Computational protocol: Genetic Heterogeneity in a Cyclical Forest Pest, the Southern Pine Beetle, Dendroctonus frontalis, is Differentiated Into East and West Groups in the Southeastern United States

Similar protocols

Protocol publication

[…] Entire specimens were used for DNA extraction with the DNeasy DNA Extraction Kit (Qiagen, www.qiagen.com). Collected individuals (n = 1198) were screened at eight microsatellite loci following the methods detailed in Schrey et al. (). Microsatellite loci were briefly amplified by PCR (10 µL final volume), electrophoresed on an ABI 377 (Applied Biosystems, www.appliedbiosystems.com), and genotypes were determined using GENESCAN 3.2.1 and GENOTYPER v 2.5 (Applied Biosystems). Allele size data were binned after visualization on scatter plots. FSTAT version 2.9.3 () was used to test each locus in each geographic sample for conformation to Hardy-Weinberg equilibrium and to test all pairs of loci for conformation to linkage equilibrium.Bayesian analysis of population structure was performed among geographic samples of southern pine beetle using three software packages. First, BAPS version 5.3 () was used to cluster discrete samples into larger groups with and without geographic data. The presence of 1–27 groups was tested, with the most likely number of genetic groups and the samples constituting each group being identified. Second, TESS version 2.3.1 (; ) was used to characterize population structure among individuals. TESS estimates the number of populations (k) present among individuals and identifies individual membership in each k using a model-based clustering approach. Geographic coordinates were estimated for each individual from the geographic coordinates of each sample location and a pilot analysis was perfomed to confirm that 50,000 sweeps with a 10,000 step burn-in stabilized the likelihood. The preferred k was tested with five runs from k = 2-10. The preferred k was selected by comparing the DIC score and individual assignments. After selecting the preferred k, 100 replicate analyses were run at that k and summarized the runs with CLUMPP (). For every TESS run, 50,000 sweeps were used with a 10,000 burn-in and a fixed interaction parameter of 0.06 (). Third, the number of genetic groups among all individuals was estimated with STRUCTURE version 2.3 (; ).The admixture model was used with correlated allele frequencies, 10,000 burn-in steps and 50,000 post burn-in steps. The likelihoods of k = 1–5 groups were determined for four runs at each k by comparing the estimated natural log probability of observing the data (x) given the number of groups, In Pr(x|k). The most likely number of groups was identified by the test that maximizes In Pr(x|k). Individuals were assigned to groups by Q-values, which indicate the proportion of their genotype that originated from each group.The θST estimate of FST () was calculated among all geographic samples and pairwise among samples with FSTAT. GENALEX-6 () was used to perform a hierarchical AMOVA to partition genetic variation among samples within Bayesian clustering defined groups PhiPR and PhiRT. A Mantel test () was performed to compare pairwise genetic differentiation estimates (as OST /(1- θST)) to pairwise geographic distance (as log10 Euclidean distance in meters) with POPTOOLS (). Statistical significance was determined by 999 permutations.Genetic diversity estimates were calculated for each sample. Allelic richness and private allelic richness were calculated with HPRARE (). Observed heterozygosity, expected heterozygosity, and the inbreeding coefficient were calculated with GENALEX-6. Genetic diversity was compared among geographic samples and among groups defined by BAPS. All statistical tests were corrected for multiple tests using the sequential Bonferroni approach (). T-tests were used to compare genetic diversity estimates among genetic groups defined by Bayesian clustering. […]

Pipeline specifications