Adaptive responses to antifungal selection depend on dose and drug combination in experimentally evolved yeast
Natural populations rarely face a single stressor at a single intensity, yet most laboratory evolution experiments impose exactly that. Theory predicts that stronger or compounded selection pressures should change the magnitude of adaptive change as well as the specific loci involved, but this prediction is rarely tested directly. Here, we use an evolve-and-resequence approach to compare genomic responses in Saccharomyces cerevisiae across antifungal environments differing in the concentration and number of drugs applied. Twelve replicate populations founded from a genetically diverse ancestor were evolved for >50 generations under low and high concentrations of caspofungin or clotrimazole alone, a combination of the two drugs, or no drug, with pooled-population sequencing at three timepoints. We found little overlap among candidate loci between concentrations of the same drug or between single and combined treatments. Alleles favored in one environment did not shift predictably in the others, departing from expectations of additivity. We also observed that most candidate loci fell outside known antifungal resistance genes, except for FKS1 in all caspofungin treatments and a putative gain of the ERG11-containing chromosome VIII in 9 of 12 populations evolved under high clotrimazole. Remaining candidates implicated diverse cellular functions that differed among environments. Furthermore, we found little evidence of fitness costs in drug-free medium across all evolved populations. Our results collectively show that stronger and combined antifungal selection redirects genomic responses rather than simply amplifying those favored under less challenging conditions, highlighting the limits of predicting adaptation from simpler selection regimes.