Fig 1.
Image of Tennessee Ambystoma barbouri.
Photograph by David Pineros.
Fig 2.
Map of surveyed populations of A. barbouri in Tennessee and Kentucky.
County lines are shown with black borders. Shaded regions indicate major watersheds as depicted in the legend. Tennessee populations are colored by ESU assignment where green circles indicate northern ESU populations, yellow circles indicate central ESU populations, and purple circles indicate southern ESU populations.
Table 1.
Population IDs, sample sizes for mtDNA/GBS analyses, and map coordinates for all populations.
Population IDs for A. barbouri collections correspond to map locations in Fig 1. For outgroup samples, AAFB denotes Arnold Air Force Base and AKT1 denotes A. texanum from Arkansas. Population IDs by county are as follows: Bedford County (B6), Davidson County (D3), Rutherford County (R1, R7, and R9), Sumner County (S2, S5, S7, and S8), Wilson County (W1, W3, and W4), and Williamson County (Wil2).
Table 2.
Primers used for PCR amplification of the mitochondrial D-loop.
Primers THR (forward) and 651 (reverse) were used for initial amplification of entire ~1300 bp. Internal primers 007 (forward), DL3 (reverse), and DL1 (reverse) were used for Sanger sequencing (Shaffer & McKnight, 1996).
Fig 3.
Phylogenetic tree of A. barbouri mitochondrial haplotypes.
Maximum likelihood phylogenetic reconstructions of unique mitochondrial D-loop haplotypes from Ambystoma barbouri and A. texanum under a GTRCAT model of evolution as performed by RaxML. Bootstrap support values above branches are shown for nodes with 70% support or greater. Values below nodes indicate posterior probabilities from Bayesian reconstructions under a T92+G model of sequence evolution as performed by MrBayes. Asterisks denote accession numbers for sequences downloaded from Genbank.
Table 3.
Standard measures of genetic diversity for 12 populations of Ambystoma barbouri in Tennessee based 586 SNP loci.
Summary statistics include the average number of individuals genotyped per locus (N) and the proportion of SNPs that were polymorphic within each population (P). Allelic richness (AR), observed heterozygosity (HO), and expected heterozygosity (He) were calculated for all loci and again for only those loci that were polymorphic within each population.
Table 4.
Estimates of effective population sizes (NE) for five populations of Ambystoma barbouri in Tennessee based on the Linkage Disequilibrium (LD) method and the heterozygosity excess method as performed by Ne Estimator.
The 95% confidence intervals were estimated by the parametric chi-squared method.
Table 5.
Population pairwise FST estimates (above diagonal) and G”ST (below diagonal) averaged across 584 SNP loci for 12 populations of Ambystoma barbouri in Tennessee.
Table 6.
Results of hierarchical analyses of molecular variation (AMOVA) for the SNP dataset from 12 populations of A. barbouri in Tennessee.
Assignment to mitochondrial clades are as follows: Clade II (S5, S7, and S8), Clade IV (D3, W1, W3, and W4), and Clade III (Wil2, R1, R7, R9, B6). Asterisks indicate significance of Φ statistics based on 1000 permutations in Arlequin.
Fig 4.
Scatter plot from discriminant analysis of principal components (DAPC).
Scatter plots of PCA 1 and PCA 2 from DAPC analysis of SNP genotypes for A) all Tennessee A. barbouri (N = 198, K = 3), B) all Tennessee A. barbouri and samples from five populations of A. barbouri in Kentucky (N = 215, K = 3) and C) all A. barbouri and A. texanum (N = 4, K = 5). The optimal number of clusters for each analysis was determined using the Bayesian Information Criteria (BIC). The final number of principal components used in each analysis was determined using the cross-validation method.
Fig 5.
Results of Bayesian Assignment tests based on SNPs generated from GBS sequencing. Barplots indicate individual assignment probabilities for samples from 12 A. barbouri populations in Tennessee (K = 2–5).
Fig 6.
Multispecies coalescent analysis of SNP genotypes.
Species-tree (A) and divergence time estimates (B) from Bayesian multispecies coalescent method as implemented in the software BPP. Numbers at nodes also indicate divergence time estimates (t). Divergence time parameter τ was converted to time in years before present (YBP) in the BPPR statistical package in R-studio. Shading in tree indicates ESU assignments as follows: Northern ESU (green), Central ESU (Yellow) and Southern ESU (Purple).