Beacharrr The use of DNA variation for stock identification of Oncorhynchus keta 



621 



Principal component 1 



Figure 4 



Plot of first two principal components ( PC ) incorporating varia- 

 tion at the Ssa-A33 and Ssa-A34 loci and including Ssa- 1 band 

 counts for 39 stocks indicated in Figure 2. North = two stocks 

 from southeast Alaska and two stocks from northern British 

 Columbia. WCVI = west coast of Vancouver Island; ECVI = east 

 coast of Vancouver Island; Mainland = southern coastal main- 

 land of British Columbia. 



for Japanese chum salmon to specific stocks aver- 

 aged 15%, but 82% were correctly identified as Japa- 

 nese in origin (Table 8). Similarly, classification ac- 

 curacy for Russian chum salmon to stock of origin 

 was 30%, and to region of origin, 47%. Yukon River 

 chum salmon were fairly distinctive and the origin 

 of about 75%> of the fish was correctly identified. Ac- 

 curate classification for individual southern British 

 Columbia chum salmon to region of origin ranged 

 from about 9 to 44% (Table 8), presumably reflecting 

 less separation among these groups compared with 

 that among Japanese, Russian, and Yukon River 

 chum salmon. With Asian and North American stocks 

 pooled into two groups, accuracate classification to 

 continent of origin for 323 Asian fish was 78%, and 

 that for 797 North American fish was 94%. The 

 minisatellite DNA variation provided a reasonable 

 measure of accuracy in classifying individuals to con- 

 tinent of origin. 



Discussion 



Allele frequencies at the Ssa -A33 and Ssa -A34 loci 

 indicated that there were three regional groups of 

 chum salmon stocks: the Japanese, Russian and 

 Yukon River, and southeast Alaska and British Co- 

 lumbia stocks. The greatest differences were ob- 

 served between stocks from Japan and British Co- 

 lumbia. Similar results were also observed in a sur- 



vey of minisatellite DNA variation with the Ssal 

 probe (Taylor et al., 1994). The differentiation be- 

 tween stocks from Japan and British Columbia 

 has also been demonstrated in studies of varia- 

 tion at protein-coding loci (Okazaki, 1982) and 

 more recently in mitochondrial DNA in the re- 

 gion coding for NADH dehydrogenase subunits 

 5 and 6 (Park et al., 1993). These results are con- 

 sistent with the conclusion that stocks that are 

 more distant geographically also show more dis- 

 tinct genetic differentiation (Taylor et al., 1994). 

 These patterns presumably reflect the dispersal 

 of salmon from different refuges after the retreat 

 of glaciers at the end of the Wisconsinian glacial 

 period (Lindsey and McPhail, 1986). The separa- 

 tion between Japanese and Russian stocks re- 

 ported by Winans et al. ( 1994) in a survey of varia- 

 tion at protein-coding loci was also observed in 

 the current study with variation at minisatellite 

 loci. A high degree of genetic differentiation has 

 been consistently observed between Japanese and 

 North American chum salmon, regardless of 



Table 4 



Estimated percentage composition of a variety of mixtures of 

 chum salmon. Each mixture of 100 fish was generated 50 

 times with replacement, and a stock composition of the mix- 

 tures estimated by using either a fixed distribution of the 

 allele frequencies or band counts in the baseline stocks, or by 

 randomly resampling each baseline stock with replacement 

 to obtain a resampled distribution of the frequencies or counts 

 for each baseline stocks, with the same sample size in the 

 new distribution as in the original one. Band frequencies used 

 included those derived from genomic DNA restricted with 

 Haelll and hybridized with Ssal, pSsa-A33, and pSsa-A34. 

 Twelve (9 Japanese, 3 Russian) baseline stocks were used to 

 resolve the Japanese and Russian mixtures, and 8 (3 Rus- 

 sian, 5 Yukon ) baseline stocks for Russian and Yukon River chum 

 salmon mixtures. Standard deviations are given in parentheses. 



Distribution (%) 



Origin 



Actual 



Fixed 



Resampled 



Japan vs. Russia 

 Japan 

 Japan 

 Japan 

 Japan 

 Japan 



Russia vs. Yukon 



Russia 



Russia 



Russia 



Russia 



Russia 

 Andreafsky 





 25 

 50 



75 

 100 





 25 

 50 



75 



100 

 100 



0.0(0.0) 

 24.8(0.4) 

 49.5(0.6) 

 74.7(0.05) 

 99.6(0.6) 



0.1 (0.4) 

 24.1 (1.4) 

 48.3(1.5) 

 72.4(1.9) 

 97.5(1.8) 

 100.0(0.0) 



5.2(4.2) 

 28.3(4.3) 

 51.3(5.4) 

 74.5(4.8) 

 99.6(0.9) 



3.6(3.5) 

 28.2(3.6) 

 49.9(4.7) 

 72.5 (3.9) 

 96.0(3.7) 

 94.5(5.0) 



