004980 melanocyte stimulating hormone receptor activity GO:0008502 melatonin receptor activity GO:0015156 melibiose transmembrane transporter activity GO:0015487 melibiose:monovalent cation symporter activity GO:0043887 melibiose:sodium symporter activity GO:0022810 membrane potential driven uniporter activity GO:0015097 mercury ion transmembrane transporter activity GO:0046873 metal ion transmembrane transporter activity GO:0051139 metal ion:hydrogen antiporter activity GO:0016530 metallochaperone activity GO:0008191 metalloendopeptidase inhibitor activity GO:0010577 metalloenzyme activator activity GO:0048551 metalloenzyme inhibitor activity GO:0010576 metalloenzyme regulator activity GO:0048270 methionine adenosyltransferase regulator activity GO:0043865 methionine transmembrane transporter activity GO:0015350 methotrexate transporter activity GO:0015452 methyl transfer-driven active transmembrane transporter activity GO:0015264 methylammonium channel activity GO:0015200 methylammonium transmembrane transporter activity GO:0015592 methylgalactoside transmembrane transporter activity GO:0015130 mevalonate transmembrane transporter activity GO:0032393 MHC class I receptor activity GO:0032394 MHC class Ib receptor activity GO:0032395 MHC class II receptor activity GO:0017082 mineralocorticoid receptor activity GO:0034246 mitochondrial transcription initiation factor activity GO:0022809 mobile ion carrier activity GO:0060089 molecular transducer activity GO:0015098 molybdate ion transmembrane transporter activity GO:0042888 molybdenum ion transmembrane transporter activity GO:0008504 monoamine transmembrane transporter activity GO:0015311 monoamine:hydrogen antiporter activity GO:0008028 monocarboxylic acid transmembrane transporter activity GO:0008439 monophenol monooxygenase activator activity GO:0015145 monosaccharide transmembrane transporter activity GO:0015373 monovalent anion:sodium symporter activity GO:0005451 monovalent cation:hydrogen antiporter activity GO:0015077 monovalent inorganic cation transmembrane transporter activity GO:0016015 morphogen activity GO:0004988 mu-opioid receptor activity GO:0015559 multidrug efflux pump activity GO:0015353 multidrug endosomal transmembrane transporter activity GO:0015239 multidrug transporter activity GO:0015564 multidrug, alkane resistant pump activity GO:0004981 muscarinic acetylcholine receptor activity GO:0005365 myo-inositol transmembrane transporter activity GO:0005366 myo-inositol:hydrogen symporter activity GO:0005367 myo-inositol:sodium symporter activity GO:0017020 myosin phosphatase regulator activity GO:0035013 myosuppressin receptor activity GO:0015571 N-acetylgalactosamine transmembrane transporter activity GO:0015572 N-acetylglucosamine transmembrane transporter activity GO:0004982 N-formyl peptide receptor activity GO:0004972 N-methyl-D-aspartate selective glutamate receptor activity GO:0051724 NAD transporter activity GO:0015547 nalidixic acid transporter activity GO:0022842 narrow pore channel activity GO:0022831 narrow pore, gated channel activity GO:0016941 natriuretic peptide receptor activity GO:0045233 natural killer cell receptor activity GO:0008148 negative transcription elongation factor activity GO:0010465 nerve growth factor receptor activity GO:0005042 netrin receptor activity GO:0016498 neuromedin K receptor activity GO:0001607 neuromedin U receptor activity GO:0008030 neuronal pentraxin receptor activity GO:0042263 neuropeptide F receptor activity GO:0008188 neuropeptide receptor activity GO:0004983 neuropeptide Y receptor activity GO:0016492 neurotensin receptor activity, G-protein coupled GO:0030379 neurotensin receptor activity, non-G-protein coupled GO:0030594 neurotransmitter receptor activity GO:0005326 neurotransmitter transporter activity GO:0005328 neurotransmitter:sodium symporter activity GO:0005030 neurotrophin receptor activity GO:0015175 neutral amino acid transmembrane transporter activity GO:0005295 neutral amino acid:sodium symporter activity GO:0005294 neutral L-amino acid secondary active transmembrane transporter activity GO:0015374 neutral, cationic amino acid:sodium:chloride symporter activity GO:0004704 NF-kappaB-inducing kinase activity GO:0015099 nickel ion transmembrane