Computational protocol: Juvenile habitat partitioning and relative productivity in allochronically isolated sockeye salmon (Oncorhynchus nerka)

Similar protocols

Protocol publication

[…] Sockeye salmon fry DNA was isolated using WIZARD genomic purification kits (Madison, WI, USA). The adult DNA extraction protocol was a modified proteinase K and phenol:chloroform:isoamyl alcohol (PCI) technique, as described in . Fry and adult sockeye were genotyped at eight highly variable microsatellite loci. Microsatellite polymerase chain reactions (PCRs) were performed as described in , using the same microsatellite loci (). PCR products from the fry and parental fish were analyzed for molecular size (±0.5 bp) using an automated DNA analyzer; allele sizes were determined using the manufacture's software (Visible Genetics, Toronto, Canada) and verified by manual allele size identification. Approximately 5% of all PCR reactions were replicated to test for repeatability (97% of the alleles agreed across the two replicates, departures were typically due to single repeat size differences).An exact test for goodness of fit to Hardy–Weinberg equilibrium was conducted for adult early- and late-run fish in 1999 and 2000, using Arlequin version 3.11 (), and adjusted for significance using sequential Bonferroni correction. Population structure was evaluated for the 1999 and 2000 adults by calculating pairwise FST using TFPGA 1.3 (). An exact test for goodness of fit to Hardy–Weinberg equilibrium was conducted at all loci for the fry samples in both years (2000 and 2001) using the Monte Carlo method (20,000 permutations). The results of the Hardy–Weinberg test were adjusted for significance using the sequential Bonferroni correction ().We used genotype assignment to assign sockeye salmon fry (unknown) to their early- and late-run parental (source) population for the 2 years of adult-fry paired sampling (e.g., 1999–2000 and 2000–2001). The genotype assignment used a two-step process in Gene Class 2.0 (). First, we used the partial Bayesian method of to exclude fish having both assignment likelihoods below a 10% threshold. Next, fish were scored as “early” or “late” run fry using the rank-based assignment method with the criterion that the likelihood score of assignment must exceed 80% for the individual to be successfully assigned. One possible source of error for this analysis would be if the early- and late-run adult sockeye were misclassified due to errant run timing to the mouth of the Klukshu River (where they were sampled). However, the early run and late run were well differentiated temporally in both years with the DNA sampling separated by more than 4 weeks (). The mean straying rates between the early- and late-run adults were estimated to be less than 4% in 1999 and less than 7% in 2000 (), making it unlikely that parental misclassification contributed substantially to the assignment error.Once the fry had been assigned to early or late populations, we tested for global differences in the spatial distribution of early- and late-run fry across the seven lake sampling sites (; sites 1–7), using two-way crosstab chi-square analyses in 2000 and in 2001. We then tested for individual site deviations from the total early- and late-run fry proportions using two-way crosstab comparisons within each year separately. Since sockeye fry were only captured at site 8 (Klukshu River, ) in 2001, we excluded that site from our analyses. Finally, crosstab analyses was used to test for significant year-to-year (2000–2001) differences in the proportion of early- and late-run fry at each site separately.The relative productivity (i.e., production of fry) of each run was determined by comparing the proportion of adults (based on numbers at the counting weir) to the proportion of fry recruits (based on the assignment results) in both the early run and late run for the replicated adult-fry sample groups. Since our estimates of the absolute numbers of early- and late-run fry are based on a subsample of the fry present at each sample site, we used proportional comparisons of adult and juvenile abundance. Thus, for example, if the relative productivity of the early- and late-run adult sockeye were equal, we would expect our random sample of fry to generate the same proportion of early- to late-run juveniles as we calculated for the returning early- and late-run adult fish (this assumes equal reproductive success and incubation/fry survival). To test for differences in the proportional relative productivity between the early run and late run, the estimated numbers of early- and late-run fry versus adults were compared using a crosstab chi-square analysis in 2000 and 2001 separately. […]

Pipeline specifications

Software tools Arlequin, TFPGA
Application Population genetic analysis