The Role of Machine Learning in Causation

There are a small number of companies that look at human clinical data and trying to discern patterns. They're mostly driven by redefined clinical hypotheses that a clinician has set the direction for. There's very few, if any, companies that actually try and kind of create that entire arc from cellular experiments through to the human clinical outcome. The fortunate matter for us here is that genetics gives us a fairly good proxy for causality because of the way in which genetics kind of lies upstream of pretty much all of the phenotypic consequences.