Pacific Salmon: Ecology and Management of Western Alaska’s Populations

Population Structure and Stock Identification of Chum Salmon from Western Alaska Determined with Microsatellite DNA and Major Histocompatibility Complex Variation

Terry D. Beacham, Khai D. Le, Michael Wetklo, Brenda McIntosh, Tobi Ming, and Kristina M. Miller


Abstract.—Microsatellite and major histocompatibility complex (MHC) variation was surveyed to evaluate population structure and the potential for genetic stock identification in chum salmon Oncorhynchus keta populations largely bordering the eastern Bering Sea and northwestern Gulf of Alaska. Variation at 14 microsatellite loci and one MHC locus was surveyed for 59 populations in the study. The genetic differentiation index (Fst) over all populations and loci was 0.023, with individual locus values ranging from 0.007 to 0.058. At least 10 regional stocks were observed in the survey area, with populations from Kotzebue Sound (mean Fst value usually >0.02 in regional comparisons) and the Alaska Peninsula (Fst >0.03) the most distinct of Alaskan populations surveyed. For stock identification applications incorporating DNA variation, chum salmon sampled in rivers’ tributaries to the eastern Bering Sea were classified into the following regions: Kotzebue Sound, Norton Sound, lower/middle Yukon River, Tanana River, Kuskokwim River, Nushagak River, north/central Bristol Bay, southwest Bristol Bay, northern Alaska Peninsula/Aleutian Islands, southwest Alaska Peninsula, and southeast Alaska Peninsula. For simulated single-region mixtures, estimated regional stock compositions were generally above 90% for the previously-listed regions except for the Kuskokwim River, Nushagak and north/central Bristol Bay regions. Incomplete characterization of DNA variation for populations in those regions most likely contributed to the lower accuracies of estimated stock compositions in the simulated mixtures. Estimated regional stock compositions of simulated samples comprising fish from several regions were within 1–3% of actual values provided that no contributions from the three under-represented regions were included in the simulated samples. Microsatellite and MHC variation has the potential to provide accurate estimates of regional stock composition for chum salmon fisheries in the Bering Sea and northern Gulf of Alaska.