Figure 1.
Photographs, karyotypes and Europe-wide distribution of spined loach sexual species (Cobitis) from this study.
(A,C,E) Photographs (scale bar = 1 cm) and (B,D,F) respective karyotypes of three widespread Cobitis species. Karyograms with diploid chromosome number (2n), metacentric (m), submetacentric (sm), subtelocentric (st), and acrocentric chromosomes (a) were modified after Janko et al. [50]; (G) Sampling localities of Cobitis taenia (light gray squares; 1–10), C. elongatoides (dark gray squares; 11–20), C. tanaitica (black squares; 21–30), C. paludica (checkered square; 31), C. fahirae (spotted square; 32), and C. vardarensis (reticulated square; 33). Insets show European species distribution with respective markings as given in squares. Note that locality no. 1 is situated more eastward, as marked by the arrow.
Table 1.
Summary of nucleotide variation.
Figure 2.
Phylogenetic comparison of gene trees constructed from nuclear and mitochondrial gene markers and mito-nuclear discordance.
Bayesian DNA gene trees constructed from nine nuclear gene markers and one mitochondrial cytb gene marker were rooted with sequences from C. paludica. Haplotype numbers correspond to Table 2. Bar represents 0.1 substitution/site. The schematic tree shows the phylogenetic conflict of C. tanaitica topology between mitochondrial and nuclear gene markers.
Table 2.
Spined loaches (Cobitis) used in this study.
Figure 3.
Probability densities of four parameters in coalescence simulation as functions of length of internode.
Graphical visualization in which (1–PConcord-mtDNA) denotes the probability density of observing discordant mtDNA phylo tree; (Pbinom-nuc) denotes the probability density of observing eight topologically concordant nuclear gene trees out nine studied nuclear loci in total; (PCoal-mtDNA) denotes the cumulative probability of mtDNA coalescence along the internode and (PCoal-nuc) denotes the cumulative probability of coalescence of nuclear locus along the internode. Note that there is very small intersection of probability densities allowing for observing eight out of nine topologically concordant nuclear loci while having discordant mtDNA tree (see the text for details).
Figure 4.
Parameters calculated from alternative tests using sequence data to explain C. tanaitica mito-nuclear discordance.
Contemporary and ancestral population sizes are denoted by (θC.tae, θC.tan, θC.elo, θC.tae,C.tan, θC.tae,C.tan,C.elo). Divergence times are denoted by (τC.tae,C.tan and τC.tae,C.tan,C.elo), and interval between those times is denoted by (γ). Migration rates are denoted by (m) with relevant index. All parameters are scaled by mutation rate μ, and can be converted to absolute values using the relations θ = 4Nμ (where N is effective population size), m = m/μ (where m is gene-flow rates per gene copy per generation, τ = tμ (where t is a time of population splitting at τ generations in the past), and γ = tμ. Parameters estimated by BPP program are denoted by ($), those by IM by (#), and those by IMa2 by (@). The parameter γ was calculated from τs given by BPP and ds programmes. C. taenia (C.tae), C. tanaitica (C. tan), and C. elongatoides (C. elo).
Table 3.
Prior and posterior distributions of parameters in the BPP Bayesian analysis of the nine nuclear loci.
Figure 5.
Posterior probability distributions for migration rates from two-population IM analysis.
Coalescent-based estimates of migration rates (scaled by mutation rate) for three studied species inferred separately from (A–C) nuclear sequence data that included nine nuclear markers, and from (D–H) one mitochondrial marker gene.
Figure 6.
Posterior probability distributions for migration rates from three-population IMa2 analysis.
Coalescent-based estimates of migration rates (scaled by mutation rate) for three studied species inferred from (A–C) nine nuclear markers and (D–F) combined mito-nuclear sequence data that included one mitochondrial marker gene and nine nuclear markers.