CanDriS A Statistical Framework for Posterior Profiling Cancer-Driving Sites based on Recurrent Somatic Mutations

Current cancer genomics databases have accumulated millions of somatic mutations that remain to be further explored. Due to the over-dominance of passenger mutations unrelated to cancers, the great challenge is how to identify somatic mutations that are cancer-driving. Under the notion that carcinogenesis is a form of somatic-cell evolution, we develop a two-component mixture model: while the ground component corresponds to passenger mutations, the rapidly-evolving component corresponds to driver mutations. Then, we use an empirical Bayesian procedure to tatistically calculate the posterior probabilities of cancer-driving for all sites of a gene. Based on these, we developed a software CanDriS, which is available from GitHub (https://github.com/jiujiezz/candris).