Ecological stoichiometry and life history theory, not the identity of genomic variants, predict rapid adaptation

Abstract

SignificanceForecasting how populations adapt to changing environments is a central challenge in evolutionary biology. We tested the predictability of rapid adaptation by tracking both phenotypic and genomic trajectories of evolution in replicate outdoor populations of Drosophila melanogaster and comparing these data to predictions derived from life history theory, ecological stoichiometry, and genomic data. We found that life history theory and ecological stoichiometry forecasted the direction of phenotypic evolution, even with temporal changes in the direction of evolution. In contrast, genomic predictors were largely unable to forecast the SNPs or genes underlying adaptation. These results demonstrate the utility of ecological frameworks to provide general, but robust, forecasts of adaptation in complex environments. Determining the scale and scope of the predictability of evolution is fundamental to understanding biological processes and to managing biodiversity. Ecological frameworks, including life history theory and ecological stoichiometry, offer testable predictions about the direction of adaptation and trade-offs between traits in response to environmental change. Similarly, well-characterized molecular pathways, genotype-phenotype linkages, and prior evolutionary genomic studies could be predictive of the genomic architecture underlying adaptation. We tested whether ecological frameworks and evolutionary genomic data can be used to forecast rapid adaptation in replicated outdoor populations of Drosophila melanogaster evolving in response to natural seasonal fluctuations from summer to late fall. Life history theory predicted the observed pattern of adaptive tracking: in summer, reproductive output increased and stress tolerance decreased, while in fall, this direction reversed, with evolution of increased stress tolerance at the expense of reproduction. Stoichiometric phenotypic evolution was also predictable, with phosphorus and magnesium content, both linked to growth rate, and alkali metal, associated with maintaining homeostasis in response to thermal stress, showing rapid and parallel evolution indicative of adaptation. Temporal genomic data revealed a complex genomic architecture of temporal adaptation and the SNPs and genes involved in adaptation were largely unpredictable. These results demonstrate that ecological frameworks, more than genomic data, have utility in forecasting adaptation in complex and variable environments.