High-resolution mapping of a rapidly evolving complex trait reveals genotype-phenotype stability and an unpredictable genetic architecture of adaptation

The extent to which adaptation can be predicted is unknown. Here, we leveraged a longitudinal sampling design to test the efficacy of genomic prediction of trait evolution in an ecologically-relevant setting. Specifically, we monitored genome-wide allele frequencies and pigmentation variation in genetically diverse populations of Drosophila melanogaster across seven generations of evolution in both field mesocosms and a controlled, lab-based setting. At two points during trait evolution, we conducted a high-powered quantification of trait architecture that produced a well-resolved genotype-phenotype map. While we were able to use this map to correctly infer the direction of pigmentation evolution in both the field and lab mesocosms, the particular loci responding to selection, and thus the architecture of adaptation itself, was largely unpredictable. Further, we quantified a striking stability of the genotype-phenotype map, even across independent and genetically diverged populations. Our results hold implications for both the promise and limitations of genomic prediction.