Genetic Footprints of Seasonal Fluctuating Selection: A Comparison With Established Selection Forms

Fluctuating selection is frequently studied in natural populations by observing allele frequency trajectories over evolutionarily brief timeframes or through theoretical analyses of frequency dynamics. However, little is known of its effect on linked neutral diversity. Here, we simulate seasonally fluctuating selection at a single locus and characterize its genomic footprint using diversity, site-frequency spectrum (SFS), and haplotype-based statistics. We find that adaptively fluctuating loci exhibit distinct signals that depend on when in the seasonal cycle the population was sampled and whether sampling occurs early or late relative to the onset of the cyclic selection pressure. Compared with other modes of selection acting on a single locus, adaptively fluctuating loci show genomic signatures that are not shared by hard and soft selective sweeps but overlap considerably with heterozygote advantage. Leveraging linear discriminant analysis, we identify a combination of summary statistics that most effectively distinguish fluctuating selection from heterozygote advantage. Our findings shed light on the characteristic genomic signatures of fluctuating selection and imply that in-depth analyses of the long-term dynamics of adaptively fluctuating loci is achievable from sequence data collected from one or two time points across a selection cycle.