SoMAS: Finding somatic mutations associated with alternative splicing in human cancers

Aberrant alternative splicing is prevalent in cancer and affects most cancer hallmarks involving proliferation, angiogenesis, and invasion. Somatic point mutations can exert their cancer-driving functions via splicing disruption. We propose “SoMAS” ( So matic M utation associated with A lternative S plicing), an efficient computational pipeline based on principal component analysis techniques, to explore the role of somatic mutations in shaping the landscape of alternative splicing via both cis - and trans -acting mechanisms. Applying SoMAS to 33 cancer types consisting of 9,738 tumor samples in The Cancer Genome Atlas, we identified 908 somatically mutated genes significantly associated with altered isoform expression in three or more cancer types. These genes include many well-known oncogenes/suppressor genes, RNA binding protein and splicing factor genes with both biological and clinical significance. Many of our identified SoMAS genes were corroborated to affect gene splicing by independent cohorts and/or methodologies. With SoMAS, we for the first time demonstrate the potential network of somatic mutations associated with the overall splicing profiles of cancer transcriptomes, bridging the genetic and epigenetic regulation of human tumorigenesis in an innovative way.