Multiple unsatisfiable classes resulting from uberon import chain

[cmungall]

go-plus became incoherent, many explanations like this:

[cmungall]

Diagnosis: uberon assumes 'reproductive structure' is metazoan (inferred from the assumption that the root class is metazoan); GO uses this also for plants

Long term fix: EITHER uberon does not bake in metazoan assumptions (effectively taking over some of the role of CARO) or we put functional groupings like this in CARO (as @dosumis did originally in FUNCARO)

Short term fix: remove uberon taxon constraints from import module

Note: the short term fix was already queued in the Makefile. However, it the rebuild of uberon_import.owl never happened due to the unsatisfiability. Race conditions in ontologies, fun. Some of this will be fixed by a clearer separation of edit vs public artefacts with #12211

[dosumis]

The explanation above shows inconsistency because of the mesoderm thing. I tried to fix that with an edit to uberon (it's not even true for metazoans), but looks like that didn't make it into a release yet.

If there is also inconsistency through 'reproductive structure' in Uberon, then maybe we should just remove the GO link to Uberon reproductive structure for now. Presumably could be functionally defined using GO. As you say, this could live in CARO rather than Uberon to keep it metazoan only.

[cmungall]

On 4 Jan 2016, at 13:37, David Osumi-Sutherland wrote:

The explanation above shows inconsistency because of the mesoderm thing.

Ah yes, you're right

I tried to fix that with an edit to uberon (it's not even true for metazoans), but looks like that didn't make it into a release yet.

will be in next one

If there is also inconsistency through 'reproductive structure' in Uberon, then maybe we should just remove the GO link to Uberon reproductive structure for now. Presumably could be functionally defined using GO. As you say, this could live in CARO rather than Uberon to keep it metazoan only.

there is less of a commitment issue than I thought. There is no problem with reproductive structure with your fix.

However, I will progress with the short term fix for now until the new uberon release is propagated, hopefully later today

---

Reply to this email directly or view it on GitHub: https://github.com/geneontology/go-ontology/issues/12223#issuecomment-168817594

[cmungall]

short term fix committed with r30654

Leaving this ticket open until new uberon release out

This eliminates the majority of unsats. Will file with remaining ones

[cmungall]

back on: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/ontology/extensions/Makefile?r1=32321&r2=32925

[dosumis]

To fix:

Possible fixes:

1. 'cardiocyte differentiation' 'part of' some 'heart development' -> 'cardiocyte differentiation' 'part of' some 'primary circulatory organ development' (GO, TBA?)
2. Relax constraints on heart in Uberon.

EDIT FROM CMUNGALL - see https://github.com/obophenotype/uberon/issues/1217

[dosumis]

Fix: Need separate terms in Uberon for 'vertebrate trachea' and arthropod trachea - both under 'respiratory tube'

EDIT FROM CMUNGALL

is the following sufficient for GO axiomatization purposes? UBERON:0005155 ! open tracheal system

[dosumis]

post-anal tail: "A tail that extends from the posterior tip of the organism to the anus, contains muscle and skeleton."

The skeleton clause excludes nematodes. Should 'post-anal tail' be broadened, or do we need a new term?

EDIT FROM CMUNGALL

See https://github.com/obophenotype/uberon/issues/1188

However, having 'generic tail tip' 'generic post anal tail tip' and 'chordate post anal tail til' just seems like propagating unneccessary groupings

Furthermore, it's not required here. The GO axiom is obviously wrong. If fixed it: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/ontology/editors/gene_ontology_write.obo?r1=r33368&r2=r33369

[dosumis]

ecotdermal placodes are common in developing arthropods. Remove this taxon restriction from Uberon.

EDIT FROM CMUNGALL

That would probably be inconsistent with how these structure are defined in uberon, let's follow up here: https://github.com/obophenotype/uberon/issues/135

UPDATE DOSUMIS 160802

Now seems to be fixed: https://github.com/obophenotype/uberon/issues/135#issuecomment-220793379

[dosumis]

Uberon needs a term for arthropod spiracle. - perhaps 'spiracle of open tracheal system' ?

EDIT FROM CMUNGALL

I'm just deleting the axiom for now (r33388). There are probably quite a few unaxiomatized development terms that are insect specific. So far we've been a bit hesitant about insect-specific classes in uberon, we should visit this in a separate ticket

[dosumis]

Unsatisfiable class module

@cmungall - Safe to turn import of constraints off until these are fixed. I think this is important to do ASAP as these won't get fixed straight away: TG is now blocked; I also urgently need an OBO version with recent commits in order to generate up-to-date doc for the synapse project.

[cmungall]

Makefile reverted with r32951

I can regenerate uberon_imports too

I recognize many of the issues above, will provide feedback shortly

[cmungall]

r32952 committed new imports module, minus TCs. GO should now be satisfiable

[tberardini]

Is GO satisfiable after changes made in r32952? If yes, then we can close the related issues, correct?

[dosumis]

epithelial cell fate determination, open tracheal system

Bit tangled - Uberon & CL - various granularities.

Problematic EC def:

'cell fate determination' and ('results in determination of' some 'tracheal epithelial cell')

From CL:

'tracheal epithelial cell' Equivalent to: epithelial cell and part of some trachea comment: This class is for the vertebrate tracheal structure. For the analagous insect cell type, see 'respiratory tube epithelial cell'

respiratory epithelial cell Equivalent to: endo-epithelial cell and part of some respiratory tract

This looks way to general, and should probably be defined functionally.

FBbt has tracheocyte: 'epithelial cell' 'part of' some 'tracheal system' (Might be a bit too broad still.)

Solution: add tracheocyte to CL (Can be done straight away, but some work needed to make appropriate terms in Uberon to wire up properly)?

But - this is sometimes used in vertebrates too. e.g. http://www.ncbi.nlm.nih.gov/pubmed/6149692

And is an exact synonym for 'tracheal epithelial cell' in CL.

Uberon has 'open tracheal system trachea'

So, we should just make CL term 'open tracheal system tracheocyte' - 'epithelial cell' that part_of some 'open tracheal system trachea'.

[dosumis]

@cmungall: should now be all fixed. Please try turning taxon constraint import back on. Ta.

[tberardini]

@cmungall - reminder. Thanks

[cmungall]

r35423

Let's see what happens tonight...

[dosumis]

Looks like there’s one remaining Uberon issue to fix. We missed this because this class was inconsistent for two separate reasons and we’ve only fixed one.

• [x] Quick fix. Delete this, it's wrong: 'optic lobe placode formation' part_of some 'optic placode formation'. This fixes the inconsistency.

Explanation: The optic placode referred to is the eye primordium in chordates. The optic lobe placode is the precurosor of the optic lobes - part of the brain in arthropods. In holometabolous insects at least, the eye develops from an imaginal disc formed by invagination of the embryonic epithelium not an ectdermal placode.

[cmungall]

good catch, I agree with the fix.

At some point we should go through and improve the labels and the definitions in GO. Everyone interested should contribute to the review of the placode hierarchy here: https://github.com/obophenotype/uberon/issues/135

[dosumis]

FIxed.