[Ontology-editors] [Go-fp] notes from MF-BP skype (Re: MF-BP links & managers call)

[Chris Mungall]

On Aug 18, 2008, at 10:46 AM, Harold Drabkin wrote:

> Just saw this;
>
> I''ve asked Jen where I need to put some new files. I've been  
> working on and off on various "simple" metabolic pathways and have  
> noted that there will be instances where the occurance of the  
> functions in a pathway is organism specific. That is, for example,  
> with the glyoxylate pathway:: the pathway is used by bacterial and  
> plants, not animals, yet it does in fact use several TCA cycle  
> enzymes. The "has_part" is fine of course, but
> is means that a person might get the wrong impression when the  
> ontology is searched. Thus, an annotation to a mamalian TCA cycle  
> enzyme might be returned if one searched on glyoxylate pathway,  
> unless there is a way to restrict the relationship searches.

There should be no danger of people getting the wrong impression if  
software is implemented correctly. Stripping out has_part would work,  
but we really need to start educating people to think in terms of  
relation composition:

http://wiki.geneontology.org/index.php/Relation_composition
http://wiki.geneontology.org/index.php/Has_part

Simply not providing a composition for has_function_in has the same  
effect as stripping has_part as far as annotation searches are  
concerned.

We have the option of naming this composition - something like  
"functions_in_same_pathway_as"...?

>
> What I really do not want to see is a separate MF entry to the  
> enzymes that are involved in glyoxylate path, and the ones that don't.
> ie
>
> aconitase activity involved in TCA
> aconitase activity involved in glyoxylate
>
> it's the same activity.
>
> Note this is sort of the opposite of two organisms doing the same  
> thing via different processes.
>
> Harold
>
>
> Midori Harris wrote:
>> <div class="moz-text-flowed" style="font-family: -moz-fixed">For  
>> that we probably want a nupdate from Harold. Is he on this list?
>>
>> m
>>
>> On Fri, 15 Aug 2008, Chris Mungall wrote:
>>
>>>
>>> Ah, I guess my question was about what's happening with F-P links  
>>> in general, specifically the low hanging fruit part (there is the  
>>> P->F regulation terms which are really LHF of course, but there  
>>> must be more LHF..)
>>>
>>> On Aug 15, 2008, at 2:20 AM, Jennifer Deegan (nee Clark) wrote:
>>>
>>>> Hi Chris,
>>>>
>>>> The next day we discussed this on the managers' call. It was  
>>>> decided that when there were two species that did the same  
>>>> process differently we should make two different process terms  
>>>> with informative non-taxon names and put the appropriate  
>>>> functions underneath.
>>>>
>>>> I have started discussing this with the electron transport  
>>>> working group and it is being met with strong opinions, both for  
>>>> and against. We haven't actually made any new terms to test it,  
>>>> but I am hoping to have an example put together to present at the  
>>>> meeting.
>>>>
>>>> Jen
>>>>
>>>>
>>>> Chris Mungall wrote:
>>>>
>>>>> Just wondering where we were with this - last commit in the fp- 
>>>>> links directory is April...
>>>>> On Jun 17, 2008, at 10:42 AM, Midori Harris wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Minutes from today's impromptu Skpe conference, such as they  
>>>>>> are ...
>>>>>>
>>>>>> Present: Midori, Chris, David, Tanya, Pascale, Harold, Amelia
>>>>>>
>>>>>> Conclusions:
>>>>>> - None of us could see a big problem with the MF-BP approach;  
>>>>>> we  think Jen's concerrns reflect the fact that electron  
>>>>>> transport is a  fairly thorny area of biochemistry, rather than  
>>>>>> a general problem  with making MF-BP links.
>>>>>>
>>>>>> - We want to continue capturing MF-BP links essentially as  
>>>>>> Harold  and Jen have started. Go for low-hanging fruit --  
>>>>>> emphasize  metabolic pathways for now, because there should be  
>>>>>> a lot of  straightforward links there.
>>>>>>
>>>>>> - Don't worry about the occasional "error" of asserting a link  
>>>>>> that  may not hold in some obscure species. The people working  
>>>>>> on the  oddballs are, by and large, the same groups who most  
>>>>>> want GO to have  the MF-BP links (e.g. JGI, TIFKAT*, other  
>>>>>> groups working on proks).  They'll tell us if we get it wrong,  
>>>>>> and then we can fix it.
>>>>>>
>>>>>> - Also continue looking at Reactome for links. Show our links  
>>>>>> to Reactome curators (Peter d'E, and maybe others).
>>>>>>
>>>>>> - Keep collecting links in tab-delimited format for now (Chris   
>>>>>> parses into OBO format). Time frame for asserting links  
>>>>>> directly in  OBO format: weeks. Time frame for including links  
>>>>>> in the ontology:  months.
>>>>>>
>>>>>> - Present this summary at tomorrow's managers' call.
>>>>>>
>>>>>> - We had fun!
>>>>>>
>>>>>> Additions, corrections, comments, etc. welcome (not to mention  
>>>>>> charitable contributions). Also, feel free to plunk this on a  
>>>>>> wiki  or tell me where to do so.
>>>>>>
>>>>>> thanks!
>>>>>> m
>>>>>>
>>>>>> *The Institute Formerly Know As TIGR
>>>>>>
>>>>> _______________________________________________
>>>>> Go-fp mailing list
>>>>> Go-fp at geneontology.org
>>>>> http://fafner.stanford.edu/mailman/listinfo/go-fp
>>>>
>>>> -- 
>>>> Jennifer Deegan (nee Clark)
>>>> EMBL-European Bioinformatics Institute
>>>> Gene Ontology Consortium
>>>>
>>>
>>> _______________________________________________
>>> 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
>> </div>
>
>