Open hkir-dev opened 1 day ago
some may be related with https://github.com/brain-bican/human_brain_atlas_ontology/issues/2
No time to look in detail now, but I would expect these to be showing up in the MBA build
Nothing to do with cell types etc being added.
(Also - PNS CNS subdivision in Uberon is a perenial pain in the ass)
Didn't we attempt to fix these before by classifying as nerve roots?
It was only done for HBA terms https://github.com/obophenotype/uberon/pull/3281
Odd that MBA ODK passes.
Re fix: are the relationships between MBA terms coming from LinkML-OWL? If so trying to think through options to over-ride. A table or tables of curated relationships --> Slightly hacky curated delete and replace using SPARQL update following LinkML-OWL build?
MBA ODK is an older version (v1.4, latest is v1.6), so probably it doesn't have those checks.
Yes, parent_structure_id
property in the structure graph is used to drive the hierarchy between MBA terms using LinkML. But we have a python script that drive this operation, so we can over-ride some relations if needed using a curation table.
WMBO build raised 11 unsatisfiable classes:
Explanation of unsats: wmbo_explanation.md
Can you recommend a solution to fix the unsatisfiable classes?