Podarcis muralis

Data , genetic distances and phylogenetic reconstructions

A total of 803 base pairs (bp) from two mitochondrial gene fragments were obtained, with 39.9% variable sites in cyt b (155 bp, 35.3% parsimony informative) and 24.1% variable sites in 16S rRNA (100 bp, 20.9% parsimony informative). When excluding the outgroups, cyt b had 11.8% variable sites and 4.3% parsimony informative sites.

Focusing into the nuclear DNA dataset, a total of 3698 bp were obtained (Pod15b 515 bp, Pod55 405 bp, MC1R 635 bp, RAG1 856 bp, NKTR 587 bp and UBN1 700 bp). Out of the 515 bp of Pod15b 72 sites (13.9%) were variable and 48 sites (9.3%) were parsimony informative. The percentages of the variable and parsimony sites of the less informative nuclear markers are shown in supplementary table S4.

The pairwise genetic distances (p-distances) within the mtDNA dataset ranged from 0-24.2% for cyt b and from 0-13% for 16S rRNA. Regarding the nuDNA dataset, the range of pairwise distances was 0-3.31% for Pod55, 0-5.11% for Pod15b, 0-1.87% for MC1R, 0-2.45% for RAG1, 0-4.71% for NKTR and 0-5.74% for UBN1.

Focusing on the genetic distances among the major lineages of P. muralis as revealed by the phylogenetic trees, the pairwise genetic distances ranged from 1.51-6.52% for cyt b, 0.52- 1.76% for 16S rRNA, 0.02-0.27% for Pod55, 0-1.2% for Pod15b, 0-0.37% for MC1R, 0.2- 0.46% for RAG1, 0-0.73% for NKTR and 0.35- 2.39% for UBN1. According to the mitochondrial dataset, the highest genetic distance is observed between the lineage of Samothrace and Northeast Central Greece, and the lowest between the lineage of Peloponnese and Central Greece, whereas the highest genetic distance based on the nuclear dataset is observed between the lineages of Peloponnese and Northeast Greece and the lowest between the lineages of Northwest Greece and Central/Western Greece. The mean genetic distances between the major lineages of P. muralis as well as the distances calculated between P. muralis and the other Podarcis species included in the analysis are given in supplementary tables S5 and S6.

The nucleotide substitution models and the best-fit partitioning scheme selected for each phylogenetic analysis are shown in supplementary table S5.

The phylogenetic reconstruction based on cyt b, which aimed to clarify the phylogenetic position of south Balkan lineages of P. muralis in regard to the rest of the European lineages, revealed five very well supported southern Balkan lineages. These lineages, which correspond to distinct geographic regions of Greece, clustered with central European, Italian and Turkish lineages. Additionally, the lineage from Samothrace (north Aegean Sea) seems to group with Turkey’s lineages with high statistical support. The cyt b tree is shown in supplementary fig. S1 View Figure 1 .

Maximum Likelihood and Bayesian inference analyses produced fairly similar topologies for the mitochondrial ( figs. 3 View Figure 3 and 4b View Figure 4 ) and concatenated ( figs. 2 View Figure 2 and 4d View Figure 4 ) datasets. In all BI analyses, all MCMC diagnostic metrics indicated that all runs reached convergence and stationarity. The average standard deviation of split frequencies was 0.017 for the cyt b dataset, 0.006 for the mtDNA dataset and 0.01 for the concatenated one (when this value approaches zero, the tree samples are more similar), the plot of generation versus log-likelihood of the data had characteristic “white-noise” morphology after burn-in. In addition, for all parameters, the Potential Scale Reduction Factor (PSRF) values were 1.000 in both datasets and the minimum Effective Sample Sizes (ESS) values were well over 200 for all parameters in both datasets.

In all phylogenetic trees, P. muralis is more closely related to P. siculus specimens from the Italian group than to the rest of the Balkan species, verifying its prior classification as a member of Italian group. Considering the rest of Balkan species of Podarcis , all analyses produced trees with similar topologies and high statistical supports, distinguishing two major groups of species, which are in agreement with previously published studies; the group of P. erhardii with P. cretensis , P. levendis , P. erhardii , P. peloponnesiacus , and P. thais and the group of P. tauricus with P. ionicus , P. milensis , P. tauricus , and P. melisellensis ( Poulakakis et al., 2003, 2005a, b; Psonis et al., 2017, 2018, 2021; Spilani et al., 2019; Kiourtzoglou et al., 2021).

Considering the focal species ( P. muralis ), the mitochondrial tree, based on dataset 2 containing cyt- b and 16s RNA sequences, revealed seven lineages corresponding to seven geographical areas within the Greek territory that had high posterior probabilities, but relatively low bootstrap support ( fig. 3 View Figure 3 ); Northeast Greece, Samothrace , Northwest Greece, Northeast Central Greece, Central Greece, Peloponnese, and Central & West Greece. However , the phylogenetic relationships among these lineages remain unresolved. When examining the phylogeny retrieved from the concatenated dataset ( figs. 2 View Figure 2 and 4d View Figure 4 ), five distinct lineages were observed in Greece, in which the mtDNA lineages of Peloponnese and Central Greece were merge, as well as the mtDNA lineages of Central & West Greece and Northwest Greece. It should be noted however that the Northwest Greece lineage is represented by only three specimens. When looking at the trees produced solely using the nuDNA markers, three poorly (with the exception of the lineage of Samothrace) resolved lineages were revealed; Samothrace, Northeast Greece and a group including the rest of the mtDNA lineages (supplementary fig. S2 View Figure 2 and fig. 4c View Figure 4 ).

According to the Median Joining Network analysis, the mitochondrial markers revealed a clear structure with the exception of an overlap in Northeast Central Greece and Peloponnese noticed in the 16S rRNA marker. In the cyt b network, a high number of mutational steps is observed, as well as a great number of inferred (unsampled) sequences. Networks derived from nuclear markers do not indicate a clear structure of haplotypes, as the majority of samples group together. The haplotypes that seem to group apart more commonly are the ones found in the samples of Samothrace and Ukraine, as opposed to the Northeast Greece, Central Greece and Peloponnese samples that share many haplotypes. Haplotype networks are shown in detail in supplementary fig. S3 View Figure 3 .

Species delimitation and divergence time estimation

The analysis of STACEY (ln L = −10 743.69) identified all P. muralis lineages as a single species (supplementary fig. S4 View Figure 4 ). This agrees with the results of A11 analysis of BPP. The A00 analysis of BPP was specifically conducted to determine the parameters for species divergence times (τ s) and population sizes (θ s). The A00 analysis revealed an Inverse Gamma distribution (IG) which resulted in values of τ s ∼ IG (3, 1.5) and θ s ∼ IG (3, 0.02). These derived values served as the prior parameters for the subsequent A11 analysis of BPP. The A11 analysis results indicated that there were no posterior probabilities exceeding 0.95 within the P. muralis species, (Posterior probability = 1, prior = 0.073), suggesting the absence of distinct clustering. Divergence time estimation was performed using the distinct lineages that arose from the species delimitation, even those with low support, as distinct clusters. The intraspecific diversification of P. muralis within Greece was estimated at 1.93 Mya (early Pleistocene) ( fig. 5 View Figure 5 ).