Global impact of somatic structural variation on the cancer proteome

Both proteome and transcriptome data can help assess the relevance of non-coding somatic alterations in cancer. Here, we combine mass spectrometry-based proteomics data with whole genome sequencing data across 1307 human tumors spanning various tissues to determine the extent that somatic structural variant (SV) breakpoint patterns impact protein expression of nearby genes. We find that only about 25% of the hundreds of genes associated with recurrent somatic SV-mediated cis-regulatory alterations at the mRNA level are similarly associated at the protein level. SVs associated with enhancer hijacking, retrotransposon translocation, altered DNA methylation, or fusion transcripts are implicated in protein over-expression. SVs combined with altered protein levels considerably extend the numbers of patients with tumors somatically altered for critical pathways. We catalog both SV breakpoint patterns associated with patient survival and genes with nearby SV breakpoints associated with increased cell dependency in cancer cell lines. Proteogenomics identifies targetable non-coding alterations. Using SVExpress, we defined genes with altered expression (by protein or mRNA) associated with nearby somatic SV breakpoints. Relative to each gene, genomic region windows considered included the within-gene regions and within 100kb upstream or 100kb downstream of the gene. For the above regions, SVExpress constructed a gene-to-sample matrix with entries as 1, if a breakpoint occurs in the specified region for the given gene in the given sample, and 0 if otherwise. We also used SVExpress to examine a 1Mb region surrounding each gene, using the “relative distance metric” option, whereby breakpoints close to the gene will have more numeric weight in identifying SV-expression associations, while breakpoints further away but within 1Mb can have some influence. Gene-level SV-expression association analyses included 15439 unique named genes, 10087 of which had protein values for at least 400 tumors. Using the geneXsample SV breakpoint matrix, SVExpress assessed the correlation between expression of the gene and the presence of an SV breakpoint using a linear regression model (with log-transformed expression values), incorporating sample cancer type and gene-level CNA.