Non-coding driver mutations in human cancer

Despite the vast size of the non-coding genome, driver mutations in this space appear to be relatively infrequent. This Review highlights recent notable discoveries and challenges in the field of non-coding driver mutations, with a focus on somatic point mutations and indels. Tumour formation involves random mutagenic events and positive evolutionary selection acting on a subset of such events, referred to as driver mutations. A decade of careful surveying of tumour DNA using exome-based analyses has revealed a multitude of protein-coding somatic driver mutations, some of which are clinically actionable. Today, a transition towards whole-genome analysis is well under way, technically enabling the discovery of potential driver mutations occurring outside protein-coding sequences. Mutations are abundant in this vast non-coding space, which is more than 50 times larger than the coding exome, but reliable identification of selection signals in non-coding DNA remains a challenge. In this Review, we discuss recent findings in the field, where the emerging landscape is one in which non-coding driver mutations appear to be relatively infrequent. Nevertheless, we highlight several notable discoveries. We consider possible reasons for the relative absence of non-coding driver events, as well as the difficulties associated with detecting signals of positive selection in non-coding DNA.

Automated summary

Despite the vast size of the non-coding genome, driver mutations in this space appear to be relatively infrequent. The transition towards whole-genome analysis is enabling the discovery of potential driver mutations outside protein-coding sequences, but detecting selection signals in non-coding DNA remains challenging.