Maintenance of genetic diversity of Atlantic salmon (Salmo salar) by captive breeding programmes and the geographic distribution of microsatellite variation

The capability of Finnish Atlantic salmon (Salmo salar L.) broodstock breeding programmes to maintain genetic diversity was assessed by comparing the levels of microsatellite diversity in wild and hatchery stocks in general, and in wild and hatchery derivatives of the same stock. The effective population sizes (N-c,) of the broodstocks and the ratio of effective size to census size (N-e/N-c) as well as the rate of loss of diversity in captive breeding were assessed. Moreover, the distribution and pattern of genetic diversity among Atlantic salmon stocks in the Baltic Sea, Barents Sea and NW Atlantic were measured. Microsatellite data were also compared with allozyme data of the same salmon stocks. Nine microsatellite loci were amplified: Ssa85, Ssa171, Ssa197, Ssa202, Ssa289, SSOSL85, SSOSL 311, SSOSL417 and SSOSL438 from 11 Atlantic salmon stocks. The effective population sizes of broodstocks were estimated with the method based on temporal allele frequency change. In short-term breeding programmes, the average rate of loss of heterozygosity was 1.4% per generation and the average observed rate of loss of alleles was 4.7% per generation. The estimated N-c's for the broodstocks were 32 and 238. The average N-e/N-c, ratio was 0.81. Changes in present-day broodstocks were not alarming and the N-e/N-c ratios were higher than in wild populations in general. The genetic D-A distance between continents was 0.64 (F-ST(theta)=0.22) and distances about half of that level (D-A distance 0.34, F-ST=0.09) were measured between European and Baltic Sea salmon stocks. A nearly diagnostic difference was observed in the SSOSL311 of the North American stocks (Maine 0.982 and Labrador 0.957) for a single allele (SSOSL311(118)) that did not occur in European populations at all. Microsatellite data showed relatively more genetic differentiation (F-ST=0.040) on a small geographical scale than did allozyme data (F-ST=0.017) indicating the higher discrimination power of this data set. (C) 2002 Elsevier Science B.V. All rights reserved.