Ontology meeting 2015-12-17
Attendees: Paola, Melanie, David OS, David H, Tanya, Harold, Chris, Heiko, Paul T
Minutes: Paola + Melanie
- 1 GH: Reinstate 'DNA supercoiling activity'
- 2 Redundant relations when cycling ontology file through OBO-Edit
- 3 GAF with inferred annotations from annotation extensions
- 4 Transitivity over 'regulates'
- 5 Term requests from from VU synapse folks, too general?
- 6 TermGenie Template for Protein Complex?
- 7 Report from the LEGO meeting in Geneva
- 8 Brief refresh on annotation policies
GH: Reinstate 'DNA supercoiling activity'
AI: Tanya will reinstate.
Redundant relations when cycling ontology file through OBO-Edit
This is still happening. Any news? (Previously discussed at http://wiki.geneontology.org/index.php/Ontology_meeting_2015-11-19#Redundant_relations_when_cycling_ontology_file_through_OBO-Edit)
UPDATE - seems fixed now.
20151217: MC: I had the issue recently via a TG template request.
Unclear: some people are still seeing these and they break the release. AI: if they show up again, let’s try leaving them and see what happens. Chris and Heiko will look into this. (Melanie's case seemed to be different, Heiko will investigate.)
GAF with inferred annotations from annotation extensions
Are these now being released. If so, where is the file and who, if anyone, is picking it up?
Heiko says this is not active yet.
DOS: What evidence code will be used? I Heiko: t’s IC, as decided by curators some time ago. Not everyone at this call is comfortable with it (IEA sounds better), but the objection is that IC is the meaning anyway. We may revisit the IC decision, but we are otherwise ready to turn the functionality on as soon as possible. AI: Heiko will do it today. Then we can see how it looks.
Transitivity over 'regulates'
(Chris) Please respond to the user.
The user has muddled transitivity of relations with grouping of annotations. AI: Chris will respond DOS: Don't we still need better doc on annotation grouping by edges other than is_a/part_of ?
RELATED ISSUE: Regulation 'algebra' property chains
Thought we had these in already. Seems not. Need to get a clear plan in place for this.
They’re not in, we had plans to but didn’t. The algebra property chains are not in GO nor RO. Property chains + and - =? - and - and - => + e.g. negatively regulates and negatively regulates => positively regulates)
AI: Chris will add the algebra chains. MC will create a ticket in the RO tracker - Done see https://github.com/oborel/obo-relations/issues/98
Potentially relevant example: https://github.com/geneontology/synapse/issues/41
Term requests from from VU synapse folks, too general?
- trans-synaptic signaling:** "Cell-cell signaling in either direction across the synaptic cleft."
- trans-synaptic signaling by lipid:** "Cell-cell signaling in either direction across the synaptic cleft, mediated by a lipid ligand."
Lack of direction means too vague?
Surely one would always know: whether the protein being annotated is pre or post synaptic; whether the function is in production/transport/secretion of ligand vs reception/transduction of signal. This is enough to infer direction (ignoring autocrine signaling).
We can see why they may want it, but we need to provide annotation guidance and suggest to use a more specific term whenever possible. AI: DOS will consult with curators to see how they’d want to use them. We need more curator input in general.
TermGenie Template for Protein Complex?
Petra wants also to create new complexes. She thinks a simple template would be nice. Thoughts?
We will simply add such terms upon request. Again, we are currently not accepting protein family binding terms; there is a warning in TG freeform to this extent.
Report from the LEGO meeting in Geneva
Those of us who weren't there would like a quick update please. If there isn't time, these documents are available:
- LEGO curation docs: https://docs.google.com/document/d/18ihslb7prB6CWtu2yjF-pMHZBTY1-AdXJAu-ZyuyXS4/edit#
- Videos https://vimeo.com/channels/Noctua
Short LEGO report
Good time for ontology group to start making generic models that annotators could flesh out “templates” someone could pull down into their own model, such as MAPK cascade. To which level of details to we go in terms of annotations? Eg transporters. Complicated for compound terms, and we still create a lot of those. => we need to have discussion boundary between ontology & LEGO templates
David: how does this all work with enrichment if we have only the models and not the term? Chris: level 1: we get a lossy gap with less specific terms (downside) level 2: smart: post composition where you take the LEGO model and use TG-like/pattern-like to create an analysis ontology level3: build new generation of term enrichment tool which makes use of the full axioms
Brief refresh on annotation policies
Experiments and papers that may not be in scope for GO?
Two issues cropped up recently, stemming from TG freeform requests. 1) We thought we agreed - and made it clear to curators - that terms describing completely non-physiological stimuli, such as exposure to microwave radiation, are not in scope for GO. But similar requests continue to arrive. We need to make sure we all agree on guidelines. 2) We thought that annotations from abstracts only (where the full text of the paper is in non-English language) were discouraged. Looks like this may not be the case for some MODs. But then, how can people verify, a posteriori, correctness of annotations? AI: DOS will go back to current curation guidelines and report. AI: Paola will ask Kimberly to bring this up at an annotation call.