Genotyping-in-Thousands by sequencing (GT-seq) genotyped data used for panel validation in GSI

Single Nucleotide Polymorphism (SNP) panels are powerful tools for assessing genetic population structure and dispersal of fishes and can enhance management practices for commercial, recreational, and subsistence mixed-stock fisheries. Arctic Char (Salvelinus alpinus), Brook Trout (Salvelinus fontinalis), and Lake Whitefish (Coregonus clupeaformis) are amongst the most harvested and consumed fish species in northern Indigenous communities in Canada, contributing significantly to food security, culture, tradition, and economy. However, genetic resources supporting Indigenous fisheries have not been widely accessible to northern communities (e.g., Inuit, Cree, and Dene). Here, we developed Genotyping-in-Thousands by sequencing (GT-seq) panels for population assignment and mixed-stock analyses of three salmonids, to support fisheries stewardship or co-management in northern Canada. Using low-coverage Whole Genome Sequencing data from 943 individuals across source populations in Cambridge Bay (Nunavut), Great Slave Lake (Northwest Territories), James Bay (Québec), and Mistassini Lake (Québec), we developed a bioinformatic SNP filtering workflow to select informative SNP markers from genotype likelihoods. These markers were then used to design GT-seq panels, thus enabling high-throughput genotyping for these species. The three GT-seq panels yielded an average of 413 autosomal loci and were validated with an average assignment accuracy of 83.03%. Thus, these GT-seq panels are powerful tools for assessing population structure and quantifying the relative contributions of populations/stocks in mixed stock fisheries across multiple regions. Interweaving these genomic-derived tools with Traditional Ecological Knowledge will ensure the sustainable harvest of three culturally important salmonids in Indigenous communities, contributing to food security programs and the economy in northern Canada.

Methods
793 fish samples were genotyped using GT-seq panel developed in this study. After filtering (SNP with >50% missing data; samples with>30% missing data) the final datasets were converted in rubias format for each of the five (5) species/region datasets used in our paper for estimating assignment accuracies for GSI studies.