Moving GSI into the Next Decade: SNP Coordination for Pacific Salmon Treaty Fisheries

Numerous projects, for both SNP genotyping and discovery, have been funded by the United States Chinook Technical Committee and the Northern and Southern Endowment funds of the Pacific Salmon Commission, as well as by a number of other funding sources. SNP applications are flourishing in local jurisdictions where dozens of new SNPs have been implemented to greatly improve resolving power. This project coordinated current efforts in SNP genotyping and discovery for Chinook salmon within the Pacific Salmon Commission area of interest and provided guidelines for the development of a common panel of 96 SNPs for coastwide use in genetic stock identification (GSI). First, a comprehensive list of over 288 SNPs was compiled in collaboration with other Pacific Salmon Commission investigators. Genotypes for all 288 SNPs were collected from four representative populations from across the range. Results for the four populations were evaluated by interested laboratories for resolution and information content to identify a subset of 192 SNPs that would work well in all laboratories. Next, an additional 38 populations ranging from the Central Valley of California to Northwest Alaska were genotyped for the 192 SNPs. Using the data from all 42 populations, SNPs were ranked for information content, and panels constructed using three separate techniques and two separate reporting group scenarios. F_ST and f_ORCA analyses produced different sets of 95 loci each for the population and the two reporting group analyses, while the principal component analysis (PCA) produced a single set of loci. We compared the results from each of these methods to the results from a random selection of 95 loci, treating the random sets of 95 loci as null hypotheses. [Standard SNP panels consist of 96 loci. We evaluated panels of 95 loci to reserve one spot for a SNP designed to identify the sex of the individual fish.] We evaluated each of the seven sets of loci by comparing their GSI performance (i.e., the proportion of correct assignments) to each other and to the random sets of loci.

All sets, including the random sets performed well. From these results, it is clear that there are many sets of 95 loci that would achieve the needed resolution for GSI analyses and fishery management. Furthermore, it is conceivable that less than 95 SNPs will be needed for assigning individual fish to, or for mixed stock analyses of, reporting groups. There are at least two new studies that are attempting to optimally construct combined GSI and PBT SNP panels (CRITFC and WDFW 2012 CTC-LOA projects). We recommend that these two studies use the results of this project as a starting point from which combined PBT and GSI panels are constructed.