The combined use of raw and phylogenetically independent methods of outlier detection uncovers genome‐wide dynamics of local adaptation in a lizard

Abstract

Local adaptation is a dynamic process by which different allele combinations are selected in different populations at different times, and whose genetic signature can be inferred by genome‐wide outlier analyses. We combined gene flow estimates with two methods of outlier detection, one of them independent of population coancestry (CIOA) and the other one not (ROA), to identify genetic variants favored when ecology promotes phenotypic convergence. We analyzed genotyping‐by‐sequencing data from five populations of a lizard distributed over an environmentally heterogeneous range that has been changing since the split of eastern and western lineages ca. 3 mya. Overall, western lizards inhabit forest habitat and are unstriped, whereas eastern ones inhabit shrublands and are striped. However, one population (Lerma) has unstriped phenotype despite its eastern ancestry. The analysis of 73,291 SNPs confirmed the east–west division and identified nonoverlapping sets of outliers (12 identified by ROA and 9 by CIOA). ROA revealed ancestral adaptive variation in the uncovered outliers that were subject to divergent selection and differently fixed for eastern and western populations at the extremes of the environmental gradient. Interestingly, such variation was maintained in Lerma, where we found high levels of heterozygosity for ROA outliers, whereas CIOA uncovered innovative variants that were selected only there. Overall, it seems that both the maintenance of ancestral variation and asymmetric migration have counterbalanced adaptive lineage splitting in our model species. This scenario, which is likely promoted by a changing and heterogeneous environment, could hamper ecological speciation of locally adapted populations despite strong genetic structure between lineages.

Environmental association modelling with loci under divergent selection predicts the distribution range of a lizard

“…The outlier analysis performed with Bayescan detected 12 outlier loci with α > 0 (0.97 < α < 1.35) and q < 0.05, while the FLK analysis identified nine additional loci with p < 0.001, none of which was previously detected by Bayescan. An MDS analysis performed elsewhere with these 21 SNPs putatively under selection, placed sampled populations along a first major axis that recovered the same pattern of differentiation observed for mtDNA and for some relevant phenotypic traits (Llanos-Garrido et al, 2019). Moreover, these phenotypes were interpreted as adaptive after a process of ecologically-driven divergence in a much wider sample (Díaz et al, 2017).…”

“…By fulfilling all grid cells with those environmental values, we could extrapolate our prediction to the overall distribution range of the species. Finally, we removed from the inferred range a few disconnected patches (in central France, coastal Italy, and the Mediterranean islands) that were too far from the main distribution range of P. algirus (>8 km from the nearest inferred distribution limit, which was the largest distance between adjacent but disconnected patches within the inferred range), whose low dispersal rate (Santos et al 2009) is supported by the fact that genetic differentiation can be detected even among forest fragments separated by 350 m of unsuitable arable land (Pérez-Tris et al 2019). This produced our second (and final) inferred distribution range (predicted range #2).…”

“…The two other sampling sites had populations of the western lineage: 4) El Pardo (P; 40.511 ºN, -3.755 ºE; 658 m asl; N = 19), a xeric, lowland evergreen forest, and 5) Navacerrada (N; 40.726 ºN, -4.023 ºE; 1,230 m asl; N = 20), a montane location covered by deciduous forest. Several particularities of these populations make them representative of a wide range of selective pressures gathered around the core of this species' range: 1) lizards from Lerma inhabit a very fragmented forest archipelago that resembles the typical habitat of western lizards (although they belong to the eastern lineage; see Díaz et al 2005;Santos et al 2008;Telleria et al 2011;and Pérez-Tris et al 2019 for further information about habitat fragmentation effects in this system); 2) Aranjuez lizards inhabit the typical hot and dry habitat of eastern lizards, and although this locality is very close to the western populations included in this study (El Pardo and Navacerrada), it receives very little gene flow from them (Díaz et al 2017), so that its isolated condition promotes the accumulation of genetic divergence subject to selection (Llanos-Garrido et al 2019); and 3) the two western populations (El Pardo and Navacerrada) are separated by a significant altitudinal gradient, and although lizards from both populations show little genetic differentiation (Díaz et al 2017), they differ in important phenotypic traits such as escape tactics, sexual dimorphism, sexual ornaments, ectoparasite loads and other life history traits (Iraeta et al 2006(Iraeta et al , 2010(Iraeta et al , 2011Llanos-Garrido et al 2017).…”

“…Our intra-population randomization approach showed that the predictive power of GEAM was much larger than expected by chance. Such high predictive power was based on the genetic diversity found in five closely-located populations with geographical distances that were not correlated with either genetic or environmental distances (Llanos-Garrido et al 2019). Yet, a small number of inter-population randomizations and randomizations by neutral loci also yielded significant models, as expected from a certain degree of environmental pseudoreplication and genetic aggregation in our data.…”

Environmental association modelling with loci under divergent selection predicts the distribution range of a lizard

“…This lizard is widespread across the Western Mediterranean region, and its range encompasses contrasting environmental conditions, extending from northern Africa in the south to southwest France in the north, and from Portugal in the west to Tunisia in the east (Figure 1a). We used 21 loci putatively under selection (hereafter outliers; Llanos-Garrido et al, 2019) to model distribution boundaries on the basis of five closely located central populations that cover a representative fraction of the environmental variation faced by P. algirus across its entire distribution range. These outlier SNPs were identified by two methods of outlier detection, one of them independent of population coancestry (an extension of the Lewontin-Krakauer test; Bonhomme et al, 2010) and the other one not (bayescan v.2.1; Foll & Gaggiotti, 2008).…”