gouttegd
commented
11 months ago What I am considering to do (unless there is objection):
-
Remove the CL bridges to the life stage ontologies. This shouldn’t be controversial. Unless we start putting some life stage terms in CL (why whould we ever do that?), those bridges will always remain empty, so they serve no purpose at all.
-
Remove the bridges to ontologies that are deprecated (e.g., AAO, BILA, EHDAA, EMAP, EV, etc.). We may keep the existing cross-references on the Uberon/CL terms (though I’be keen to remove them as well), but they would no longer be used to generate bridge files.
-
I’d be inclined to do the same for ontologies that are inactive (e.g. AEO, MFA, OlatDv, etc.), but maybe this is going a bit too far.
-
For prefixes that are “unknown to the OBO Foundry” (e.g. ABA, BSD, OG, OGES, etc.): unsure. All those prefixes are also unknown to Bioregistry. If someone knows which resources these prefixes refer to, please share your knowledge and indicate whether you think having bridges with these resources is still relevant.
-
For the bridges where the source of truth is currently unknown (meaning we have the bridges, but no idea where they come from), I’d suggest moving the source of truth to cross-references in Uberon/CL, to align them with most of the other bridges. That would concern at least BFO, and possibly ABA, NIFSTD, VAO, KUPO (if we decide to keep the bridges with those).
matentzn
I will soon overhaul the way the bridges between Uberon/CL and foreign resources (mostly foreign ontologies but not only) are handled. Before I do that, here is an inventory of those bridges as they currently exist. Some of them may not be relevant anymore, so I’d like people to weigh in on whether they can be removed or not.
“Last updated” refers to the last time the bridge file had any changes committed to the repository (regardless of whether those changes were significant or not; for example, merely changing the version IRI without any edit to the bridging axioms still counts as an “update” and indicates the bridge file is actively maintained); when the last update is “ODK migration (2021)”, it means the history of the bridge cannot be traced back any further than PR #1798 because of the huge repository-wide changes in that PR – the real last update may be much older.
The “source of truth” indicates where the mappings that underlie the bridging axioms come from. A “?“ means that the source of truth is unknown: we have the bridge file, but we have no indication of how that file was generated in the first place.
The “bridge types” are:
equivalent: the mapped foreign term is said to be equivalent to the Uberon/CL term (FOREIGN:YYYY EquivalentTo: UBERON:XXXX);has-subclass: the mapped foreign term is said to be a subclass of the Uberon/CL term (FOREIGN:YYYY SubClassOf: UBERON:XXXX);is-a: the Uberon/CL term is said to be a subclass of the mapped foreign term (UBERON:XXXX SubClassOf: FOREIGN:YYYY);reverse genus differentia (NCBITaxon:ZZZZ): the mapped foreign term is said to be equivalent to the intersection of the Uberon/CL term and an existential restriction with the indicated taxon (FOREIGN:YYYY EquivalentTo: UBERON:XXXX and (BFO:0000050 some NCBITaxon:ZZZZ)in most cases, except when the Uberon term represents a life stage, in which case the property is BFO:0000066 instead of BFO:0000050);complex: custom-made bridge that does not obey simple rules like the previous ones.Uberon bridges
CL bridges
Of note, it is completely expected that the CL bridges with foreign ontologies such as FBdv, HsapDv, MFO, MmusDv, WBls, or ZFS are empty: those ontologies only contain terms representing life stages, and CL has no such terms (all terms representing life stages are in Uberon).