Abstract 2346: ARDE: Detecting selectively expressed cancer driver mutations through integration of exome and transcriptome sequencing data

Abstract

Abstract Integrating multi-omics data from cancer patient samples can potentially advance the discovery of cancer driver mutations (CDM) towards establishing a knowledgebase for cancer precision medicine. However, it remains unclear whether CDMs are selectively expressed at RNA levels unexpected from DNA copy numbers and whether identifying such selectivity will enable more accurate discovery of CDMs. To answer this question, we developed a novel statistical algorithm, ARDE (Allelic RNA DNA expression), which estimates tumor purity, allele-specific copy number and intratumor heterogeneity (ITH) from matched whole-exome and whole-transcriptome sequencing data, deconvolutes tumor allele-specific expression levels (ASEL) and uses a two-component mixture probability model to identify selectively expressed alleles (SEAs) in the tumor compartment of heterogeneous tumor tissues. We applied ARDE to analyze melanoma and breast samples from The Cancer Genome Atlas (TCGA). Tumor purities estimated by ARDE were highly correlated with estimation from pathology review, DNA methylation array, gene expression microarray and SNP array data. ITH scores can significantly stratify breast cancer subtypes and survival time of melanoma patients. The deconvoluted tumor ASEL showed consistently higher correlation with the averaged gene expression levels in tissue-matched single tumor cells than in normal cells. Higher proportions of somatic mutations are selectively expressed than germline. Top SEAs are highly enriched in deleterious and gain-of-function (GOF) mutations. Stratifying melanoma patients by the expression selectivity of BRAF V600E mutation resulted in a 10-fold improvement of significance in survival analysis. These results indicate that ARDE will likely be an effective, widely applicable approach for accurate discovery of CDMs from tumor tissue sequencing data. (See Table 1.) Table 1. Top 10 significant positive-selectively expressed somatic mutations MelanomaBreast cancerGenepSNVFunctionSIFTGenepSNVFunctionSIFTBRAFp. V600EGOFDAKT1p.E17KGOFDNRASp. Q61RGOFDPIK3CAp.H1047RGOFTCDKN2Ap. R114LGOFDERBB2p.L755SGOFDIDH1p. R132CSwitch of functionDTP53p.R273HGOFDEEF1A1p. H349YDGATA3p.M293KDRAC1p. P29SGOFDTP53p.P278TDNRASp. Q61KGOFDFOXA1p.S250FDSF3B1p. R625HGOFDTP53p. R175HGOFDNRASp. Q61LGOFDPIK3CAp.H1047LGOFTMAP2K1p. P124SGOFDERBB2p.V777LGOFT Citation Format: Fang Wang, Shaojun Zhang, Yiwen Chen, Traver Hart, Kenna Shaw, Funda Meric-Bernstam, Gordon Mills, Ken Chen. ARDE: Detecting selectively expressed cancer driver mutations through integration of exome and transcriptome sequencing data [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 2346.