Pharmacological stress exposes hidden allelic background effects in genetic interaction screen normalisation

Synthetic Genetic Array (SGA) analysis comprises the high-throughput crossing of a query deletion strain against a genome-wide deletion library to score fitness interactions in thousands of double mutants. SGAs have produced comprehensive genetic interaction maps in yeasts and have emerged as a leading platform for pharmacogenomics: mapping genetic modifiers of drug response, identifying synthetic lethal targets and illuminating mechanisms of drug action and resistance. We have previously demonstrated that in fission yeast, the ade6 mutant is functionally neutral relative to the parental library and can serve as a standard negative control for SGA screens. Here, while we confirm our previous observation, we show that this neutrality fails under pharmacological stress. Using Torin1, an ATP-competitive TOR kinase inhibitor, we demonstrate that the ade6 SGA fitness profile diverges from that of the parental library in a dose-dependent and genomically widespread manner. At 2 M Torin1 only 12.2% of scored genes exceed a 1.5-fold fitness difference between backgrounds; at 3 M this proportion rises to 43.2% -a 3.5-fold increase driven by qualitative reorganisation of the genetic interaction landscape rather than simple scaling of pre-existing differences. Gene ontology analysis of divergent genes implicates autophagy, iron starvation responses, central carbon metabolism, and vesicle trafficking, consistent with TOR-regulated nutrient adaptation being differentially affected by the ade6-M210/M216 point mutations in the library versus the ade6 null in the SGA control. Our results have implications in fission yeast and beyond and we propose solutions towards reliable retrieval of genetic interactions.