transporter activity GO:0015663 nicotinamide mononucleotide transmembrane transporter activity GO:0034257 nicotinamide riboside transporter activity GO:0004889 nicotinic acetylcholine-activated cation-selective channel activity GO:0070553 nicotinic acid receptor activity GO:0010542 nitrate efflux transmembrane transporter activity GO:0015112 nitrate transmembrane transporter activity GO:0009671 nitrate:hydrogen symporter activity GO:0030184 nitric oxide transporter activity GO:0030235 nitric-oxide synthase regulator activity GO:0015514 nitrite efflux transmembrane transporter activity GO:0015113 nitrite transmembrane transporter activity GO:0015513 nitrite uptake transmembrane transporter activity GO:0001626 nociceptin/orphanin-FQ receptor activity GO:0004926 non-G-protein coupled 7TM receptor activity GO:0001566 non-kinase phorbol ester receptor activity GO:0001571 non-tyrosine kinase fibroblast growth factor receptor activity GO:0016252 nonspecific RNA polymerase II transcription factor activity GO:0005334 norephinephrine:sodium symporter activity GO:0005333 norepinephrine transmembrane transporter activity GO:0051032 nucleic acid transmembrane transporter activity GO:0015205 nucleobase transmembrane transporter activity GO:0015932 nucleobase, nucleoside, nucleotide and nucleic acid transmembrane transporter activity GO:0015391 nucleobase:cation symporter activity GO:0005487 nucleocytoplasmic transporter activity GO:0005337 nucleoside transmembrane transporter activity GO:0010174 nucleoside transmembrane transporter activity, against a concentration gradient GO:0015395 nucleoside transmembrane transporter activity, down a concentration gradient GO:0015471 nucleoside-specific channel forming porin activity GO:0060589 nucleoside-triphosphatase regulator activity GO:0015506 nucleoside:hydrogen symporter activity GO:0005415 nucleoside:sodium symporter activity GO:0016502 nucleotide receptor activity GO:0001608 nucleotide receptor activity, G-protein coupled GO:0015215 nucleotide transmembrane transporter activity GO:0005338 nucleotide-sugar transmembrane transporter activity GO:0005340 nucleotide-sulfate transmembrane transporter activity GO:0045735 nutrient reservoir activity GO:0004989 octopamine receptor activity GO:0043563 odorant transporter activity GO:0004984 olfactory receptor activity GO:0033155 oligogalacturonide transmembrane transporter activity GO:0015198 oligopeptide transporter activity GO:0015157 oligosaccharide transmembrane transporter activity GO:0015478 oligosaccharide transporting porin activity GO:0004924 oncostatin-M receptor activity GO:0005252 open rectifier potassium channel activity GO:0004985 opioid receptor activity GO:0001847 opsonin receptor activity GO:0016499 orexin receptor activity GO:0015274 organellar voltage-gated chloride channel activity GO:0005342 organic acid transmembrane transporter activity GO:0005343 organic acid:sodium symporter activity GO:0008514 organic anion transmembrane transporter activity GO:0015101 organic cation transmembrane transporter activity GO:0015548 organomercurial transporter activity GO:0015605 organophosphate ester transmembrane transporter activity GO:0015315 organophosphate:inorganic phosphate antiporter activity GO:0042978 ornithine decarboxylase activator activity GO:0008073 ornithine decarboxylase inhibitor activity GO:0042979 ornithine decarboxylase regulator activity GO:0005034 osmosensor activity GO:0015271 outward rectifier potassium channel activity GO:0019531 oxalate transmembrane transporter activity GO:0000227 oxaloacetate secondary active transmembrane transporter activity GO:0015131 oxaloacetate transmembrane transporter activity GO:0017077 oxidative phosphorylation uncoupler activity GO:0015453 oxidoreduction-driven active transmembrane transporter activity GO:0015367 oxoglutarate:malate antiporter activity GO:0005344 oxygen transporter activity GO:0004990 oxytocin receptor activity GO:0015569 p-aminobenzoyl-glutamate transmembrane transporter activity GO:0015558 p-aminobenzoyl-glutamate uptake transmembrane transporter activity GO:0015450 P-P-bond-hydrolysis-driven protein transmembrane transporter activity GO:0015405 P-P-bond-hydrolysis-driven transmembrane transporter activity GO:0001602 pancreatic poly