Add a gene-disease edge to rules?

[cbizon]

We have a number of rules that are of the form (chemical)<-(gene)->(chemical)-treats->(disease).

There's a disease and it's treated by a chemical and some gene has a common effect on these things. So given that commonality, maybe the first chemical also treats the disease.

These rules are found by the rule engine, they have some predictive power but

1. the power is not high
2. They're not mechanistic for some users

I think we can improve these rules in both ways by adding an edge from the gene to the disease. I think it can be a related to.

NCATSTranslator/Feedback#465

[capasfield]

Chris wants to test.