Ontology meeting 2016-09-15

From GO Wiki
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Attendees: DavidH, DavidOS, Paola, Harold, Tanya, PaulT, Jim, Melanie

Regrets: Chris

Minutes: David H

How to report potential/necessary changes to annotations resulting from edits

There's been some renewed discussion of the need to be able to report annotation issues in one place.

Some ideas:

Can we get around the reluctance of some to use the go-annotation ticket tracker by use of GitHub CCs? Paola has done this quite effectively using the go-ontology tracker: This still requires knowing who to CC for each source though - we'd need documentation of that. Should we ask each group for a volunteer?

Alternatively, can we get at least one person from each group to sign up to a tracker with slightly less volume than the current one?

Side issue: It would help if we made sure everyone potentially has a GitHub account we know about. Not everyone is on here: https://github.com/orgs/geneontology/people?utf8=%E2%9C%93&query= Shall we put out a renewed call for people to sign up and share their logins?

Do we need a specific tracker for interpro mappings? We would have a general tracker for other issues. Paola likes keeping the issues on the ontology tracker. We shouldn't have to open a separate ticket on the annotation tracker and another issue at the specific help desk. It's extra work for us to know everyone that needs to be contacted for every different kind of issue. In github not everyone we need autocompletes, so their names never show up. We need to add them to the teams for this to happen. UniProt has a separate dedicated help desk. We need to find out who the UniProt contact would be with whom to coordinate. We should try to get a PI to encourage the curators from other resources to join the GitHub groups. We would set the options so that people would only get the emails from the tickets in which they were specifically mentioned. Do we want a separate low-volume tracker or do we want to keep everything in one gigantic place. Then we would mention all the appropriate groups on the item. We need a table of contacts for each group that includes GitHub ids.

Update: table is here https://docs.google.com/spreadsheets/d/1nAquF0oXHU6d638Q97ZeQ7K7DOhuru84-rOMEL4A114/edit#gid=0

Daily report of terms and stats

See http://wiki.geneontology.org/index.php/Ontology_meeting_2016-08-18#Daily_report_of_terms_and_stats

Chris, is there a quick fix to reinstate this please, or should a ticket be filed?

Ticket: https://github.com/geneontology/go-site/issues/217

The ticket says that this is fixed, but no one has seen the report.

Metadata update

Chris: general plan is to move everything to yaml including metadata. All agree that’s a good plan.

Need Chris for this one.

Modified proteins

There have recently been a few GH items about modified proteins particularly with respect to binding. We should reopen this discussion and formulate a proposal for the meeting at USC. [[1]] [[2]] [[3]]

Editors will further discuss the topic, and will then formulate a proposal to present to the wider group.
Preliminary plan:
   1. Merge all the terms for modified protein binding  (e.g. gylcosylated protein binding) into 'protein binding'
   2. Add comment to 'protein binding':   Please do not annotate to this term when it is known that binding is to a protein modification (e.g. to the oligosaccharide component of glycoprotein).   In these case, please instead annotate to the an appropriate term for the bound structure (e.g. polysaccharide binding)
Please contribute to and comment on the draft proposal here: https://docs.google.com/document/d/17U1TD3n_QI9coCycJcOUBDa2nRf3ggPR-tKjDMa_ff8/edit

Please comment on the proposal! Keep in mind that not everyone attends the annotation calls and not everyone will be at USC.

Protein complexes



  • macromolecular complex: update definition to say it contains at least one protein;
  • make ‘protein-containing complex’ a synonym of this term
  • protein complex: add user friendly comment along the lines of “these are complexes containing *only* proteins, if you are looking for a more general term please consider using the parent ‘macromolecular complex’”
  • Classification of complex by function: switch genus to 'macromolecular complex' (for example of why this is necessary, see: https://github.com/geneontology/go-ontology/issues/12620#issuecomment-247048464)

These edits have now been completed (DOS branch), but have not yet been committed. Are editors happy to commit ASAP?

Are we excluding non-coding RNAs here? No. The proposal is reasonable. We will wait to hear from Birgit, then David will go ahead with this plan.

We need to make sure that someone from Intact is aware of the discussion at the GO meeting regarding complexes. We need to have a transition plan.

FAO update

  • FAO now has a home on github
  • We have a partial axiomatization of some fungal CCs and BPs using FAO - now in import chain
  • See Inferences Changes Report for: 2016-08-24 - e.g.
+id: GO:0075251 ! uredospore formation
+is_a: GO:0044711 {is_inferred="true"} ! single-organism biosynthetic process


New high level terms for fungal processes (population of unicellular organisms vs. multicellular organism)

Take to GO annotation call or GO meeting?

NTR: population of unicellular-organism process terms [[4]]

Discussion is ongoing and recorded in the GH ticket above.

NEXT CALL first topic and Tanya will ask Diane to join us.

Implement a system for keeping partonomies in sync (Uberon & CC -> BP)


We have many cases where we manually maintain partonomies between Uberon and BP (most notably under the development branch, and between CC and BP (most notably in the organization branch). Ideally we'd have some way to keep these part hierarchies in sync automatically. Not having such a system causes problems and adds unnecessary work for editors.

We can solve the problem by using GCIs to infer partonomy

  • These GCIs can be incorporated into Design patterns, TermGenie etc.
  • GCIs are visible in Protege 5 under terms used in them
  • We have a system for generating inferred partonomies. This will allow an edit file pipeline analogous to the one we use to generate automated classifications at each commit and add them to an imported file so that they are visible in OE or Protege.

I propose we test this out on the organization branch.