Rule stability + minimum gene set notebook

[envest]

We want to examine how stable the predicted rules are between repeats (using baseline kTSP or RF models, maybe lasso).

Given those rule sets, what rules (and genes) show up consistently? Can that gene set be used as a minimum model? How does it compare to random gene sets of the same size (negative control)?