Computational protocol: Postglacial range shift and demographic expansion of the marine intertidal snail Batillaria attramentaria

Similar protocols

Protocol publication

[…] All raw DNA sequences were aligned using AlignIR v.2.0 (LI-COR Inc., Lincoln, NE), and ambiguous basecalls and indels were manually adjusted. Then, these sequences were compared to COI sequences (507 specimens of 384 bp) from the Japanese population (GenBank Accession: No. AB164326 to AB164358) (Kojima et al. ). All statistics related to population genetic indices such as the number of mutations, parsimonious sites, and haplotypes were calculated in DnaSP v.4.0 (Rojas et al. ). In addition, various molecular density indices including haplotype diversity (h), the probability of two randomly chosen haplotypes being different (Nei ), nucleotide diversity (π) (Tajima ; Nei ), the mean number of pairwise differences (k), and their corresponding variances were calculated in Arlequin (Schneider et al. ). The population genetic structure of the Korean B. attramentaria was further investigated as follows. Differences in haplotype frequencies between populations were examined by the exact test of population differentiation (Raymond and Rousset ) in Arlequin v3.5.1.3 (Excoffier and Lischer ). The unbiased fixation indices of FST (Weir and Cockerham ) were estimated and tested for significance with 10,000 permutations. For hierarchical comparisons of the degree of genetic differentiation among populations, we assigned the populations to four groups according to their geographical locations, north to south and mainland to island (Table). That is, the populations of Dobido, Gonam, Byunsan, and Jindo were assigned to the YS group; the populations of Yeosu, Geoje, and Masan were assigned to the SS group; the populations of Shinchang, Hyubje, Hanlim, Samyang, and Sewha were assigned to the NJI group; and the populations of Shinyang, Kangjeong, and Pyosun were assigned to the SJI group (Fig. and Table). Tajima's D (Tajima ,) and Fu's FS (Fu ) were used to test the assumption of neutral evolution. To confirm the result of the neutrality test, we also conducted the mismatch test based on the frequency distributions of pairwise differences between sequences (Rogers and Harpending ; Ray et al. ; Excoffier ). We reconstructed a genealogy of the unique haplotypes of B. attramentaria and compared the clades to two previously identified divergent clades, referred to “Kuroshio” and “Tsushima,” from Japan (Kojima et al. ). Phylogenetic trees were constructed by the neighbor-joining method in MEGA v.4.0 (Kumar et al. ), and the sister batillariid species, Batillaria multiformis, was used as an outgroup (GenBank Accession: No. AB054364) (Kojima et al. ). The relationships among haplotypes were also examined through the network-tree method with two separate sets of data: the Korean B. attramentaria dataset (619 bp) and another pooled Korea–Japan dataset (337 bp). Network trees based on the most parsimonious connections of haplotypes were constructed in TCS v.1.21 (Clement et al. ). [...] To investigate the demographic history of B. attramentaria populations around the Korean peninsula and the Japanese archipelago, Bayesian skyline plot (BSP) analyses were conducted in BEAST v.1.7.5 (Drummond et al. ). These analyses were separately performed with genetically distinct geographic groups of populations identified in this study and in the previous study (Kojima et al. ) as follows: four Korean groups (YS, SS, SJI, and NJI) and two Japanese groups (Tsushima and Kuroshio). As explained above, we also excluded the three individuals with the Bat17 haplotype in the SJI. The HKY+I model of nucleotide substitution was selected for all groups through the Akaike information criterion (AIC) with PartitionFinder (Lanfear et al. ). Coalescent genealogies of BSP were constructed with a lognormal relaxed-clock model for a total of 100 million generations, sampling every 1000 steps. As the mutation rate of COI gene of B. attramentaria is not known, this uncertainty was taken into account in our BSP analyses such that a range of mutation rates (0.86–2.4% per nucleotide per million years) was used as an a priori uniform distribution of the mutation rate of COI. Convergence to the stationary distribution and sufficient effective sampling sizes for each estimated parameter were checked using Tracer v.1.5 (Rambaut and Drummond ). After discarding 10% of samples as burn-in, historical estimates of effective population size were summarized by Tracer based on the previously constructed coalescent genealogies. Finally, as a comparative method for the test of population expansion, we examined mismatch distribution analyses by applying a sudden population expansion model (Rogers and Harpending ) on the same six geographic groups in Arlequin (Excoffier and Lischer ). […]

Pipeline specifications