Precompute inferred taxon constraints and add to ontology release

[balhoff]

No description provided.

[cmungall]

We can use this to show TCs in amigo, see:

geneontology/amigo#56 geneontology/amigo#58

The requirements are that a class should have a direct lookup to non-redundant only-in and never-in.

E.g for cameraa-type-eye-development we may have

• only-in vertebrata, from uberon:camera-type-eye
• only-in metazoa from uberon:generic-eye

We only want to show the more specific one.

similarly never-ins should only be added if they add info on top of the MRCA of only-ins. Note that redundancy checking for never-ins is inverted

Note: this would also be useful for the ontology group, so they can see in Protege the TCs

Things to decide

• do using OWL reasoning/materialization vs graph walking
• should this go in robot or a separate util?
• note for robot, it would be nice if these would show up when we hav inferences in exports, see Add INFERRED as relationship column qualifier for robot export ontodev/robot#697
• would this be inserted back as a regular owl axiom? What provenance should we add on the axiom? Ideally we could inherit e.g. PMIDs from Uberon TCs but we can tackle this incrementally

[vanaukenk]

See also: geneontology/noctua#261
geneontology/noctua-form#100

for reference to using this information in Noctua either solely for information purposes, or perhaps eventually to prioritize or restrict terms based on the taxon of an annotation entity.

[pgaudet]

Hi @balhoff

just wondering where this stands ?

[balhoff]

I haven't made any progress here; I focused on the Protégé plugin first.

[pgaudet]

No problem, just wanted to check.

Thanks, Pascale

[pgaudet]

I think this is done @balhoff ?

[ValWood]

@balhoff ?

[balhoff]

This is not done. I can try to prioritize this in the next few weeks.

[ValWood]

OK, should I leave taxon constraint tickets until this is completed?

[balhoff]

No, this is just a rendering of the logical consequences of all the taxon constraints, for external folks who aren't using an OWL reasoner. So if you add more taxon constraints now, they will still operate within our reasoner-based QC.

[pgaudet]

Thanks, I've added this to the next ontology call.

[kltm]

@pgaudet Unmoved in "PLANNING Phylogeny-based Taxon Constraints"; In Progess in "Ontology technical essentials". Issues should belong in one project, so removing mention in the one that's not being used.

[balhoff]

Done in #26180.