Abstract

Identifying complementary genetic drivers of a given phenotypic outcome is a challenging task that is important to gaining new biological insight and discovering targets for disease therapy. Existing methods aimed at achieving this task lack analytical flexibility. We developed Candidate Driver Analysis or CaDrA, a framework to identify functionally-relevant subsets of binary genomic features that, together, are associated with a specific outcome of interest. We evaluate CaDrAs sensitivity and specificity for typically-sized multi-omic datasets, and demonstrate CaDrAs ability to identify both known and novel drivers of oncogenic activity in cancer cell lines and primary tumors.