Add supporting study method types slot to association class

[kevinschaper]

Summary

• Add supporting study method types to the association class's slots list in biolink-model.yaml

The supporting study method types slot is already defined in the schema and inherits from supporting study metadata → association slot, but it was not explicitly declared in the association class's slots list. This adds it so that any association can explicitly declare supporting study method types.

Closes #1698

[sierra-moxon]

apologies - I am not sure here; should all of supporting_study_metadata descendents be available on association or just supporting_study_method_types?

Also - there is has_supporting_studies - which has a range Study -- should we be creating a has_supporting_studies inlined list of Study objects where the Study objects have supporting_study_metadata hierarchy of slots?

  supporting study metadata:
    description: >-
      Information about a **study** used to generate information used as evidence to support the knowledge expressed in an
      Association. In practice, data creators should use one of the more specific subtypes of this property.
      ....
     - Note that these concrete 'supporting study metadata' slots are used only when a more normalized model that leverages
        the 'supporting studies' slot and 'Study' class to link to and describe the Study itself are not possible or preferred,
        such that this study metadata must be captured directly on the edge.

I think we do create Studies in KGX for other ingests (see tmkp, etc.)

[kevinschaper]

I'm not sure - this is motivated by wanting to move curies for gene expression assay types out of 'qualifier' and into a better slot and @mbrush suggested supporting study method types

I wouldn't object to supporting a more detailed nested structure for supporting study metadata, but it's more than I need for the specific problem I'm aiming to solve

[sierra-moxon]

I think we should add assay or assayed_with or expression_assay to:

biolink-model/biolink-model.yaml

Line 10341 in 9e2c2dc

gene expression mixin:

with a supporting enumeration of assay types to keep the values simple for now. Later, we should discuss including a more rigorous assay class in Biolink. Some efforts towards this in the toxicology space are here: https://github.com/EHS-Data-Standards/outcomes-working-group/blob/e1fbe1feb9f077313e53693d0ae434692c6885cb/src/outcomes_working_group/schema/assay_base.yaml#L47 but would need to be significantly generalized to import directly here.

I think using supporting study method types causes a bit of inconsistency by not populating it in the Study object (of course slots can be re-used in many classes, but this one is named, defined, and currently only used in a Study object that is referenced on an edge instead of all of its properties being pushed up a level to the edge directly, and is also often used as a collection of studies, each of which contain this property individually) and also in the gene_expression_mixin which seems to serve the purpose of grouping attributes of an expression edge.