Ontology meeting 2015-11-19

Attendees: Paola, PaulT, DavidH, Harold, Heiko, Chris, DavidOS, Tanya

Minutes: Tanya

Redundant relations when cycling ontology file through OBO-Edit

We've been fixing this manually so far. Any development?

 result of adding extended UBERON import chain, both zombie relations and cycles seem to stem from this
 pushing limits of what can be done with OE but Protege development has been slower than hoped
 revert to older UBERON without TCs?, try a different strategy for the paper in progress, but how to implement this alternate
 strategy before paper needs to be submitted
 inquire for later submission to JBMS?
 have TCs come in not as part of main UBERON import but as a separate job that runs on Jenkins and also ends up in GO plus?
 have the OBO version of the UBERON file be simple (won't break OE), but OWL version is used for Protege and for pipelines

 AI: Chris will change the pipeline so that OBO version of UBERON wouldn't have the TCs

Additional taxon constraints are now being inferred from Uberon

There was an issue because UBERON synonyms with extended characters were creating fatal errors and preventing us to commit taxon rules in OBO-Edit. Now fixed (see https://github.com/obophenotype/uberon/issues/1173) by changing the module extraction pipeline to not include synonyms (or definitions, for good measure) for the uberon import module. Chris, when you're back, could you please add the definitions back? Useful to have them there if possible.

 No guarantee that the defs won't have the extended characters. Safer not to do this.

Newly added constraints do not have PMIDs (or other types of references) associated with them. UBERON is usually quite rich in this kind of information. Would be possible to inherit PMIDs from UBERON constraints? Not a super-high priority, but useful - and in line with our approach to try to add references to rules whenever possible.

 UBERON inherited constraints are not visible without reference ids for Tony/QuickGO/Protein2GO. Should be there but not visible,
 could be fixed later. No GOTAXIDs.

Also, would it be possible to have an idea of how many annotations across all MODs have been flagged as wrong following the change?

 There is no single report for annotations across all MODs.

There may still be issues with this. Try clicking on odontogenesis in obo-edit, or search on a child term and have a look at the graph. There are lots of unknown self-referential (circular) relationships.

 Once pipeline is fixed, this behavior should disappear.

Follow-up: Adding logical defs to cellular response terms

Background: DOS has started to do this. What's the syntax, and should we look into filling all gaps.

Decision: All agree that "occurs_in some cell" is sufficient for cellular response to X terms. While this pattern wil not work for all 'cellular' terms, it is likely to be applicable in multiple branches.

AI: DOS to add 'response to' patterns to GitHub design pattern repo

Follow-up: where are we with this?

 Pattern for cellular response is on GitHub repo.  
 Bigger issue is whether "occurs_in some cell" (or something similar) is suitable for defining cellular processes in other branches.  
 A general pattern for all cellular process terms is not possible given the vague, granularity-based definition of GO: cellular process.

Follow-up on Inference management

Background: Documentation of ETINE flagging & relationship to redundancy stripping.

STILL TO DO: Document how ETINE flagging works :Chris & Heiko:

AI: DOS to try to track down examples of potential bugs found during previous discussion of ETINE results.

Problem: How to break down assessment of existing ETINEs? Need to work through all of these before moving to automated deletion. There are over 1500.

Follow-up: where are we with this?

 Down to 1460.
 Need ways to separate out all:  DOS has split into sets that could be treated similarly. Then much more tractable.
   chebi  Accept all?
   uberon Accept all?
   CL Accept all?
   single org  How much do we care?
   cellular  - Look into which branches might be formalised.

 Get rid of SOP and MOP?  Probably not.
 
 DOS will post files of grouped ETINEs. GOAL: end of January.  Then move to import of inference.

Follow-up on Design Patterns

Background: Next steps

Need to work on automating DP validation via Travis. Requires running Jython on Travis while pulling in relevant libs.

AI: DOS and Heiko to meet next week & screenshare to work through how to do this.

Follow-up: where are we with this?

 DOS and Heiko have worked together on the Travis/Jython issue and have made some progress.  
  Builds now work with OWLtools alone or Brain alone.  But differences in OWL-API version make it difficult 
 to combine the two whilst pulling from maven.  One solution would be to update Brain to the latest version of the OWL-API.
 AI: Heiko will look at updating BRAIN to current OWL API.
 AI: DOS + Chris + Heiko to have a call tomorrow to check in.  8:30 am /4:30 pm

Protege 5

Is everyone getting the 'array out of bounds' error?

Melanie: P5 always gives the error, I use P4 for editing.

Chris: has error been reported? DOS: think so.

GH + TG coverage

DOS will take over GH + TG next Wednesday from DH.

synapse check in

DOS going out to Amsterdam week after next. Private GH repository made under GO umbrella, issues being created for discussion purposes. The GH move should allow tracking. May need to consult with GO eds during this process.