Abstract LB-330: Somatic copy number alterations at oncogenic loci show diverse correlations with gene expression

Abstract

Abstract Somatic copy number alterations (SCNAs) affecting oncogenic drivers have a firmly established role in promoting cancer. However, no agreed-upon standard exists for calling locus-specific amplifications and deletions in each patient sample. Here, we report the correlative analysis of copy number amplitude and length with gene expression across 6,109 samples from The Cancer Genome Atlas (TCGA) dataset across 16 cancer types. Using specificity, sensitivity, and precision-based scores, we analyzed over 30 recurrently focally altered loci that had sufficient pan-cancer penetrance for robust statistical assessment. We were able to assign optimized amplitude and length cutoffs for nine recurrent SCNAs affecting known oncogenic drivers, using mRNA expression as a functional readout. These cutoffs captured the majority of SCNA-driven, highly-expression-altered samples. The majority of oncogenes required only amplitude cutoffs, as high amplitude samples were almost invariably focal; however, CDKN2A and PTEN uniquely required both amplitude and length cutoffs as primary predictors. For PTEN, these extended to downstream AKT activation. In contrast, the remaining 20+ assessed SCNA genes, particularly ones located peri-telomerically or in fragile sites, showed poor expression-copy number correlations. Finally, we describe several novel associations between copy number and gene expression that have implications for how SCNAs are assessed as harboring candidate cancer driver genes. Overall, our analyses identify optimized amplitude and length cutoffs as efficient predictors of gene expression changes for specific oncogenic SCNAs, yet warn against one-size-fits-all interpretations across all loci. Our results have implications for cancer data analyses and the clinic, where copy number and mutation data are increasingly used to personalize cancer therapy. Citation Format: Jason Roszik, Ching Jiun Wu, Alan E. Siroy, Alexander J. Lazar, Michael A. Davies, Scott Woodman, Lawrence N. Kwong. Somatic copy number alterations at oncogenic loci show diverse correlations with gene expression. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr LB-330.