Imbalanced expression of somatic alleles in
cancer can suggest functional and selective features, and can therefore indicate possible driving potential of the underlying genetic variants. To explore the correlation between allele frequency of somatic variants and total gene expression of their harboring gene, we used the unique data set of matched
tumor and normal
RNA and
DNA sequencing data of 5523 distinct single
nucleotide variants in 381 individuals across 10
cancer types obtained from The
Cancer Genome Atlas (TCGA). We analyzed the allele frequency in the context of the variant and gene functional features and linked it with changes in the total gene expression. We documented higher allele frequency of somatic variants in
cancer-implicated genes (Cancer Gene Census, CGC). Furthermore, somatic alleles bearing premature terminating variants (PTVs), when positioned in CGC genes, appeared to be less frequently degraded via nonsense-mediated mRNA decay, indicating possible favoring of truncated
proteins by the
tumor transcriptome. Among the genes with multiple PTVs with high allele frequency, ARID1, TP53 and NSD1 were known key cancer genes. All together, our analyses suggest that high allele frequency of
tumor somatic variants can indicate driving functionality and can serve to identify potential
cancer-implicated genes.