A Generalized Partial Canonical Correlation Model to Measure Contribution of Individual Drug Features Toward Side Effects Prediction

摘要: Identification of potential drug-side effects is an open problem importance for drug development. Side are related to a variety interlinked aspects such as chemical properties drugs, drug–target interactions, pathways involved, and many more. Existing statistical methods machine learning models toward creating that incorporate features predict adverse reactions. One the challenges in these efforts disentangle interdependence identify contribution individual specifying side effects. We present partial canonical correlation analysis (PCCA) model facilitates enumeration from prediction class effects, irrespective on other features. The combination analytical numerical strategies, can be used arrive at most effective set starting range available descriptors. For eye nose we demonstrate implementation our identification best 2D closely linked with organ-specific Despite presence large number drugs simultaneously associated both organs, could discern distinct specifically each class. With availability amounts data array interdependent descriptors, value discovery process it enables dealing multidimensional space.