Evolving copy number gains promote tumor expansion and bolster mutational diversification

Abstract

Charting the evolutionary history of rampant somatic copy number alterations (SCNA) is an indispensable step toward a deeper understanding of their roles in tumor development. However, the existing SCNA timing analysis is limited to low copy number states and initiating gains, which are not necessarily close to the onset of the malignant proliferation. Moreover, it remains a critical question if the timing of an SCNA reveals the corresponding variants fitness effect. Here we propose a framework to estimate the arrival time of a clonal SCNA, i.e., the time delay from its last alteration to the start of population expansion, in addition to its initiation time when the first alteration occurs. Our method leverages the bias that a genomic segment, when resting on a copy number (CN) state, accumulates somatic single nucleotide variants (SSNV) at a rate proportional to its CN. From the whole genome sequencing data of 184 samples from 75 patients across five tumor types, we commonly observed late clonal CN gains following early initiating events, which appear right before the expansion leading to the observed tumor sample(s). Genome doubling (GD) can be late, but post-GD CN evolution is prevalent in the genealogy of the most recent common ancestor of patient tumors. Notably, mathematical modeling suggests that late evolving events could contain rate-limiting drivers. The advantage of evolving gains could arise from the dosage increase of cancer genes in proliferative signaling and amplification of early functional variants. In addition, evolving SCNAs bolster the diversification of SSNVs between sub-populations, exacerbating the vicious circle between malignant growth and accumulation of genomic heterogeneity. Our findings have broad implications for resolving the trajectory of SCNAs, discerning the CN markers of malignant growth in vivo, and properly quantifying tumor heterogeneity in aneuploid tumors.