Abstract 1304: Whole-exome analysis of 1,135 melanomas reveals new significantly mutated cancer genes and copy number alterations

Abstract

Abstract Introduction: Identifying driver mutations and significantly mutated genes (SMGs) is critical to characterize therapeutic targets and understand biological subtypes. Past studies in melanoma have identified ~20 SMGs and 4 genomic subtypes (BRAF, NRAS, and NF1, and triple wildtype). Critically, these studies have been powered to detect SMGs only at a frequency of >~10%. However, clinically targetable drivers can occur at much lower frequencies (e.g. ALK-rearrangmenet in lung cancer). Thus, we hypothesized that a uniform meta-analysis of all publicly available melanoma whole exome sequences would provide power to identify clinically SMGs not previously associated with melanoma. Methods: We assembled and uniformly analyzed a cohort of 1,135 tumor and matched germline primary and metastatic non-uveal melanoma whole exomes. All samples were realigned to hg19 using the Picard best practice realignment pipeline. These samples passed several QC metrics that included coverage (DepthOfCoverage, 50X tumor and 20X normal), cross sample contamination (ContEst, < 5%), tumor purity (FACETS, > 20%), and tumor-in-normal (deTiN, < 30%). Mutational significance analysis was performed using MutSigCV2 followed by filtering via biological priors using OncoKB. CNA profiles were generated using GATK4 best practices pipeline. Results: We identified 77 SMGs (MutSig q < 0.1) that were implicated in either melanoma or other cancer types according to OncoKB's curated list of 1,018 cancer genes, some of which that have not been considered SMGs or previously reported in melanoma. Novel genes were involved in immune response [B2M (1.7%) and PTPRC (12.1%)], DNA damage checkpoints [MDC1 (7.5%)], negative regulation of the MAPK pathway [CBL (4.6%)], MYC suppression [MGA (9.4%)], and the PI3K/AKT pathway [PIK3CA (3.4%) and PIK3CB (3.5%)]. Within the “triple-wt” subset of melanomas (n=212), we identified additional SMGs including IDH1, SF3B1, CTNNB1, and MAPK2K1 present in 3-5% of tumors, suggesting additional potential therapeutic targets. Copy number profiles of the cohort were consistent with previous findings. By increasing the resolution from cytobands to 1Mb regions, we identified a 1Mb region in chromosome 3 (chr3:138Mb-139Mb) that was enriched in amplifications (29.2%). Genes in this region include MRAS, PI3KCB, FOXL2, and FAIM. Conclusion: Through assembly and uniform analysis of the largest melanoma whole exome cohort to date, we identified significantly mutated known cancer genes not previously associated with the disease, and identified additional drivers within an incompletely characterized subtype of melanoma. This also enabled us to find genome regions enriched in copy number alterations. Taken together, our findings may inform the additional classification of melanoma and identification of novel drivers to shed light on biological subtypes and identify potential therapeutic targets. Citation Format: Jake R. Conway, David Liu, Stephanie Wankowicz, Amaro Taylor-Weiner, Felix Dietlein, Eliezer Van Allen. Whole-exome analysis of 1,135 melanomas reveals new significantly mutated cancer genes and copy number alterations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1304.