Phylogeography of Limia vittata (Cyprinodontiformes: Poeciliidae): geographical distribution of mitochondrial haplotypes is comparable to other Cuban poeciliids

This is the supplementary material associated with the article "Phylogeography of Limia vittata (Cyprinodontiformes: Poeciliidae): geographical distribution of mitochondrial haplotypes is comparable to other Cuban poeciliids", published by the Biological Journal of the Linnean Society (https://doi.org/10.1093/biolinnean/blad040). The following documents (pdf format) are included:

Supporting Information 1

Table S1. Sampling size (n) and haplotype (COI+CR) information for each sampling locality of L. vittata included in this study. Numbers correspond to those indicated in Figure 1. A star (*) represents a new locality record for L. vittata. ANC: acronym of the Acuario Nacional de Cuba Collection, La Habana, Cuba.

Supporting Information 2

Table S1. Prior distribution of parameters used for the analysis of the L. vittata data based on the ABC approach using DIYABC v.2.1.0 (Cornuet et al., 2014). Time is in generations. W-PR: Western Pinar del Río population, W-C: Western-Central population, C-E: Central-Eastern population.

Table S2. Summary statistics between L. vittata haplogroups for each scenario based on 3 × 106 simulated datasets. NHA: number of alleles, NSS: number of segregating sites, MPD: mean of pairwise differences, VPD: variance of pairwise differences, MP2: mean pairwise differences within samples, MPB: mean pairwise differences between samples, HST: FST between samples.

Table S3. Posterior distributions of the parameters based on scenario B for L. vittata. Data were obtained from 1% of the simulated dataset (1 × 106). Time is in generations. W-PR: Western Pinar del Río population, W-C: Western-Central population, C-E: Central-Eastern population. RMAE: relative median of absolute error.

Figure S1. PCA of the summary statistics of the observed dataset and the dataset generated from the prior distribution of parameters to evaluate the three biogeographic scenarios explaining the current distribution of L. vittata. The analysis was performed using DIYABC v.2.1.0 (Cornuet et al., 2014).

Figure S2. Posterior probability of the biogeographic scenarios tested to explain the current distribution of L. vittata. Logistic regression was used to compute the posterior probability using the ABC approach as implemented in DIYABC v.2.1.0 (Cornuet et al., 2014).

Supporting Information 3

Figure S1. Bayesian tree depicting the relationships of 160 partial COI+CR sequences of L. vittata. The red dots depict samples from eastern localities (Yateras, Sabanalamar, and Yacabo Abajo) sharing the W-C haplogroups (H9 and H10). Bootstrap support (≥ 93) and Bayesian posterior probabilities (≥ 0.95) are shown for the main clades.

Supporting Information 4

Figure S1. Mismatch distribution of pairwise haplotype differences from partial COI and CR sequences for the three haplogroups recovered in L. vittata. Dashed lines represent the distribution of the observed pairwise nucleotide differences, whereas solid lines represent the values expected after a population growth or decline model.

Supporting Information 5

Figure S1. Maximum Likelihood tree and haplotype network of Girardinus falcatus, Girardinus metallicus, and the Gambusia punctata species complex based on cytb partial sequences. Each colour represents a geographic region in Cuba. Green: westernmost Cuba, blue: western Cuba, orange: central Cuba, red: eastern Cuba. Numbers on tree branches are bootstrap values ≥ 95% and hatch marks on the networks are the number of mutations. n: number of individuals, H: number of haplotypes, h: haplotype diversity, π: nucleotide diversity, SD: standard deviation.

Table S1. Sampling size (n), GenBank accession numbers, and studies that made available the sequences of cytochrome b used for the phylogeographic comparisons of the different poeciliids in Cuba.