Computational protocol: A signature of dynamic biogeography: enclaves indicate past species replacement

Similar protocols

Protocol publication

[…] We Sanger sequenced a 658 bp mtDNA fragment for 625 individuals (following []) and were able to sequence a 110 bp internal fragment for 23 individuals following []. The clear-cut geographical distribution of mtDNA allowed us to infer mtDNA type for the 16 remaining individuals (electronic supplementary material, table S1). The 658 bp mtDNA fragments were added to the mtDNA haplotype database of [] and collapsed into haplotypes with MacClade 4.08 []. To assign new haplotypes to species, we constructed a neighbour-joining phylogeny with 1000 bootstrap replicates in MEGA 5 [], with the Pyrenean newt Calotriton asper and the marbled newt T. marmoratus (from []) as outgroups (electronic supplementary material, figure S1; table S2). Internal 110 bp mtDNA fragments were aligned with the set of 658 bp haplotypes and this set was then trimmed accordingly. By removing redundancy in MacClade, the internal fragments could be allocated unambiguously to species-diagnostic mtDNA type. [...] We used Structure 2.3.3 [] to estimate the fraction of ancestry for each individual derived from the four parental species based on nuclear DNA data. We used the admixture model in combination with the correlated allele frequency model with 1 000 000 iterations, after 250 000 iterations of burn-in, and ran 10 replicates. An initial Structure analysis on the set of reference individuals, in which we allowed the number of genepools k to vary from 1–20 (with the upper limit defined by the total number of localities included), confirmed k = 4 as the most likely number of genepools under Evanno's Δk criterion [], as implemented in Clumpak []. Next we conducted a Structure run on the full set of individuals, in which we fixed k to 4. The lowest Q-value with which a reference individual was allocated to its respective species was Q = 0.9794. Localities allocated to two (or in some cases three) species with Q > 0.0206 (i.e. 1–0.9794) were considered genetically admixed. Structure scores (with replicates summarized in Clumpak) are presented in electronic supplementary material, table S1 and locality averages are plotted in b. [...] We determined the evolutionary origin of alleles based on our 15 reference individuals per species (see section on sampling). We found 23 nuclear DNA markers to be diagnostic, i.e. exhibiting fixed allelic differences between T. ivanbureschi (the species with the enclave) and the other three species. We determined the proportion of diagnostic T. ivanbureschi alleles present at each locality, i.e. the mean hybrid index (electronic supplementary material, table S1), and plotted this on a map (c). For pairwise species comparisons of T. ivanbureschi versus the other three Triturus species, we determined individual heterozygosity (the fraction of markers heterozygous for alleles from each parental species) and ancestry (the fraction of alleles derived from each parental species) using the R [] package ‘HIest’ []. Locality averages (electronic supplementary material, table S3) were plotted in Statistica 7 ( (). To test if introgression between species pairs was significantly asymmetrical we determined, for each locality, whether there was introgression of nuclear DNA alleles and what the fraction of introgressed alleles was (electronic supplementary material, table S3). We used a Fisher's exact test for the presence/absence data and a one-tailed Mann–Whitney U test for the quantitative data (electronic supplementary material, table S5). To avoid the confounding effects of gene flow from other Triturus species, we excluded localities showing ancestry of additional Triturus species (based on the Structure analysis or, in the case of low frequency introgression, inferred from the genetic composition of neighbouring Thiessen polygons; electronic supplementary material, table S4) from pairwise comparisons. Figure 3. […]

Pipeline specifications