Detecting fine-scale population structure in the age of genomics: a case study of lake sturgeon in the Great Lakes

Abstract

Great Lakes-wide population structure analyses using neutral markers have provided an understanding of broad-scale genetic structure of lake sturgeon. To assess the fine-scale genetic structure of lake sturgeon populations in two different rivers, we combined both microsatellites and genome-wide SNP markers. The St. Clair-Detroit River System (SCDRS) is entirely freshwater with no known impediments to sturgeon movement. In contrast, the St. Lawrence River (SLR) outlets into the Atlantic Ocean and is fragmented by hydroelectric dams. Both microsatellites and SNPs provided evidence of differentiation between the rivers. When applied to fine-scale structure, microsatellites failed to detect population structure using a Bayesian approach for within either river and FST values using microsatellites identified only a low level of differentiation between the upper and lower St. Clair River. Using the full set of SNPs for each comparison yielded similar results to the microsatellite results. Discriminant analysis of principal components using both markers partitioned the samples into spatially structured clusters. The SNP datasets filtered for high FST values had greater success for detecting fine-scale population structure and had the highest accuracy for reassignment to prior populations. This reduced SNP dataset may represent a more meaningful set of loci that can be used to estimate lake sturgeon fine-scale population structure, which is complicated by their long-generation times.