The study team reported single tumors containing individual cells whose metaphase chromosome counts ranged from 25 to over 200. That spread is striking. It implies that, in some cancers, chromosome segregation fidelity can fall to ultra-low levels, producing an extraordinary numerical heterogeneity among neighboring cells. The authors argued that standard models—where catastrophic genome rearrangement is followed by clonal stabilization—do not account for this pattern when observed at single-cell resolution.
Data indicated that this extreme karyotypic variability is recurrent across multiple cancer types, but it was most common in acral melanoma. The cohort showed that acral tumors frequently harbor cells with wildly divergent chromosome counts. The investigators noted that acral melanoma tumors display a constellation of structural abnormalities alongside the numerical instability, a combination that distinguishes them from many other tumor samples the team examined.
Bulk genomic sequencing failed to resolve the subpopulation-level structural variability, the authors reported. To overcome that limitation, the study group applied single-cell whole-genome sequencing. Single-cell profiles revealed ongoing structural variation rather than a single locked-in karyotype. Cells within the same tumor showed different structural rearrangements, indicating that changes in chromosome number and architecture continue to arise across cell generations rather than being confined to one catastrophic event followed by stable expansion.
Sequence patterns in affected genomes carried the signature of microhomology-mediated break-induced replication (MMBIR), according to the authors. They described MMBIR as a backup DNA repair pathway that can trigger chromoanasynthesis—complex, templated rearrangements arising from replication-based template switching. The single-cell data therefore suggested a mechanistic contributor: repeated, error-prone repair events that generate structural complexity and maintain a state of flux rather than allowing long-term clonal stabilization.
Cytogenetic analyses documented the presence of dicentric chromosomes, translocations, and complex structural variants in the tumors with extreme numerical instability. The authors reported that these substrates likely promote continued mis-segregation during mitosis. Dicentric chromosomes, by their nature, can generate breakage–fusion–bridge cycles and ongoing rearrangement; translocations and complex SVs further rewire chromosomal architecture and may create new loci for subsequent instability.
The investigators proposed a model in which some cancer genomes do not pass through a single catastrophic rearrangement and then settle into a stable lineage. Instead, these genomes undergo continuous evolution of chromosome number and structure across generations. The result is extreme intratumoral heterogeneity and a shifting repertoire of gene fusions, which the authors observed in their single-cell data. This continuous-evolution state challenges the dominant punctuated model of karyotypic change that has guided much of cancer genome interpretation.
The preprint did not provide peer-reviewed clinical correlations or therapeutic-response data. The authors posed testable questions: whether tumors exhibiting continuous karyotypic fluctuation can be classified as a distinct biological category and whether that state confers selective dependence on backup repair pathways such as MMBIR. The team also noted, implicitly, that standard bulk sequencing pipelines cannot detect these dynamics and that single-cell approaches are required to quantify intra-tumor structural flux.
Limitations reported in the manuscript included the preprint status of the work and the absence of validated therapeutic dependencies; the study did not report clinical outcome measures or detailed intervention studies. The final sentence of the abstract framed next steps as experimental tests of dependency on backup repair pathways—a specific, concrete research agenda rather than a broad clinical recommendation. Readers should note the manuscript had not been certified by peer review at the time of posting.