Combining RADseq and contact zone analysis to decipher cryptic diversification in reptiles: insights from Acanthodactylus erythrurus (Reptilia: Lacertidae)

Abstract

Linnaean and Wallacean shortfalls (Uncertainties on species taxonomy and distribution, respectively) are major factors hampering efficient conservation planning in the current context of biodiversity erosion. These shortfalls concern even widespread and abundant species in relatively well-studied regions such as the Mediterranean biodiversity hotspot which still hosts a large fraction of unrecognised biodiversity, notably in small vertebrates. Species delimitations have long been based on phylogenetic analyses of a small number of standard markers, but accurate lineage identification in this context can be obscured by incomplete lineage sorting, introgression or isolation by distance. Recently, integrative approaches coupling various sets of characters or analyses of contact zones aiming at estimating reproductive isolation (RI) have been advocated instead. Analyses of introgression patterns in contact zone with genomic data represent a powerful way to confirm the existence of independent lineages and estimate the strength of their RI at the same time. The Spiny-footed Lizard Acanthodactylus erythrurus (Schinz, 1833) is widespread in the Iberian Peninsula and the Maghreb and exhibits a large amount of genetic diversity, although the precise number and distribution of its genetic lineages remain poorly understood. We applied a RADseq approach to obtain a genome wide SNPs dataset on a contact zone in central Morocco between the previously described Rif and Middle-Atlas lineages. We show that these two lineages exhibit strong RI across this contact zone, as shown by the limited amount and restricted spatial extant of gene flow. We interpret these results as evidence for species-level divergence of these two lineages. Our study confirms the usefulness of RADseq approaches applied on contact zones for cryptic diversity studies and therefore to resolve Linnaean and Wallacean shortfalls.

Phylogeographic breaks and how to find them: An empirical attempt at separating vicariance from isolation by distance in a lizard with restricted dispersal

“…However, several of the inferred barriers to gene flow were not unambiguously supported, highlighting a need for further investigations. Surprisingly, this was the case for Rif and Middle-Atlas, two groups that show very little to no traces of admixture at their The red star highlights the contact zone between the Rif and Middle-Atlas lineages, that is inconclusive in our analysis but is identified as a species-level barrier in Doniol-Valcroze et al (2022).…”

“…Using linear regressions, we could identify at least four geographic groups separated by barriers to gene flow (summarized in Figure 5, discussed in-depth in the next section), within which we could not reject the null hypothesis that populations differentiate through IBD. It is interesting to note that our results suggest that similar levels of mitochondrial divergence can arise from both IBD and allopatric divergence (and even allopatric speciation, as highlighted at the Rif/Middle-Atlas transition by Doniol-Valcroze et al, 2022). In that context, interpreting mitochondrial splits as evidence for phylogeographic breaks is potentially misleading and biased toward an over-estimation of the number of phylogeographic lineages.…”

“…Surprisingly, this was the case for Rif and Middle-Atlas, two groups that show very little to no traces of admixture at their contact zone, near Ifrane, where individuals belonging to either group occur 1 km apart. An in-depth analysis of this contact zone based on genome-wide data byDoniol-Valcroze et al (2022) showed a very steep transition, demonstrating that the Middle-Atlas and Rif are strongly differentiated groups with restricted admixture, consistent with the existence of intrinsic barriers to gene flow see Doniol-Valcroze et al (2022)…”

“…For example, it seems that the inconclusive results for the Rif/Middle-Atlas tests could be due to the asymmetric sampling around the contact zone, with only a few samples available in the former group, and a large sampling gap between the contact zone and the core area. In-depth analyses of this transition by Doniol-Valcroze et al (2022) also identified sampling gaps as a limiting factor, even when considering genome-wide data. Likewise, our IBD tests for coastal populations were very limited because of “lineages” represented by single localities, which made thorough comparisons of within-groups and between-groups distances impossible.…”

“…N-Atl, from Souk-Tnine to Larache); 2) the northern part of the Rif mountains (N-Rif, from Martil to Ketama); 3) the southern part of the Rif (S-Rif, Tazzeka and the North of Ifrane); 4) the coastal population of Moulay-Bousselham (M-Bou); 5) the coastal populations of Blad-Dendoun and Mehdia (Blad-Mehd); 6) the inner-most parts of the Middle-Atlas mountains (Mid-Atl1, around Azrou and Ifrane); 7) the rest of the Middle-Atlas mountains (Mid-Atl2, from Oulmès to Ouaourioud and Boumia); 8) the population of Sidi-Yahia (S-Yah); 9) the coastal population of Sidi-Boumaine (S-Bou); 10) the populations surrounding Marrakech (Mara; including Sidi-Chiker and Timlite) and 11) the Southern Atlantic coast (S-Atl; from Ouled-Ghanem south to Essaouira).All lineages are geographically isolated by sampling gaps, except for S-Rif and Mid-Atl1 which come into contact north of Ifrane (Middle-Atlas). Of the three samples from this locality, two cluster with the former and one with the latter (seeDoniol-Valcroze et al 2022 and FigureS1).The 11 lineages cluster into three main clades: i) Mara, S-Atl, S-Bou; ii) Rif N, Rif S (as well as the Iberian samples) and iii) N-Atl, M-Bou, Blad-Mehd, Mid-Atl1, Mid-Atl2. The aLRT support for these clades ranges from 77 to 100, but the relationships among them remain unresolved.Finally, S-Yah harbors a remarkably divergent mitochondrial lineage which is basal to all other Moroccan mitochondrial clades of A. erythrurus.…”