The study team followed de novo genome evolution from individual cells using single-genome sequencing. They sampled 34 primary and established cancer cell lines and organoids. The cohort covered five cancer histologies: high-grade serous ovarian carcinoma, pancreatic ductal adenocarcinoma, colorectal cancer, breast cancer and sarcoma. This design allowed investigators to measure short-timescale generation of chromosomal changes rather than only inferring events over years.
Data indicated that Chromosomal instability (CIN) operates as a highly active diversification process in cancer. The authors reported a rate of roughly one to five new copy number alterations (CNAs) generated each cell cycle. That is a rapid mutational input on a per-division basis, implying that genome structure can change measurably over days in cultured cancer cells.
Investigators reported a tight correlation between CIN rates and measures of tumour genomic complexity within their cohort. In other words, samples with higher ongoing CNA generation showed greater overall genomic complexity. The study team presented this relationship as validation of commonly held assumptions linking chromosomal instability to complex tumour genomes.
The authors examined whether newly generated CNAs clustered near candidate genomic features. Data indicated that most new CNAs were not readily explained by proximity to those examined features. Instead, the distribution resembled a largely stochastic process across the genome. This suggests that, at the timescale measured, emergent CNAs arise broadly rather than concentrating at predictable hotspots defined by the candidate features the authors tested.
The cohort showed evidence of negative selection acting on large CNAs during the short-term experimental evolution. The investigators observed fewer large-scale alterations persisting than would be expected from generation rates alone, indicating that large CNAs can incur fitness costs in these cellular contexts and are therefore less likely to be retained.
The authors reported that whole genome duplication (WGD) events acutely increased CIN rates per cell. That acute increase was not transient in the population-level view; a broader relationship emerged across the dataset linking higher ploidy with elevated CIN across cancer types. Put another way, WGD amplified the per-division generation of CNAs, and samples with high ploidy tended to show higher ongoing chromosomal instability.
The abstract indicates that the authors also tracked cancer genome evolution under the selection pressure of chemotherapy, with further observations implied. The publicly available abstract text available from the source was truncated and ends mid-sentence; therefore, the source did not report the full chemotherapy-related results in the provided abstract. No additional experimental details, numerical breakdowns, methodological parameters (for example, sequencing depth, timepoints, or exact selection protocols), or complete findings beyond those summarized above were present in the accessible abstract. Readers should consult the full preprint and supplementary material—linked by the source—for the complete methods and any chemotherapy-selection data.
Limitations documented in the source abstract include that the report is a preprint and has not been peer reviewed. The abstract does not provide detailed methods for CNA calling thresholds, the precise candidate genomic features tested for proximity effects, or the experimental durations for selection assays. Those methodological specifics were not reported in the abstract text and must be verified in the full manuscript.
Unanswered questions left in the available abstract include which genomic features were tested and excluded as explanatory for CNA localization, the quantitative metrics used to define ‘‘large’’ CNAs subject to negative selection, and the mechanistic basis by which WGD increases per-cell CIN rate. The preprint format indicates these topics and any chemotherapy-evolution findings may be addressed in the full article, but the abstract alone does not supply the data or detailed protocols needed to evaluate those items.