Abstract 1558: Data-driven approach to detecting novel gene interactions in cancer with applications to drug response prediction and cancer stratification

Abstract

Abstract The effect of a gene’s activity on a phenotype depends on the context of a complex network of functionally interacting genes. Several genetic interactions (GIs), such as Synthetic lethality and Synthetic Rescues, have been reported to have a significant functional role in cancer progression and provide potential candidates for selective cancer treatments. However, numerous other types of GIs with potential clinical significance are yet to be explored. In this work we generalize the concept of GI and detect ~70,000 GIs of different types with both molecular and clinical signature. We demonstrate their clinical predictive value as well as their ability to stratify breast cancer patients into refined clinical subtypes that might allow for better diagnosis and modified course of treatment. These results compare favorably with previous sequence based approaches and provide evidence for the importance of context specific genomic events and their effect on tumor progression. Additionally, the GI network accurately predicts patients’ drug response, where difference GI types are found to be predictive of distinct drugs in a complementary manner. This work provides the basis for future exploration of novel GI types as well as individual interactions with major impact on cancer progression. Citation Format: Assaf Magen, Avinash Das, Joo Lee, Sridhar Hannenhalli, Eytan Ruppin. Data-driven approach to detecting novel gene interactions in cancer with applications to drug response prediction and cancer stratification [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1558. doi:10.1158/1538-7445.AM2017-1558