Developmental bridge files causing unsats

[shawntanzk]

Unsats caused by:
occurrent DisjointWith part_of some continuant

Follows the following pattern:
FBdv_00005336 EquivalentTo larval stage and instar larval stage and (part_of some Drosophila melanogaster)
HsapDv_0000006 EquivalentTo blastula stage and (part_of some Homo sapiens)
ZFS_0000044 EquivalentTo post-juvenile adult stage and (part_of some Danio)
XAO_1000001 EquivalentTo zygote stage and (part_of some Xenopus )
FMA_24120 EquivalentTo life cycle stage and (part_of some Homo sapiens)
MmusDv_0000017 EquivalentTo neurula stage and (part_of some Mus musculus)
WBls_0000093 EquivalentTo post-embryonic stage and (part_of some Caenorhabditis)

Potential fix:

1. part_of should be changed to in_taxon

Potential strategies:

1. change all bridge file intersection_of: part_of NCBITaxon: to intersection_of: in_taxon NCBITaxon:
2. change only developmental bridges
3. change only unsats

Related to #2367

[matentzn]

Share with Chris on slack, this blocks everything right now. I vote for 1, and Anita can do it.

[cmungall]

I don't understand 3, I think the choice is between 1 and 2

In OWL it's hard to have rules like "if X in-taxon T, and X is a continuant, then X part-of T". So implementing (1) could lose inferences. These will likely affect composite-metazoan.. but then we don't have composite metazoan anyway..

I think the best solution is to add two axioms

1. equivalentTo T and in-taxon T
2. subClassOf R some T, where R is determined by the category: part-of for continuants and occurs-in for processes

If this involves developing new code then go ahead and implement 1 for now and we will create an issue to return to make the more complete axioms

[shawntanzk]

I think the best solution is to add two axioms

Just to confirm does this mean something like:

HsapDv_0000006 EquivalentTo blastula stage and in_taxon some Homo sapiens
HsapDv_0000006 subClassOf occurs-in some Homo sapiens

or

MBA_710 EquivalentTo abducens nerve and in_taxon some Mus musculus
MBA_710 subClassOf part_of some Mus musculus

Where blastula stage is a process and abducens nerve is a continuant

[shawntanzk]

Note to self: this has implications on #2384

[shawntanzk]

Bridge process (to happen):
SSSOM will generate xrefs (for those that use SSSOM like fbbt)
xrefs generate bridge (ideally this step will be removed to instead use SSSOM to generate them)

To implement the solution above:

1. Look into the script (https://github.com/obophenotype/uberon/blob/master/src/scripts/make-bridge-ontologies-from-xrefs.pl) and see if we can go into it and change to above -> see if we can find if a namespace is a stage or a continuant and implement either occurs_in or part_of accordingly
2. replace this goal https://github.com/obophenotype/uberon/blob/master/src/ontology/uberon.Makefile#L1226

ALTERNATIVELY (quicker fix):
Fix after bridge is built to above with SPARQL

MAYBE BETTER SOLUTION:
Replace everything by a single SPARQL construct
how does it know the generate multiple bridge file -> specific thing in the script that breaks it up according to what the xrefs namespace

@ubyndr to look into these - we will decouple this from the release

[shawntanzk]

Fixed - though unsure if QC checks are reactivated if they were turned off - if so, open new ticket on it