Identifying patient-specific root causes with the heteroscedastic noise model

Complex diseases are caused by a multitude of factors that may differ between patients even within the same diagnostic category. A few underlying root causes nevertheless initiate development disease each patient. We therefore focus on identifying patient-specific disease, which we equate to sample-specific predictivity exogenous error terms in structural equation model. generalize from linear setting heteroscedastic noise model where Y=m(X)+ɛσ(X) with non-linear functions m(X) and σ(X) representing conditional mean absolute deviation, respectively. This preserves identifiability but introduces non-trivial challenges require customized algorithm called Generalized Root Causal Inference (GRCI) extract correctly. GRCI recovers more accurately than existing alternatives.