Incorporate the Compound Function script in ontobio into the pipeline

[dougli1sqrd]

In Issue https://github.com/biolink/ontobio/issues/341 we made a script for producing enables o has_part -> enables annotations in ontobio. We ran the results in a specialized jenkins job and posted the results to skyhook. Details in the ticket link above.

These should be run in the pipeline on all sources.

The script returns a log and inferred annotations in this way.

We need:

• [ ] a location to put each file in the directory structure
• [ ] a place to run this from. (Makefile? Directly in Jenkinsfile?)

[kltm]

The question would seem to be whether or not this is an intermediate production for the annotation production. If so, I'd say they go on the pile in /products/annotation; if not, some new top-level directory in /products, TBD.

If it is this--https://github.com/biolink/ontobio/issues/341#issuecomment-499292690--but for all files, I would be tempted not to load any more into the Makefile, but to add it as a separately optimizable step that is making a distinct product.

[dougli1sqrd]

This is an earlier version of what would be added to the pipeline

sh 'materialize.py --log target/groups/mgi/mgi-inferences.log infer --ontology target/go-ontology.json --target target/groups/mgi/mgi.inferred.gaf --gaf target/groups/mgi/mgi.gaf'
sh 'materialize.py --log target/groups/pombase/pombase-inferences.log infer --ontology target/go-ontology.json --target target/groups/pombase/pombase.inferred.gaf --gaf target/groups/pombase/pombase.gaf'
sh 'materialize.py --log target/groups/wb/wb-inferences.log infer --ontology target/go-ontology.json --target target/groups/wb/wb.inferred.gaf --gaf target/groups/wb/wb.gaf'<br class="Apple-interchange-newline">

The way we'd want to do this moving forward would have changed. Adding here as the branch that this work was on as an experiment will be removed.

[pgaudet]

@kltm I'd like to propose to put this in a current project - this would address

https://github.com/geneontology/go-ontology/issues/19721 and generally, allow curators to sop making double annotations.

Thanks, Pascale

[kltm]

@pgaudet It depends on how you might want to look at this. IIRC, this might be a bigger ask as it was not as far along as we thought when work stopped on it. It has been a few years(!) since this was in the fire, so my memory is a little hazy. Either way, depending on priorities, this might be something that @dustine32 could help us out with in March. @pgaudet , how about if we touch bases on this on Monday?

[cmungall]

@pgaudet - I could have sworn we had a GO rule for this? I can't find it

We shouldn't be implementing this if there is no clearly documented rule

[pgaudet]

Sure I can make a rule, I didn't know that this is how we wanted to implemented this.

It would be nice to distinguish between an annontation rule (what is allowed and not allowed) versus post-processing we do on annotation, like interontology links and asserting has part annotations.

[pgaudet]

(@kltm I put the other ticket in the pipeline QC and maintenance - up to you where these go)

[kltm]

@pgaudet Is it a separate issue or related to this?

[kltm]

Noting creation of GORULE:0000062 for this https://github.com/geneontology/go-site/pull/1810

[kltm]

@dustine32 @pgaudet Okay, I think I have a handle on this. I think the things that need to be done are:

• [ ] dust off eric's work, add any needed tests, etc.
• [ ] make sure that this covers all the use cases of new GORULE:0000062 (above); i.e. David's and @cmungall 's formulationa from https://github.com/biolink/ontobio/issues/341#issue-446340056, https://github.com/geneontology/go-site/issues/109#issue-110412421 (well stated by @pgaudet 's new GORULE)
• [ ] wire into test pipeline as new temp "filter" step so that it can be applied to all files; run test pipeline
• [ ] when the results look good, merge into production pipelines

[kltm]

I think this may be out of scope for the "cleanup" project, so bumping over to QC for the time being.