Clarify schema for marker relations

[dosumis]

we have marker relations 2 right now:

• is marker for RO:0002607 (in Cell_Phenotype_KG_Schema_v2)
  http://purl.obolibrary.org/obo/RO_0015004 'has characterizing marker set' in CL_KG

Proposed solution: Reserve former for single genes, use latter for combos?

Also - I think these should really be Annotation Properties in OWL as ObjectProperty restrictions are inherited down the class hierarchy. Markers are not. Can we get them changed in RO?

[dillerm]

I reviewed this with @scheuerm and the main argument we have for keeping the transcripts as nodes is because the genes they're associated with are important connection points to drugs and diseases, which may be more easily handled if the transcripts are actual graph node. We also discussed the concern that you raised about inheritance and aren't too concerned about that being a problem, although I am curious about any examples that should give us concern.

[dosumis]

They will still be nodes in the graph. Using an AP to link two nodes (classes or individuals) just ensures we don't get unwanted inferences of markers down the graph/class-hierarchy.