Sensitivity of outlier loci detection methods to population range expansions: investigating the effects of population size, dispersal, and timing of sampling

Background/Question/Methods

Landscape genomics aims to identify correlations between spatial genetic variation and environmental features in natural populations, providing insights on adaptive evolution. Identifying loci potentially under selection can help us to better understand environmental drivers of evolution. However, in rapidly expanding or invading populations, neutral loci can exhibit spatial variation (e.g., clines) that resemble adaptive loci. Due to successive bottlenecks, drift, and allele surfing, neutral genetic variation can be difficult to distinguish from adaptive variation. Understanding the mechanisms underlying range expansion can provide insights into outbreaking and/or invasive species; however, detangling these processes is not trivial. Neutral clines can inflate false discovery rates, leading to potentially inaccurate conclusions regarding the role of adaptation in expansions and invasions. Although many tools exist to identify loci under selection (outlier loci), their performance in the context of range expansions has not been investigated. Here, we examine how five different outlier detection methods (LFMM, PCAdapt, outFLANK, RDAadapt and BayEnv) perform under strictly neutral simulated populations undergoing range expansion using individual-based, forward simulations in SLIM3. Specifically, we examine how often these methods incorrectly identify neutral loci as potentially adaptive in response to varying levels of: a) population size; b) reproductive rate; and c) dispersal capacity.

Results/Conclusions

Potentially adaptive loci were incorrectly identified detected in all simulation treatments using landscape genomics methods based on spatial trends in allele frequencies. In contrast, spatially agnostic, Fst-based genome scan approaches (e.g., outFLANK) identified few outliers, as expected in purely neutral systems. Range expansions and other complex spatial population dynamics pose unique challenges to identifying loci potentially under selection. These findings highlight the importance of considering the spatial demographic dynamics of expanding populations when exploring for loci under selection. Non-adaptive outlier loci can confound efforts aiming to better understand the role of adaptive evolution on the expansion or invasion success of outbreaking and invasive species.