Fig. 7From: Learning self-supervised molecular representations for drug–drug interaction predictionScaffold side effects profiles and Cluster Analysis. a Binary interaction profiles matrix: Each cell represents the presence (black pixel) or absence (white pixel) of association side-effect: scaffolds. The four plain black lines are named L1 up to L4, respectively. b Scaffold-Scaffold similarity matrix: Scaffolds are clustered based on the similarity of their side effect profiles. Darker cells indicate higher similarity between scaffold interaction profiles. c Side effect—Side effects similarity matrix: Side effects are clustered based on how similar the list of scaffolds that cause them are. Darker cells indicate higher similarity between side effects scaffold profilesBack to article page