A kinase inhibition map approach for tumor sensitivity prediction and combination therapy design for targeted drugs

Drugs targeting specific kinases are becoming common in cancer research and are a basis for personalized cancer therapy. Some of these drugs have the capacity to target multiple kinases. Promiscuous kinase inhibitors can be effective but the "off-target" effects can bring in toxicity for the patient. Thus the success of targeted cancer therapies with nominal harmful side effects is dependent on administering a single or multiple combinations of kinase inhibitors that targets the minimum number of kinases required to inhibit the tumor pathways. This requires a framework to predict the tumor sensitivities of a drug or drug combination based on the knowledge of the kinase inhibitors of a drug. In this article, we present a novel approach to predict the tumor sensitivities of a drug based on the generation of deterministic and stochastic Kinase Inhibition Maps. We build sensitivity maps or truth tables for a cell line from experimentally generated tumor sensitivities to kinase inhibitor drugs and use them to predict the sensitivity of a new drug or drug combinations based on known kinase inhibitor targets. We test our algorithms on a dataset of a dog osteosarcoma cell line with 317 possible kinase inhibitor targets after application of 36 targeted drugs. Our proposed algorithms are able to predict the sensitivities with high accuracy based on the given kinase inhibitor targets.