Primary tumor datasets

About

Abstract: It is now common to sequence for somatic mutations in patient tumors before treatment, but the identification of mutation-specific therapies remains a pivotal challenge in precision medicine. A particularly promising approach is to identify alternative therapies that do not target the mutation directly. For instance, a mutation may increase dependence on a second gene that can easily be targeted instead. In this case, the mutation and the second gene are called a synthetic lethal (SL) pair, since a defect in either gene is compatible with cell viability, but defects in both are lethal to the cell, and the second gene is a synthetic lethal partner of the original mutation. Stanford researchers have developed a novel statistical algorithm that mines large pan-cancer patient datasets (mutation, copy number and gene expression) to identify candidate synthetic lethal partners for cancer mutations. Given a mutation and a cancer type, the output of the algorithm is a relatively short, high-quality list of candidate synthetic lethals that must then be validated experimentally to find the true synthetic lethals. This method has identified candidate synthetic lethal partners for approximately 3000 mutations in 12 different cancers, and can greatly accelerate identification of pharmacologic targets associated with specific somatic mutations in specific tumor types for all kinds of mutations. Stage of Research: The algorithm results have been validated for multiple mutations in multiple cancer types using well-established synthetic lethal interactions and comparison to large-scale functional screens. Key new findings are: Identification of a novel synthetic lethal interaction for a common mutation in acute myeloid leukemia. This interaction has been extensively validated in cell-lines, primary cells and patient-derived xenografts. Identification of a novel synthetic lethal interaction for a recurrent mutation in pancreatic cancer. This has currently been validated in in-vitro assays, and is undergoing more extensive testing. Identification of predictive genetic biomarkers for existing targeted drugs. The biomarkers identified for existing targeted drugs have been thoroughly validated using pharmacological data from cell line panels. Applications: Identification of novel drug targets Identification of targeted therapies for cancer patients based on their genomic profile Repurposing existing drugs Improved design of clinical trials Advantages: Expands the set of actionable mutations in human cancers, including mutations that are not well-represented in existing cell lines Provides a complementary approach to functional screens on cell lines for identifying synthetic lethal interactions Identifies synthetic lethal interactions from primary tumor data and are likely to be more representative of in vivo tumor biology