Transatlantic Secondary Contact in Atlantic Salmon: Comparing Microsatellites, a SNP Array, and Restriction Site Associated DNA Sequencing for the Resolution of Complex Spatial Structure

Restriction site associated DNA sequencing (RAD-seq) represents a powerful approach for screening large numbers of genome-wide markers in non-model organisms, yet its utility in the context of population genetic inference remains unclear.  Here we compare RAD-seq, a single nucleotide polymorphism (SNP) array, and a microsatellite panel, to resolve spatial structuring under a complex scenario of post-glacial recolonization and possible trans-Atlantic secondary contact in Atlantic salmon.  Differentiation among samples was similar with SNPs (F_ST=0.099-0.080) and higher than microsatellites (F_ST=0.055). Multivariate analyses indicate that the three methods displayed significant (p<0.001) similarly in both scale and nature of structuring observed.  Bayesian clustering with each dataset supported two regional groups dividing the south coast.  SNP-based tests for selection identified outlier loci associated with these east and west groups.  Annotation of outlier SNPs support hypotheses of adaptive divergence, but also implicate a role for secondary contact.  Comparison with existing mtDNA data supports the hypothesis of eastern Atlantic affinities of salmon in southeastern Newfoundland.  This work highlights the utility of RAD-seq based approaches for the resolution of complex spatial patterns, further resolves a region of trans-Atlantic secondary contact in Atlantic salmon, and reinforces the difficulty in identifying loci associated with adaptive divergence under complex demographic scenarios.