Cancer evolves dynamically, as clonal expansions supersede or overlap one another, driven by shifting selective pressures, mutational processes and disrupted cancer genes. These processes leave distinctive trails in the genome, which are increasingly capable of deciphering using massively parallel sequencing technology. Bespoke algorithms have been developed to identify somatically acquired point mutations, copy number changes and genomic rearrangements in cancer genomes, and understand their emergence in a temporal context. The findings from our first handful of genomes illustrate the potential for next-generation sequencing to provide unprecedented insights into mutational processes, cellular repair pathways and gene networks associated with cancer development.