Figure 1
From: Explainable drug sensitivity prediction through cancer pathway enrichment
Illustration of the PathDSP method and validation plan. The chemical and cell line omics data is processed using Morgan fingerprints (chemical) and pathway enrichment (drugs targets, gene expression, mutation and CNV) is fed into FNN (A). The trained model on GDSC is tested on CCLE data (B) and SHAP scores are used to identify chemical and pathway features that contribute most to the prediction scheme (C).
