Ontology meeting 2014-06-05
Attendees: Paola, David OS, David H, Heiko, Tanya, Harold, Paul, Chris, Judy, Jane
- 1 Setting priorities for ontology group
- 2 MF overhaul
- 3 SF tidbits
- 4 Adding binding terms defined by activity of the enzyme they bind
- 5 InterPro in GO (related to above)
- 6 CHEBI
- 7 Prefixing 'obsolete' to all obsolete classes
- 8 Change relation for metabolism terms
- 9 Links to annotation guidance
- 10 Follow-ups from last week
Setting priorities for ontology group
As we didn't get a chance to do this yet we should discuss our priorities for the next quarter. The quarterly call for all GOC where we report on this is on June 18.
Link to wiki for progress reports for 2nd quarter:
Agreed on the following plan, as reported in the google doc above (refer to that for details):
Report of Priorities for Current Quarter (March, April, May):
- term requests, Term Genie and SourceForge metrics
- Ontology Developers moving to OWL
Report on Priorities for Next Quarter (June, July, August):
- Transition to Protege
- Assessment of MF refactoring
AI for all managers/groups: Put down FTEs for every quarter.
OWL file now on Web Protege:
Please sign up to WebProtege and tell me your account name. I'll add you to the editor list.
Reminder: I've also put Paul's justification and prioritisation doc onto google docs. Again, anyone with the link can comment:
Paul is reviewing David OS' comments and finding them useful.
AI for all: sign up to WebProtege and tell DOS your account name, so we can add comments AND edit the ontology; all have a try.
Judy: MF overhaul is not a done deal. Not saying it's bad, but we're still looking it over.
David OS: We need to do some explanatory work.
Judy: agreed. This could be a goal for the next quarter: requirements and value
Jane: for this quarter we'll just do an assessment
All: agreed. Then we'll present at the October GOC mtg.
obsoleting homo/heterodimerization terms, who remembers? SF10909
- We did not reach an agreement or resolution, and we're not going to do it now, sorry.
upper level regulation terms SF10900
- David H will add 'positive biological regulation' and 'negative biological regulation', and will make them 'do not annotate'. But will not add activation/inhibition terms (could make them narrow synonyms, but that may not help with text searches?)
obsoleting term based on retracted paper SF10892
- Agreed that: Annotations should be removed if paper is retracted - no longer PMID to associate with. Similarly if paper which is the sole primary reference for a term is retracted, the term should be obsoleted. The retraction must be noted in a comment on the obsoleted terms.
Adding binding terms defined by activity of the enzyme they bind
Should we be doing this?
- All agreed: No. And lets obsolete the ones we already have. Should be using domain/family binding terms. See next item.
- David H Example of why these terms are especially bad for the more abstract cases:
- Everything that binds to a receptor tyrosine Kinase - ligand, dimerising receptor, effectors - are all kinase binding.
- Solution to above ticket: protein binding - IPI to what. Then make annotation extension to express context.
For example protein domain binding terms. Chris has made a set of test mappings and equivalence axioms:
Shall we make it a priority to implement these?
- Paul suggests - best to only use InterPro families that also have a Pfam entry.
- AI: Make adding templates and design patterns for protein families and protein domains a priority
- starting with domains, and targeting proteins that are not already in GO + checking those already in domo
Only got as far as here - Everything below bumped to next week
(Stemming from the 'secondary metabolite' merge issue) Jane to report from the CHEBI workshop.
Prefixing 'obsolete' to all obsolete classes
Change relation for metabolism terms
Links to annotation guidance
Chris added 'curator guidance link' RO_0002484.
To discuss today:
- how should we record the propagation rule (I,P,R) - axiom annotation
- using axioms rather than regexps
- consistent location of docs: go wiki, drupal, or go nuts?