Novel Global Network Scores to Analyze Kinase Inhibitor Profiles

Abstract We propose new developments of global network scoring methods previously introduced to estimate the importance of genes in diseases. We apply these methods to drug proteomics profiles, which consist of drug targets, to determine the part of the human interactome perturbed by a drug. Drug and disease network scores can be combined to obtain novel side-effect and pathway association prediction strategies. We illustrate our methods comparing four kinase inhibitor profiles (dasatinib, bosutinib, imatinib, bafetinib) ranging from very promiscuous to highly specific. We predict and identify the cause of plausible side-effects of bosutinib