Identification of cancer drivers from tumor-only RNA-seq with RNA-VACAY
Detecting somatic mutations is a cornerstone of cancer genomics and clinical genotyping; however, there has been little systematic evaluation of the utility of RNA sequencing (RNA-seq) for somatic variant detection and driver mutation analysis. Variants found in RNA-Seq are also expressed, reducing the identification of passenger mutations and would not suffer from annotation bias observed in whole-exome sequencing (WES). We developed RNA-VACAY, a containerized pipeline that automates somatic variant calling from tumor RNA-seq data, alone, and evaluated its performance on simulated data and 1,349 RNA-seq samples with matched whole-genome sequencing (WGS). RNA-VACAY was able to detect at least 1 putative driver gene in 15 out of 16 cancer types and identified known driver mutations in 5 and 3 UTRs. The computational cost and time to generate and analyze RNA-seq data is lower than WGS or WES, which decreases the resources necessary for somatic variant detection. This study demonstrates the utility of RNA-seq to detect cancer drivers.