unsatisfiable class: lymph node marginal reticular cell is part_of an immaterial entity

[balhoff]

This came up in a GO issue (there is an explanation there): geneontology/go-ontology#23510 (comment)

The Ubergraph pre-release coherency check missed this because it turns out there are a few CARO-to-BFO bridge axioms found only in an RO import. I've since added this import to the check: https://github.com/INCATools/ubergraph/runs/6926464224?check_suite_focus=true

[shawntanzk]

lymph node marginal reticular cell SubClassOf part of some subcapsular sinus of lymph node
subcapsular sinus of lymph node EquivalentTo anatomical space and (part of some lymph node) and (adjacent to some cortex of lymph node) and (adjacent to some capsule of lymph node)

This is the issue I think - instead of part_of sinus, perhaps adjacent_to? -> worry being that adjacent_to is subclass of spatially disjoing from and i don't know if ppl will expect that cell when searching for cells of the subcapsular sinus of lymph node

[matentzn]

@shawntanzk can we hotfix this now in whatever way, I can run a new release and we make a new issue to figure out the correct modelling? All worries should be dwarfed by unsats.

Since you define the sinus of lymph node as immaterial, it does not make sense to say its part of something material. It could be contained in it, or adjacent to it, or located in probably?

[shawntanzk]

I'll change it to adjacent to for now -> located in wont work due to range that is imported into CL

'independent continuant'
 and (not ('spatial region'))

[dosumis]

The correct relation is located_in (we need to add this here https://github.com/obophenotype/cell-ontology/blob/master/docs/relations_guide.md#recording-location). We use this in multiple places for cells --> spaces especially sinuses. The range is a confusing legacy of obsolete BFO concepts and should be fixed in RO. It has no logical consequences under ELK reasoning.

Please change from adjacent_to ASAP, as this will cause fails for HubMap and scAtlas.