The advent of high throughput sequencing has revolutionized cancer research. Whole genome sequencing and exome sequencing are routinely used for the identification of somatic mutations, and both enable reliable detection of single nucleotide variants (SNVs), small insertions and deletions, and copy number alterations (CNAs). RNA-Seq is typically used to provide data on gene expression, gene fusions and isoform usage, which is not available from DNA. RNA-Seq is not typically used for detecting point mutations, and almost never for copy number alterations.
I will present superFreq a method for calling absolute and allele-sensitive copy numbers from RNA-Seq. Performance was assessed on two The Cancer Genome Atlas (TCGA) cohorts with either low or high levels of CNAs. Using RNA-seq alone, we found that accuracy and sensitivity was comparable to exomes for large scale copy number alteration, with decreasing sensitivity for focal alterations covering fewer than 100 genes, corresponding to 12Mbp. SuperFreq also calls somatic point mutations from RNA-Seq, with high sensitivity in transcribed regions of expressed genes.
There are tens of thousands of public RNA-Seq cancer samples, of which only a small fraction has been analysed for point mutations, and almost none for copy number alterations. SuperFreq can be execute without a matched normal control, which unlocks reanalysis of these cohorts. I will demonstrate how SuperFreq can be applied to investigate new links between point mutations, copy number alterations and expression profiles.
As a proof of concept, I will demonstrate a comparative analysis of point mutations and CNAs detected in the TCGA Colon (CRC) cohort using only the RNA-Seq cancer samples. We recapitulate the recurrent copy number landscape of CRC and confirm the established observation of low levels of CNAs in patients with hypermutation.