Abstract 238: Comprehensive study of gene expression outliers and their regulation mechanisms in pan-cancer

Abstract

Abstract Cancer is a disease characterized by remarkable heterogeneity, with many molecular features significantly associated with tumor progression in specific subsets of patients. Gene expression varies drastically between tumors and within cells of a single tumor. This variability in gene expression can be caused by gene-regulatory mechanisms such as DNA copy number changes, extrachromosomal DNA, or aberrant methylation, amongst many others. These mechanisms sometimes generate extreme outliers: transcripts that show atypically high or low gene expression in a small percentage of cancers. These outliers increase the molecular and phenotypic diversity between individuals, contributing to tumor heterogeneity. Gene expression in cancer has been well studied with many reports of differential gene expression patterns between specific cancer types or subtypes. Importantly, many of these differences were strongly associated with tumorigenesis and tumor progression. For example, the BCR-ABL fusion was discovered in chronic myeloid leukemia with drugs targeting this fusion gene dramatically increasing the survival rate. The EML-ALK fusion gene is another example of a gene-expression outlier and drug target. It occurs in about 4% of non-small-cell lung carcinomas and has been routinely screened for. These examples demonstrate the critical role gene expression outliers play in cancer progression and highlight their potential as biomarkers for diagnostics and identifying novel drug targets. Despite their importance, there has not yet been a comprehensive pan-cancer study of gene expression outliers and their general properties. We lack the answers to fundamental questions such as how many outliers exist in a typical tumor, whether this differs across cancer types, what mechanisms generate the most outliers, whether specific clinical or somatic mutational features correlate with the number or type of outliers, and how many recurrent outliers exist. To answer these questions, we performed a comprehensive and systematic analysis of cancer gene-expression outliers using molecular data from multiple cancer genomics projects. We have created a new statistical outlier detection method and applied it to transcriptomics and proteomics data across 33 cancer types. We used this resource to describe the fundamental landscape of gene-expression outliers, including the most likely genetic and epigenetic mechanisms driving them. The resulting outliers will be further studied for their impact on cancer progression and validated hits will serve as clinically relevant biomarkers and targets for future functional and therapeutic investigations. Citation Format: Jee Yun Han, Stefan Eng, Jaron Arbet, Paul Boutros. Comprehensive study of gene expression outliers and their regulation mechanisms in pan-cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 238.