Drift and Gene Flow in Shaping Spatial Patterns of Genetic Differentiation

21 Spatial patterns of neutral genetic diversity are often investigated to infer gene flow in wild populations. 22 However, teasing apart the influence of gene flow from the effect of genetic drift is challenging given that both 23 forces are acting simultaneously on patterns of genetic differentiation. Here, we tested the relevance of a 24 distance-based metric-based on estimates of effective population sizes or on environmental proxies for local 25 carrying capacities-to assess the unique contribution of genetic drift on pairwise measures of genetic 26 differentiation. Using simulations under various models of population genetics, we demonstrated that one of 27 three metrics we tested was particularly promising: it correctly and uniquely captured variance in genetic 28 differentiation that was due to genetic drift when this process was modelled. We further showed that (i) the 29 unique contribution of genetic drift on genetic differentiation was high (up to 20 %) even when gene flow was 30 high and for relatively high effective population sizes, and (ii) that this metric was robust to uncertainty in the 31 estimation of local effective population size (or proxies for carrying capacity). Finally, using an empirical dataset 32 on a freshwater fish (Gobio occitaniae), we demonstrated the usefulness of this metric to quantify the relative 33 contribution of genetic drift and gene flow in explaining pattern of genetic differentiation in this species. We 34 conclude that considering Isolation-by-Drift metrics will substantially improve the understanding of evolutionary 35 drivers of observed spatial patterns of genetic variation. 37 Genetic drift, a major evolutionary process compounded in small populations, is the random alteration of allelic 38 frequencies over generations. It ultimately leads to a local loss of genetic diversity and to an increase in genetic 39 differentiation among populations. Genetic drift is however often overlooked in spatial genetic studies, in which 40 measures of genetic differentiation are thus equated to gene flow only. In this study, we reviewed the distance-41 based metrics having been proposed to account for genetic drift. We then used simulations and an empirical 42 dataset to evaluate the relevance of these metrics. These distances-based metrics are based on estimates of (or 43 proxies for) population sizes, and is easily implemented in any regression-like analysis. We showed that these 44 metrics can be efficiently used to quantify the contributions of both gene flow and genetic drift in measures of 45 genetic differentiation in a large panel …