Quantifying the selection regime in a natural Chironomus riparius population

While the evolutionary fitness of natural populations is affected by a multitude of environmental factors, theory predicts that selective responses are in principle limited. However, we lack empirical knowledge on the magnitude of different selection pressures natural populations adaptively track. Here, we developed a framework to investigate the quantitative and qualitative complexity of the effectively acting selection regime using population genomic time series data. We applied the approach to a natural population of the multivoltine midge . Using six seasonal samples over three years from the same natural population, we could show with fitness experiments that the population continuously evolved in response to a highly variable environment. Analyses of genome-wide allele-frequencies revealed that tens of thousands of haplotypes responded at least once to selection during the monitored period. Clustering the temporal haplotype frequency trajectories revealed 46 different patterns, i.e. selection pressures. Some of these co-varied with measured environmental variables known to be selective factors for the species. Our results demonstrate that 1) adaptive tracking of multiple fluctuating selection pressures occurs in natural populations, 2) the estimated minimum number of simultaneously acting selective pressures is quite high but appears to be limited and 3) changes in intensity and direction of selective responses can be frequent. This shows that adaptation in natural populations can be rapid, pervasive and complex