Request for 'target for' predicate between gene and disease

[sierra-moxon]

we have "gene to disease association" but it doesn't restrict down to a certain predicate. With the definition:

gene in which variation is correlated with the disease,
may be protective or causative or associative, or as a model

From @cbizon : the idea of being a target is more specific: that in addition to having a genetic association, it's also a druggable protein and that a drug interaction with the gene could have a therapeutic effect.
The context here is in terms of the more high-level questions that NCATS wants to enable ARAs to work on. "Suggest gene targets for a given disease" is one of the questions. We can imagine writing queries to try to infer the Protein/Disease edge, but no way to encode the edge directly. IMO, this is something that we should add to biolink as a predicate. It's a common description in the drug development community

after talking with Chris, we came up with a short term proposal:
add 'target for' predicate between gene and disease.
add a child association 'gene is target of disease association' of parent 'gene to disease association'
add enumerated list of target detail categories as a slot on 'gene is target of disease association' -- using pharos as a guide here, but DrugBank also has this information.

a longer term proposal:
take this use case to predicates and make sure this fits the qualifier model.