Natural selection and parallel clinal and seasonal changes in Drosophila melanogaster

Spatial and seasonal variation in the environment are ubiquitous. Covariation between a trait and space is often interpreted as the result of selection, but demography can complicate inference. Finding seasonal changes driven by selection is challenging, given effect sizes are small and the environment is subject to stochastic changes within seasons. Drosophila melanogaster is known to harbor polymorphisms that change with latitude and seasons. Identifying the role of selection in driving latitudinal and seasonal change is not trivial, but one promising approach is to look for parallel clinal and seasonal variation. Here, we test whether there is a genome-wide relationship between clinal and seasonal variation, and whether the pattern is consistent with selection. We investigate the role of linked selection in driving these patterns, and conservatively estimate the proportion of the variants that should be under selection to explain our observations. Allele frequency estimates were obtained from pooled samples from seven different locations along the east coast of the US, and 13 samples collected in the spring and fall in Pennsylvania. We show that there is a genome-wide pattern of clinal variation mirroring seasonal variation, which cannot be explained by linked selection alone. This pattern is stronger for coding than intergenic regions, consistent with natural selection. We find that the genome-wide relationship between clinal and seasonal variation could be explained by about 2.8% of the common variants being under selection. Our results highlight the contribution of natural selection in driving allele frequency changes with latitude and seasons in D. melanogaster.