Computational protocol: Population Genetic Patterns of Threatened European Mudminnow (Umbra krameri Walbaum, 1792) in a Fragmented Landscape: Implications for Conservation Management

Similar protocols

Protocol publication

[…] Microsatellite markers () and primers published by Winkler and Weiss [] were checked in GenBank before PCR and microsatellite analysis. Based on the available sequences, it was suggested that Umbra krameri microsatellites UkrTet2 (GenBank: FJ228219.1) and UkrTet3 (GenBank: FJ228220.1) sequences are identical, the reverse complements of each other. The DNA strand direction was mistakenly reversed for these. Data analysis of 156 fish samples (the first 14 populations) further strengthened this hypothesis. Allele numbers (n = 18) and allele frequencies, allele specific F st values (F st = 0.156) and the proportion of null alleles (prop. = 0.040372339) were equal for the UkrTet2 and UkrTet3 loci. Consequently, the UkrTet2 marker was left out of further statistical analyses.Fisher’s exact test of linkage disequilibrium and tests for deviations from Hardy-Weinberg equilibrium were conducted using GENEPOP 4.2.2 [] for each locus using a Markov chain of 10,000 dememorization steps, 20 batches, and with 5,000 iterations per batch. MICROCHECKER 2.2.3. [] was used to estimate the frequency of null alleles for each locus. Number of alleles, and locus specific F st values, were calculated using GenAlEx 6.5 []. Similarly, this software was used to estimate the mean of allelic richness (Ar), Shannon's Information Index (I), observed and expected heterozygosity (Ho, He), fixation index (F) and list the private alleles for each population. [...] The Lynch & Ritland [] estimator was used to calculate between-specimen pairwise relatedness. From this semimatrix, mean within-population pairwise values were calculated with 999 permutations, and 1,000 bootstraps in GenAlEx 6.5 []. Because in some cases null alleles were detected, their effect on pairwise F st values were corrected using the ENA procedure of Chapuis and Estoup [, ]. Genetic distances among populations were calculated using the Cavalli-Sforza and Edwards [] estimator after INA correction [], and presented in PCoA ordination. GenAlEx 6.5 [] was used to perform the analysis of molecular variance (AMOVA), with 999 permutations for population differentiation and hierarchical partitioning of genetic variation among and within regions and populations (F-statistics: F rt –H0: individuals are shuffled among regions, F sr –H0: individuals are shuffled within regions, F st –H0: populations are shuffled among regions, F is–H0: individuals are shuffled within populations, and F it–H0: individuals are shuffled in the whole sample).Genetic population structure was inferred using the hierarchical approach [] of the STRUCTURE analysis [] to estimate the most probable number of genetic groups (clusters, K) for all analysed individuals. Namely, the STRUCTURE analyses was first run including all samples, then samples were separated by river drainage basin, and finally by region. (A, B, C, etc.) Values of K were investigated from 1 to 20, with a burn-in period of 100,000 followed by 100,000 MCMC iterations and 10 runs for each K using an admixture model with correlated allele frequencies. Results of these Bayesian statistics were evaluated by STRUCTURE HARVESTER [], implementing the (deltaK) Evanno method []. Results of the 10 repetitions were combined using the software CLUMPP 1.1.2. []. For the genetic assignment of the studied individuals, Bayesian cross validation tests [] were carried out on drainage basin, regional, and population levels using GeneClass2 [] software. Cross validation tests were carried out similarly on the clusters defined at various levels by the hierarchical STRUCTURE analyses. […]

Pipeline specifications