Population Genetic Analysis of Chinook Salmon

Population genetic tests for Hardy–Weinberg equilibrium, linkage disequilibrium, Analysis of Molecular Variance (AMOVA), average heterozygosity, and F-statistics were carried out using standard population analysis software applications bundled in GenAlEx v.6.5 [17 (link)] and ARLEQUIN v. 3.5.2 [18 (link)]. Tajima’s D statistic was measured using MEGA X [19 (link)]. This analysis involved 205 amino acid sequences. The coding data were translated assuming a genetic code table. All ambiguous positions were removed for each sequence pair (pairwise deletion option). Individual sequence reads were aligned utilizing the CLUSTAL-codon V algorithm bundled in the MEGA X software package [19 (link)] for each of four mapped MHC exonic regions captured by the two sets of primers summarized in Table 1. Haplotype sequence variation in MHC loci was analyzed for evidence of reduced polymorphism in nucleotide substitution spectra using the molecular evolutionary analysis programs in MEGA X. The results of the present Chinook salmon study were compared against Chinook salmon and other closely related salmonid species for the same MHC class I and II loci accessioned in NCBI/Genbank. The MHC primer sequences are designed to amplify the exons encoding the peptide binding region located on the α1 chain of the MHC class I and the β1 chain of the MHC class II. Sequences of cleaned PCR products were run through the BLAST program for annotation confirmation against the NCBI publicly available genomic database prior to being entered into the MEGA X program and aligned using the “MUSCLE” exonic sequence alignment toolkit. From these alignments, anchored phylogenetic trees were developed for MHC class I and II genes utilizing the Atlantic salmon as the anchored outgroup. Neighbor-joining trees were constructed using p-distance including all the amino acids, with bootstrap values obtained after 1000 iterations. Independent assessment of NJ tree results was tested using Maximum Parsimony methods of phylogenetic inference and bootstrapping.
STRUCTURE v 2.3.4 [20 (link)] was used to infer genetically distinct groups among the two populations. The program was run using the admixture model utilizing a burn-in length period of 10,000 iterations for 10,000 Markov chain Monte Carlo (MCMC) repetitions and testing for K (number of populations) between 2 and 7 for 10 simulations. The creation of minimum spanning and transitive consistency score (TCS) haplotype networks was conducted using the POPART (v. 1.7) software [21 (link),22 (link),23 ]. The TCS methodology uses nucleotide sequence data to infer population level genealogical networks, even in instances where divergence is low by collapsing closely related sequences into haplotypes allowing the program to estimate frequencies of haplotypes and probable outgroups [21 (link)]. A minimum spanning tree connects all present haplotypes without further inference of ancestral nodes, so that total length of branches is kept minimal [22 (link)].