[Ontology-editors] haspart documentation (meeting with Chris (fwd))

[Karen Christie]

Hi,

I was busy doing final proofs on a GO paper for Trends in Micro yesterday, 
so didn't have time to read Amelia's documentation or comment on this 
thread yesterday, but I wanted to add a couple comments today.

For starters, the documentation is *really nice*. I have a few specific 
suggestions and might be able to contribute ideas to Midori's suggestions 
for examples on the has_part documentation.

Regarding terminology, I personally like using parent/child. I think it is 
very intuitive with respect to understanding what Amelia says in the 
documentation: "parent refers to the [broader] term closer to the root 
node, and child to the more specific term closer to the leaf nodes".

I don't think incoming/outcoming is a substitute for parent/child. 
Parent/child gives information about whether the terms are more general 
(closer to root) vs more specific (further from root) while 
incoming/outcoming gives information about the directionality of the 
relationship. Until has_part, it looks like all the relationships used in 
GO are child -> parent in direction, but has_part is parent -> child. 
Anyway, I think I'd be OK with using incoming/outcoming in addition to 
parent/child, but not as a substitute since these two sets of words do not 
provide the same information.

So, with that said, below are my more specific comments on the 
documentation.

-Karen

1. I am completely in agreement with the current convention of always 
showing the parent on the top or on the left.

2. I think I might add in the conventions, that the arrows show the 
directionality of the relationship, mostly because it provides the 
opportunity to say early on that relationships can go either from the 
child to the parent or vice verso from the parent to the child.

3. With the addition of the has_part relationship, it might be good to 
insert a small tidbit to the first sentence of the part_of info, so that 
it would read something like:

"The relation part of is used to represent part-whole relations in the 
Gene Ontology, from the perspective of the child term."

4. In the diagrams, I think it would really help me if the term bubbles 
were labeled with letters 'A', 'B', 'C', etc. to correspond with the text, 
in addition to the names of the actual terms used, e.g. mitochondrion. 
Especially since most of the relationships are converted to sentences in 
the child -> parent direction, it means that you need to look at the 
diagram from right to left, rather than from left to right. I realize the 
arrows show directionality too, but when I'm trying to correlate text 
like:

"if term A is a term B, and term B is a term C, term A is a term C"

to the diagram, it would really help if I can visually see which term is 
term A, etc.




On Wed, 17 Jun 2009, Jane Lomax wrote:

> As an aside, ChEBI also had the parent/child nomenclature issue, and they now 
> use "incoming" "outgoing" instead e.g.
>
> http://www.ebi.ac.uk/chebi/searchId.do;9F09D4D09954F8AFEDE8721A3ECD91C0?chebiId=CHEBI:32816
>
> Although not sure this is a great solution either.
>
> I think the documentation looks great btw!
>
> Jane
>
> Amelia Ireland wrote:
>> Hi again all,
>> 
>> re: comments on ordering the terms in the diagrams: I can see both sides of 
>> the argument for the order in which the terms are displayed, but I think 
>> that for the sake of clarity and consistency, I am going to going to stick 
>> to the following conventions:
>> 
>> for vertical diagrams, the root nodes will be at the top, and the leaf 
>> nodes at the bottom (as in AmiGO and OBO-Edit)
>> 
>> for horizontal diagrams, the root nodes will be on the left, and the leaf 
>> nodes on the right (as in AmiGO's browse mode, where the child terms are 
>> indented with respect to the parents).
>> 
>> The arrowheads provide a visual cue as to whether the relation goes parent 
>> --> child or vice versa.
>> 
>> 
>> re: has part examples... our example for "has part" from ye olde 
>> documentation is
>>
>>     nucleus has part chromosome
>> 
>> Is this definitely, unequivocally, 100% true? When I did the chromosome / 
>> chromatin example, I was trying to find something that all chromosomes 
>> have, but it seems that just about everything is expendable! Maybe 
>> chromosome has part DNA would be easiest...
>> 
>> 
>> Tanya said:
>>> I have a question about the examples in general. Do we want them to 
>>> reflect the GO as it stands or not?  For example, the very first example 
>>> on one term having two different relationships to two different parents 
>>> isn't quite right, if we look at the ontology literally.
>> [snip]
>> I went for an example that was relatively easy for anyone with basic 
>> biological knowledge to understand. I think in general it's probably better 
>> to go for clarity rather than fidelity to the exact ontology structure 
>> (especially given that the structure changes over time). I also wanted to 
>> try to avoid the "artificial" GO terms like 'cytoplasmic part'... at least 
>> for the first diagram!
>> 
>> 
>> I'll do the various suggested minor edits and send out a new version later 
>> on.
>> 
>> Cheers!
>> 
>> -- 
>> Amelia Ireland
>> GO Editorial Office
>> http://www.berkeleybop.org || http://www.ebi.ac.uk
>> BBOP Plant Project: http://bbopgarden.blogspot.com
>> 
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> 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
>
> _______________________________________________
> Ontology-editors mailing list
> Ontology-editors at geneontology.org
> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>