Computational protocol: Impacts of biogeographic history and marginal population genetics on species range limits: a case study of Liriodendron chinense

Similar protocols

Protocol publication

[…] Unique combinations of cpSSR alleles within a chloroplast genome were defined as haplotypes. The number of haplotypes (NA), effective number of haplotypes (NE), the haplotype richness (HR), and the haplotype diversity (H) were calculated using the program GenAlEx 6.5. Since these genetic diversity parameters were highly correlated, we performed a principal component analysis (PCA) in SPSS 13.0 to extract the genetic diversity indices as a single vector (following Garner et al.; PC1 eigenvalue = 3.79, variance explained = 94.7%, ) and later map these estimates.The partitioning of genetic variation was examined by hierarchical analysis of molecular variance (AMOVA) under 10 000 permutations in Arlequin version 3.1. The standardized measure of genetic differentiation G’ST was calculated in MSA 4.05.The presence of phylogeographic structure was determined by whether the value of RST (considering the mutational distances between haplotypes) was significantly higher than the GST (depending on frequencies of haplotypes) under 1000 random permutations in PermutCpSSR version 2.0.A haplotype neighbor joining (NJ) tree was constructed based on 1 000 bootstraps of Nei’s distance with L. tulipifera included as the outgroup using PowerMarker v3.25. Intraspecific genealogies of haplotypes were constructed using a median-joining network in Network 4.610 with all length polymorphism in microsatellite sites included and weighted equally. The demographic history of L. chinense was assessed by neutrality tests with Tajima’s D and Fu’s Fs and mismatch distribution analysis (). Chloroplast microsatellite data were coded as “DNA” type according to Pereira et al. by their nucleotide repeat numbers. Tests were examined in two levels: the whole range and each of the six putative refugia regions. [...] Genotyping errors were identified by the software Micro-checker (available at http://www.norwichresearchpark.com/ourresearch/researchgroups/elsa/software/microchecker.aspx). Linkage disequilibrium (LD) and Hardy-Weinberg equilibrium (HWE) for each population and microsatellite locus pair was assessed in Genepop version 1.2. Population genetic diversity parameter such as effective number of alleles (NE), allelic richness (AR), expected genetic diversity (HE) and observed genetic diversity (HO) were calculated using the software GeneAlEx 6.5 and FSTAT version 2.9.3.The extraction of genetic diversity and estimation of population differentiation were performed as described in cpSSR markers (PC1 eigenvalue = 4.51, variance explained = 90.1%, ). The population genetic structure of L. chinense was determined using STRUCTURE v2.3.3. An admixture ancestry model with correlated allele frequencies and no prior information of population origin was used. Ten independent runs for each K (1–33) were carried out by 50 000 burn-in and 100 000 iterations of the MCMC. The suitable number of clusters (K) was chosen at the largest rate of change in the log probability of data between successive K values.An additional cluster analysis was performed with TESS v2.0, which detects spatial correlation between individuals. In contrast to STRUCTURE, it assumes that all of the individuals are equally unrelated. The program was run with 100 replicates under the admixture (BYM) model, using 50,000 MCMC sweeps after a burn-in of 10,000 for K = 2 to 29 and with a spatial interaction parameter of 0.6. The optimal K for the data is at the point where the deviance information criterion (DIC) curve switches from a sharp decrease to a plateau. The 20 runs with lowest DIC scores were combined using software CLUMPP and the graphical representation was visualized using DISTRUCT 1.1.Extended Bayesian Skyline Plots (EBSPs) were used to test the hypothesis of demographic expansion at Yungui Plateau. The EBSPs were constructed in BEAST v1.80, and settings were referred to Teske et al.. The mutation rate of microsatellites was set as 5 × 10−4 per locus per generation. Two independent runs were carried out for a chain length of 1 × 108 and a logging frequency of 1 × 106. […]

Pipeline specifications