Experiment: materialize phylo-specific uberon classes as lattice

[cmungall]

For context see https://github.com/obophenotype/uberon/issues/1635

The goal here is to make species-specific views as well as a pan-taxon view with materialized species-specific classes

These would have IDs like UBERON:nnnnnn.9606 (possibly new namespace.. not relevant yet as this is an experiment).

• materialize the cross product of all Uberon x TaxonSubset (including intermediates, e.g Mammalia)
• filter unsatisfiable (i.e. human feather) analogous to #1635
• push down phyloGCIs
  • E.g. (C and part-of some T) sub R some D ==> CT sub R some D,
    • E.g. 'mammal naris part-of some mammal head`
• insert horizontal homology axioms, from both curated table and inferred
• merge with ssAOs if it exists (but note we will have nodes for ssAOs that do not exist). This is tricky as there is the possibility of bringing in things that mess with overall consistency and coherency, better if everyone is conforming to DPs etc

The results would be similar to the composite views https://github.com/obophenotype/uberon/wiki/Multi-species-composite-ontologies however, it would include taxa for which there is no curated ssAO.

The result would be very latticey which would confuse a lot of tools. We can set up a local OLS and test this as @matentzn did for upheno2.

We could also simply remove the taxon-generic groupings and have artefacts that are at a single taxonomic level

Same strategy could be applied to CL - this may be where we are headed with provisional CLs, transcriptomic-cluster CLs, marker-based CLs, all of which are rightly ss. cc @dosumis

[cmungall]

It looks like Wikidata are starting to materialize a similar lattice but it is v incomplete right now https://twitter.com/chrismungall/status/1453026180647383040

[cmungall]

Chatting with @fbastian this would be useful for bgee

what they currently do is take composite-metazoan and then reason over it each species S, injecting the only-in-taxon-X constraint and materializing the inferred results

[matentzn]

Talking to @cmungall right now: this may be less of a priority right now given resources, but may become much more interest when the Uberon grant comes along.

[gouttegd]

when the Uberon grant comes along.

So, never, then.