Predicting drug response in oncology: the WIN approach

In this talk the need and main chllanges in developing precision oncology scores will be presented. We will describe in details the two scores developed by the WIN consortium: the WINTHER1.5 and the SIMS sscores. The first, WINTHER1.5, quantifies the fit between a given expression profile and the literature describing drugs and expression. The main challanges in developing this score was the lack of data to estimate parameters, and the need to develop a structured knowledgebase summarizing the literature about drug efficacy. For the second, pathway based score, the main challanges were to organize the main response-related genes in "nodes", and coming up with a measure of node involvment in cancer. FInally, we will explain why we think that computational prediction of score accuracy is key to the development of precision oncology, and describe the progress we have made to address this problem.