Widespread genetic diversity in the yeast knockout collection reflects gene-specific selection pressures and resembles cancer evolution

One largely unfulfilled expectation of the human cancer genome sequencing effort is that it has not identified common early mutations that initially drove cancer progression by Darwinian selection. A potential explanation is that a mutation in any of the many different genes can drive convergent evolution to a common cancer phenotype. If so, strains of the yeast knockout (YKO) collection generated several years ago may similarly have been subjected to selective pressures for new mutations in response to the specific deleted gene.;Previously, we found that deletion of the conserved mitochondrial fission protein Fis1 selects for a specific secondary mutation in the stress-response gene WHI2. This loss of function mutation in Whi2 causes increased cell death' sensitivity, overgrowth on low amino acids, increased overgrowth sensitivity to rapamycin, and reduced oxygen consumption. These phenotypes are reminiscent of human tumor cells that are sensitive to anti-cancer agents, fail to respond to nutrient limitation, are more TOR-dependent, and undergo the Warburg shift. It is suggested that the selection for WHI2 mutations in Deltafis1 yeast is to compensate the mitochondrial defect caused by Fis1 deletion. We suggest that this selection process is reflective of early tumorigenesis.;Therefore, I screened the ∼5000 strains in the YKO collection for cell death, low amino acid growth and rapamycin sensitivity. Results indicate that 751 different knockout strains have an overgrowth phenotype on low amino acid media. Among these, 259 different knockouts exhibit the same tumor-like phenotypes as the Deltafis1 and 141 of these 259 knockouts have mutations in the same gene WHI2. Strikingly, more than 50% of the 751 knockouts appear to be quasispecies composed of mixed genotypes that are not due to knockout construction errors or other mishaps. These findings suggest that a primary mutation in a large number of different genes could drive evolution to a similar phenotype, implying that nearly every gene is critically important. I thus conclude that the selection pressures for growth and survival in yeast have considerable overlap with the same pressures leading to tumorigenesis in humans, and yeast will be a powerful tool to study these initial steps in tumorigenesis.