Signatures 1 and 17 show increased propensity to create mutational hotspots in the human genome

Abstract

AbstractRecurrence of somatic mutations at the exact same position across patients (hotspots) are often identified as potential cancer drivers, assuming that they are unlikely to be generated by neutral mutagenesis. Recent studies have challenged this by identifying examples of mutational processes that generate passenger hotspots. However, no comprehensive study to identify and quantify the determinants of hotspots formation across tumours has been carried out to date. In this work, we conducted a systematic analysis of passenger hotspot events across more than 7,500 whole genome sequences from different malignancies. We found that mutational signatures 1 (SBS1) and 17 (SBS17a and SBS17b) have the highest propensity to form hotspots, generating 5-80 times more than other common somatic mutational processes. The trinucleotide mutational probabilities and genomic sequence composition partially explain the high SBS1 hotspot propensity. Strikingly, the vast majority of hotspots (46-96%) contributed by different signatures remain unexplained after correcting for their sequence context preferences and the large-scale mutation rate variability. This finding reveals the extension of our lack of knowledge about how mutations occur, and highlights the need of identifying and subsequently modelling additional sequence and chromatin features that influence mutation rate at base resolution. This is key to accurately modelling the mutation rate under neutrality, an essential step for identifying cancer drivers, and –given the known activity of SBS1 across other organisms and in germ cells– also for reconstructing evolutionary histories and studying genome evolution.