Experiment with minimal/"GPAD" ShEx

[kltm]

Take the existing ShEX and use that as a starting point to decide by what criteria models would be able to be fully viewed (i.e. all information is displayed) in a standard annotation UI and thus marked with appropriate metadata.

The first pass of this is to use GPAD-acceptability as the baseline.

Once the proposed ShEX is available, run the proposed criteria against our existing models and create a public list for further exploration and discussion.

[kltm]

From morning Noctua workbenches call.

[vanaukenk]

Noting here some models that we are using to test the initial implementation:

Runx2 (mouse): http://noctua.geneontology.org/editor/graph/gomodel:MGI_MGI_99829 should have validation errors

Eed (mouse): http://noctua.geneontology.org/editor/graph/gomodel:MGI_MGI_95286 should validate

retina morphogenesis enabled by crb2a and crb2b Drer (zebrafish): http://noctua.geneontology.org/editor/graph/gomodel:65288b2200000125 should validate

kctd10 in heart development PMID:24430697 (zebrafish) http://noctua.geneontology.org/editor/graph/gomodel:56170d5200000012 should have validation errors

SPT6 promotes origin licensing through MCM helicase loading during initiation of DNA replication (budding yeast): http://noctua.geneontology.org/editor/graph/gomodel:641ce4dc00000384 should validate

CDC14 localization during the mitotic cell cycle (budding yeast):
http://noctua.geneontology.org/editor/graph/gomodel:6348a65d00002793 should validate, but need to allow the CC extension

lin-4 (WB:WBGene00002993) (worm):
http://noctua.geneontology.org/editor/graph/gomodel:WB_WBGene00002993 should validate

C. elegans mitochondrial UPR (worm):
http://noctua.geneontology.org/editor/graph/gomodel:5b528b1100001416 should have validation errors

[vanaukenk]

See also:
#307

[balhoff]

@vanaukenk I updated the ShEx file to allow primary input/output extensions. Should I allow direct connections from gene products to biological processes? Or do we always want at least a root MF?

[vanaukenk]

Should I allow direct connections from gene products to biological processes? Or do we always want at least a root MF?

From a GPAD perspective, we can definitely have direct connections from gene products to biological processes. The connection to a BP via MF is one of the main things we want to remove with the Noctua standard annotation UI.

Let me know if I've not understood something here.

Thx.

[balhoff]

@vanaukenk thanks, I allowed those. I wasn't sure if they would exist in the current models, but it's okay if they do. I have constrained things to disallow both styles simultaneously amongst the same nodes, though.

[vanaukenk]

Non-conformant models as of 2024-02-20