Fig. 1From: Discovery of moiety preference by Shapley value in protein kinase family using random forest modelsOverview of establishing kinase family inhibitor model and identifying key features. A Main procedure. B Dataset collection and identification of kinase family inhibitors. C Feature encoding with the MMIFs, which combined the moieties of Checkmol, ring moiety, PubChem, and in-housed moieties. D Identification of important moieties by Shapley value and validation by kinase-inhibitor structure complexesBack to article page