Abstract

We genotyped nuclear and mitochondrial single nucleotide polymorphisms (SNPs) in six paired archived and contemporary collections of Alaskan sockeye salmon (Oncorhynchus nerka) to evaluate the stability of allele frequencies over 25-42 years (4.9-8.4 generations). First, our results show that temporal changes were dramatically (between 40- and 250-fold) smaller than spatial differences in allele frequencies when based on nuclear SNPs. Second, the magnitude of temporal change was consistent with a model of genetic drift: (i) SNPs with high levels of differentiation (large theta) and candidates for diversifying selection were not more likely to show significant temporal changes than small-theta SNPs; and (ii) the fraction of single-locus significant tests was consistent with theoretical predictions relating sample size and the annual number of breeders (N-b). Third, estimates of N-b were bound by infinitely large upper 95% confidence intervals, except for one paired collection with unique life-history attributes of both a smoltification phase and generation time shorter than the other paired collections. Use of multigenerational SNP data sets seems a safe practice in management of Alaska sockeye salmon that could be extended to other large, wild aquatic populations.