Computational protocol: High Levels of Genetic Connectivity among Populations of Yellowtail Snapper, Ocyurus chrysurus (Lutjanidae – Perciformes), in the Western South Atlantic Revealed through Multilocus Analysis

Similar protocols

Protocol publication

[…] The DNA sequences were edited and aligned in BIOEDIT v. 7.1.3.0 []. In the case of the nuclear loci, the individual heterozygotes were detected when double peaks were observed at the same position in both directions in the chromatograms. Heterozygotic events caused by indels were diagnosed through visual analysis of the chromatograms, with the alleles being reconstructed in INDELLIGENT v. 1.2. (http://imperialis.inhs.illinois.edu/dmitriev/indel.asp) [].The gametic phase of each nuclear marker was defined using the PHASE algorithm [], available in DNAsp v 5. 10.01 []. The runs consisted of 1,000 burn-in iterations and 1,000 principal iterations, with a thinning interval of 1. The algorithm was applied five times, with the fifth chain being ten times longer than the others. The haplotypes that returned a probability of less than 0.8 were excluded from the analyses.For the nuclear data, the minimum number of recombination events was estimated via the Rm method [], available in DNAsp v 5. 10.01 []. As the results of this analysis are strongly affected by homoplasy [], the significance of the number of recombination events was evaluated using the ФW test [], available in SPLITS TREE v. 4.12.6 []. Linkage disequilibrium was analyzed the EM algorithm in ARLEQUIN v. 3.5.1.3 [], which was run five times, with 20,000 permutations. [...] Determination of the number of polymorphic sites and the identification of possible stop codons (in the case of codifying regions) were performed in MEGA 6 []. The identification, quantification, and distribution of the haplotypes were determined in DNAsp v 5. 10.01 []. The genetic variability of the populations was evaluated based on the haplotype (h) and nucleotide (π) diversity indices [] obtained from ARLEQUIN v. 3.5.1.3 [].The haplotype network was generated using HAPLOVIEWER [] based on a maximum parsimony tree produced in DNAPARS, available in the package PHYLIP v. 3. 6 [], in accordance with Salzburger et al. [].The genetic homogeneity of the O. chrysurus populations was initially evaluated through an analysis of molecular variance (AMOVA) [] for each marker individually and subsequently through a multilocus approach, both with 10,000 permutations. This analysis permitted partitioning of the results into within- and between-population variation. In addition, Fst values [] were used to evaluate the gene flow between pairs of populations. These analyses were run in ARLEQUIN v. 3.5.1.3 [], with a subsequent adjustment of the p values using the False Discovery Rate test []. To verify the existence of isolation by distance, Mantel tests were performed using a matrix of genetic (Fst/(1-Fst)) and geographic (km converted to Ln) distances []. Negative Fst values were expressed as zero. These analyses were conducted in IBDWS (http://ibdws.sdsu.edu/∼ibdws) [], with 10,000 permutations.For comparison between Brazilian and Caribbean populations, the control region sequences used by Vasconcellos et al. [] (accession numbers EF624354—EF624359; EF624361—EF624373) were included in the network, AMOVA, and pairwise Fst analyses as well the Mantel test.Bayesian methods, using STRUCTURE v. 2.3.4 [], were applied to assign individuals to populations. This procedure places individuals into K clusters, where K is chosen in advance but can be experimentally varied throughout independent runs. K values between 1 and 8 were tested, using a model with admixture and no locprior (i. e., only genetic data is used for the assignment of individuals to a given K). For this analysis, only nuclear data were employed. Each run consisted of 1,000,000 steps (burn-in = 10%, and each value of K was implemented 10 times). The number of K was inferred by comparing the mean values of Ln Prob obtained in Structure Harvester (http://taylor0.biology.ucla.edu/structureHarvester/) [].Cluster analyses were also conducted in STRUCTURAMA []. For this analysis, the mitochondrial and nuclear data were grouped. The runs consisted of 2,000,000 generations, (burn-in = 20%). For the values of K, we employed the following distribution (K = expk (2)). The runs were summarized using the "showtogetherness" command.To check the fit of the historical population dynamics to a model of exponential growth, we used a mismatch distribution [] together with the SSD and raggedness indices. Mismatch analyses were conducted in DNAsp v 5 10 01 [], rates of SSD and raggedness, were implemented in Arlequin v. 3.5.1.3 [] based on 10, 000 permutations.Historic fluctuations in the demography of O. chrysurus were visualized using a Bayesian Skyline Plot (BSP) [] and an extended Bayesian Skyline Plot (EBSP) []. These procedures were run in BEAST v.1.7.4 [], based on evolutionary models suggested by JMODELTEST 2.1.1 [,] (HKY for Cytb, IGF 2, GH 5, ANT 1 and HKY+ I + G for CR). The analyses were based on the strict molecular clock used for the teleost control region, with a substitution rate of 10% per million years [,].Two runs were performed using different random seeds, including 200 million generations for each BSP run and 400 million for each EBSP run, with samples taken at intervals of 10,000 generations, 10% of which were discarded as burn-in. The convergence and mixing of the chains were inspected visually in TRACER v.1.5 []. The convergence and mixing were considered to be appropriate when all of the ESS values for each of the parameters analyzed were above 200.Tajima’s D [] and Fu’s Fs [] were also calculated, given that in addition to the detection of possible deviations from neutrality, these values may be used to evaluate demographic patterns, such as population expansion. These analyses were run in ARLEQUIN v. 3.5.1.3 [], with their statistical significance being assessed using 10,000 permutations. […]

Pipeline specifications