Variation of Amplified Fragment Length Polymorphisms in Yukon River Chum Salmon: Population Structure and Application to Mixed‐Stock Analysis

Abstract

AbstractThe population structure of fall‐run Yukon River chum salmon Oncorhynchus keta has been studied previously using allozyme, microsatellite, and mitochondrial markers. However, genetically similar populations from tributaries near the U.S.‐Canadian border render mixed‐stock analyses (MSAs) difficult in the fisheries from lower portions of the Yukon River; MSA simulation apportionment estimates are less than 90% accurate for the border region divided by country of origin. To increase the accuracy and precision of contribution estimates to harvests in the Yukon River and to improve our understanding of the population structure of fall‐run chum salmon, we investigated the variation of amplified fragment length polymorphisms (AFLPs). Our results show that Yukon River chum salmon populations are structured by both seasonal race and geographic region. As expected, the MSA is most successful when mixtures are allocated to geographic regions. Both AFLP and microsatellites have better than 80% apportionment accuracy in MSA simulations for the U.S. and Canadian border regions, but neither approach clearly or consistently outperforms the other. In general, the population structure resolved by AFLP is similar to that observed for other genetic markers. Relatively weak population divergence, rather than shortcomings of the previously studied genetic marker systems, appears to be the limiting factor in attaining high levels of accuracy and precision in MSA.