Scaling the fitness effects of mutations with respect to differentially adapted Arabidopsis thaliana accessions under natural conditions

Mutations are the ultimate source of genetic variation for natural selection to act upon. A major question in evolutionary biology is the extent to which new mutations can generate genetic variation under natural conditions to permit adaptive evolution over ecological time scales. Here we collected fitness data for chemically induced (ethylmethane sulfonate, EMS) mutant lines descended from two Arabidopsis thaliana ecotypes that show differential adaptation to the local environment of our common garden plot. Using a novel nonparametric Bayesian statistical approach, we found that both ecotypes accumulated substantial proportions of beneficial mutations. The poorly adapted ecotype showed higher variance in the fitness effect of mutations than the well-adapted ecotype. Furthermore, we predict that it takes less than 4000 generations for the fitness space of the two ecotypes to overlap through mutation accumulation, and that a single founder, through mutation accumulation, is able to achieve the species-wide genetic variation in less than 10,000 generations. Our results provide evidence for relatively rapid local adaptation of Arabidopsis thaliana in natural conditions through new mutations, as well as the utility of nonparametric Bayesian method for modeling the distribution of fitness effects for field-collected data.