Time-Series Sequences for Evolutionary Inferences.

Abstract

The application of genomics to evolutionary biology has provided unprecedented power and resolution to investigate processes like mutation (Nachman and Crowell 2000; Rozhok and Degregori 2019), genetic drift (Whitney and Garland 2010; Funk et al. 2016), gene flow (Gallego-García et al. 2019; LaCava et al. 2021), and natural selection (Brauer et al. 2016; Martins et al. 2018). Evolutionary genomics has historically used contemporary or single-timepoint samples to study microevolutionary processes that most often depend on idealized model assumptions (e.g., Wright–Fisher model) (Butlin 2008; Hoban et al. 2016). For example, genome scanning methods aim to detect natural selection by assuming that the impact of selection can be discerned from the effects of neutral evolutionary processes on genetic differentiation (Hoban et al. 2016). However, different demographic scenarios can also produce signatures of selection, leading to false positives (Lotterhos and Whitlock 2015; Haasl and Payseur 2016). Similarly, when using measures of genetic differentiation to estimate gene flow, most models assume that population size is large enough that genetic drift is negligible—that is, not driving neutral divergence between populations (Whitlock and McCauley 1999; Ma et al. 2015).