Just as gardens reflect the presence—or absence—of careful maintenance, genomes bear the imprint of their caretakers. By watching yeast “gardens” fall into different states of disrepair, we can decode the signatures thus generated and correlate these to signatures appearing in human cancers, allowing us to trace them back to their molecular roots.
In this study, our collaborators and us systematically disrupted 13 different “caretaker” genes in yeast—each one responsible for a distinct aspect of DNA repair, replication, or genome organization. Over many generations, we let these populations grow and then sequenced their genomes to see what kinds of “weeds and structural collapses” accumulated. The result was a set of mutational signatures—unique patterns of DNA changes that acted like fingerprints of the underlying defect.
What emerged was striking: each disabled gene produced its own characteristic mutational trajectory. Some defects led to a slow accumulation of small errors, like scattered weeds across the garden path. Others produced sudden structural failures, akin to fences toppling or entire plots being overtaken by a single species. Still others triggered wholesale chaos, where both small weeds and large collapses occurred together. These patterns were not random—they reflected the specific molecular job that each missing caretaker would normally perform.
Crucially, several of these yeast-derived signatures closely matched those catalogued in human cancers. For example, the loss of Rad51, a key player in homologous recombination, created a mutational profile nearly identical to COSMIC Signature 3, a hallmark of BRCA1/2-deficient breast and ovarian tumors. Similarly, yeast lacking Msh2, a mismatch repair protein, mirrored the patterns seen in colorectal cancers with mismatch-repair deficiency. In essence, yeast served as a simplified testbed in which to watch mutational processes unfold in real time, revealing the direct causal links between repair defects and their mutational fingerprints.
Beyond point mutations, we also tracked loss of heterozygosity (LOH)—where stretches of the genome lose one parental copy and become uniform. In garden terms, it is as though a once diverse flowerbed becomes dominated by a single species, crowding out the others. In yeast lines lacking certain caretakers, especially Rad27 and Tsa1, LOH occurred at high rates, reshaping the genome’s landscape in ways that parallel clonal sweeps in human cancers.
By connecting the dots between yeast mutational trajectories and human tumor signatures, our work provides a mechanistic framework for interpreting the “molecular archaeology” of cancer genomes. It shows that the distinctive scars left on DNA are not arbitrary but can be traced back to specific failures in genomic maintenance. In the same way that a gardener can tell whether neglect came from broken irrigation, absent pruning, or collapsed fencing, we can now begin to read cancer genomes and infer which caretakers went missing along the way.
In the end, the message is clear: the health of a genome, like the health of a garden, depends on careful and coordinated upkeep. When caretakers falter, the signatures of neglect remain etched into the landscape. By learning to read these signatures, we gain not only a window into the evolutionary forces that shape life but also powerful new tools for diagnosing and treating human disease.