Annotation Conf. Call July 9, 2013

Agenda

QC check for Enzyme binding

• We would like to propose that there shouldn't be any direct annotations to 'enzyme binding' term. We will add a 'not for direct manual annotation' tag to this term.

Protein oligomerization

Continuing our discussion on Protein oligomerization, we want to bring up some options to curtors.

• Protein di/tri/oligomerization terms: In many cases where the functional unit is a hetero or homo dimer or oligmer, the best option would be to use col-16 to indicate the functional unit using a PRO ID (or other type of complex ID).
• In cases where geneproduct A is involved in the oligomerization of another gene product B, then the appropriate term would protein complex assembly and indicate the complex in col-16.

Protein binding annotations

• Mapping protein binding terms to finer terms. This idea was discussed last year - http://gocwiki.geneontology.org/index.php/Improving_protein_binding_annotations_using_InterPro_domains
  • We can look into implementing this. We could provide a mapping file for groups to update their annotations.
    • May have to look at cases where there is more than one more granular term to map to

• Moving protein binding to its own file. Chris came up with the idea of keeping the protein binding terms in the GO, but pulling all protein binding annotations from various curation streams into one GAF file. This seems like a good solution that will make both parties happy.

Discussion

• Enzyme Binding check:
  • Ruth- How specific should the enzyme binding term be? i.e. is it okay to annotate to ligase as opposed to a specific ligase.
  • Rama- general guideline should be to annotate to specific term. If the authors show what type of ligase the gp bound to, then the term should be specific to what was reported.
  • While requesting new enzyme binding terms, make sure the specific enzyme activity term exists to begin with. May be a TermGenie template can be made.
  • We don't want to add enzyme binding terms for all enzyme activity terms proactively. We will add them on a need to have basis.

• Protein oligomerization terms
  • Some people find these annotations useful, some don't. GOC should come up with a policy to make sure everybody is doing it consistently.
  • One option would be to make complex terms for these dimers/oligomers and then start annotating complexes since the complex is the functional unit
  • Since there are several curation and publishing policies may be the best solution would be to be all inclusive which lead to the next agenda item.

• Consolidating Protein binding curation
  • There are several curation pipelines to curate interactions. It doens't make sense to duplicate efforts and come up with rules for when to annotate protein binding
  • IntAct captures self-binding (dimerization etc)
  • Consolidate all interaction data (BioGRID, IntACT, ?) into a single GAF file.
  • Look into mapping some of the protein binding annotations to more specific binding terms (rules need to be worked out)
  • What does this mean for GO curators? Answer: You don't have to capture protein binding using GO.
  • We will circulate a proposal for all of these.