Salmonids have emerged as important study systems for investigating molecular processes underlying parallel evolution given their tremendous life history variation. Kokanee, the resident form of anadromous sockeye salmon (Oncorhynchus nerka), have evolved multiple times across the species’ pan-Pacific distribution, exhibiting multiple reproductive ecotypes including those that spawn in streams, on lake-shores, and at lake depths >50 meters. The latter has only been detected in five locations in Japan and British Columbia, Canada. Here, we investigated the multiple origins hypothesis for deep-spawning kokanee, using 9,721 SNPs distributed across the genome analyzed for the vast majority of known populations in Japan (Saiko Lake) and Canada (Anderson, Seton, East Barrière Lakes) relative to stream-spawning populations in both regions. We detected 397 outlier loci, none of which were robustly identified in paired-ecotype comparisons in Japan and Canada independently. Bayesian clustering and principal components analyses based on neutral loci revealed six distinct clusters, largely associated with geography or translocation history, rather than ecotype. Moreover, a high level of divergence between Canadian and Japanese populations, and between deep- and stream-spawning populations regionally, suggest the deep-spawning ecotype independently evolved on the two continents. On a finer level, Japanese kokanee populations exhibited low estimates of heterozygosity, significant levels of inbreeding, and reduced effective population sizes relative to Canadian populations, likely associated with transplantation history. Along with preliminary evidence for hybridization between deep-spawning and stream-spawning ecotypes in Saiko Lake, these findings should be considered within the context of on-going kokanee fisheries management in Japan.

9021_SNPs_neutral.vcf: genotypic data for 167 individuals at 9021 SNPs (neutral only) generated via RADseq

9721_SNPs_all.vcf: genotypic data for 167 individuals at 9721 SNPs (all) generated via RADseq

2165_SNPs_neutral_SaikoOnly.vcf: genotypic data for 34 Saiko Lake individuals at 2165 SNPs (neutral only) generated via RADseq

9721_loci_sequences.txt: text file containing sequences for 9721 loci

metadata.txt: individual metadata including sample identifiers, location, and ecotype