Genetic variability and the ecology of geographic range: A test of the central‐marginal hypothesis in Australian scincid lizards

Abstract

Whether they are big or small, species geographic ranges can be divided into core and edge populations. Verbal models from macroecology and population genetics predict that patterns of population abundance, individual fitness, and genetic diversity should differ across core and edge populations (Brown, 1984;Sagarin & Gaines, 2002), which can then contribute to the formation of range limits (García-Ramos & Kirkpatrick, 1997;Gaston, 2003).At the centre of their range, species are hypothesized to be optimally adapted for the habitat, allowing them to maintain large, interconnected populations characterized by high local abundance (Brown, 1984;Sagarin & Gaines, 2002). Moving away from the centre, the habitat becomes more marginal, leading to decreased reproductive output (Angert, 2006;Gaston, 2009;Pigott & Huntley, 1981). Or, favourable habitat may become increasingly patchy towards the range edge, leading to depression of regional abundance relative to range centres. Populations towards

Boundary effects cause false signals of range expansions in population genomic data

“…Tests for range expansions and estimation of the center of origin followed the same approach for the stepping stone models and the continuous space models unless otherwise stated. As in recent studies (Singhal, Wrath, and Rabosky 2022; Jaya et al 2022; He, Prado, and Knowles 2017), we used the original functions and pipelines from the R-package rangeExpansion (v.0.0.0.9000; Peter and Slatkin 2013, 2015; https://github.com/BenjaminPeter/rangeexpansion) to estimate ψ and the centers of origin. Some modifications of the original code were, however, necessary to correct some bugs, streamline the pipelines and improve computational speed.…”

“…Tests for range expansions and estimation of the center of origin followed the same approach for the stepping stone models and the continuous space models unless otherwise stated. As in recent studies (Singhal, Wrath, and Rabosky 2022;Jaya et al 2022;He, Prado, and Knowles 2017), we used the original functions and pipelines from the R-package rangeExpansion (v.0.0.0.9000;…”

“…Previous simulation work has shown clines in ψ to be more sensitive and robust to detecting signatures of REs and estimating their origins compared to methods based on clines in genetic diversity (Peter andSlatkin 2015, 2013), where the expansion origin is expected to be the location with the strongest positive correlation between geographic distance and ψ. This approach has been increasingly applied in recent population genetic studies as it only requires one diploid individual to be sampled per population (Zhan et al 2014;Maisano Delser et al 2019;Prior et al 2020;Fifer et al 2022;Hemstrom et al 2022;Lesturgie et al 2023;Singhal, Wrath, and Rabosky 2022;Walsh et al 2022;Ioannidis et al 2021;He, Prado, and Knowles 2017;Jaya et al 2022). A further potential benefit of the methodology introduced in Peter andSlatkin (2013, 2015) is the use of the Time Difference of Arrival (TDoA) -a ranging technique regularly used in the Global Positioning System (GPS) that allows for the inference of range expansion origins also from unsampled geographic regions.…”

“…Furthermore, the TDoA method relies on testing for correlations between two matrices (geographic distance and ψ) not accounting for the non-independence among the pairwise measures raising the possibility of severe p -value inflation (Peter and Slatkin 2015). Yet once the null-hypothesis of ψ≠0 has been rejected, finding significant correlations in TDoA have regularly been used to further support REs in population genomic data (Maisano Delser et al 2019; Jaya et al 2022; Peter and Slatkin 2015, 2013; Singhal, Wrath, and Rabosky 2022).…”

“…We further demonstrate that the mean |ψ|-values in equilibrium populations is proportional to the overall levels of genetic structuring in the data as measured by the fixation index ( F ST ) and test to what extent the levels of ψ, F ST and clinal variation ψ independently and jointly can predict genetic signatures of REs in the simulated data using linear models. These models were then tested on 28 empirical data sets of Australian scincid lizards with large numbers of significant ψ-values and strong geographic clines in ψ but with no known historical records of range expansions (Singhal et al 2022), on a previously analysed blacktip reef shark data from the Indian and Pacific Oceans (Maisano Delser et al 2019), as well as on data from the recent and rapid invasion of the cane toad in Australia (Selechnik et al 2019) that functioned as a positive control.…”