Abstract 219: Identifying genetic interactions resulting form diverse biological mechanisms to inform cancer drug development

Abstract

Abstract Identifying gene interactions, such as synthetic lethal relationships, has led to actionable, clinical insights, enabling targeted oncology therapies. However, identifying these interactions has proven difficult, in large part due to the multiple mechanisms by which two genes can interact, such as occurring in the same pathways or conversely occurring in parallel pathways, but involved in the same biological functions. Therefore, experimentally testing all possible sources of gene interactions is costly and time consuming, however a robust computational approach would allow researchers to investigate these interactions efficiently. Here, we describe a machine learning approach that integrates multiple metrics, such as gene-gene pairwise expression correlations, similarity metrics for genetic biological functions, and pathway co-occurrence, to derive a global interaction score. Our model can simultaneously achieve significant predictive performance (Area under the receiver operator curve = 0.8), while also elucidating the underlying mechanism of predicted gene interactions, informing future experimental validation protocols. Additionally, we leveraged transcriptomic and genomic profiles of cancer patients to identify clinically actionable genetic interactions. Our tool provides a novel multi-faceted approach for the detection of gene interactions and can be used to identify biomarkers, novel oncology targets and overall enhance our understanding of biological mechanisms. Citation Format: Manuel Garcia-Quismondo, Olivier Elemento, Neel Madhukar, Coryandar Gilvary. Identifying genetic interactions resulting form diverse biological mechanisms to inform cancer drug development [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 219.