Novel signals of adaptive genetic variation in northwestern Atlantic cod revealed by whole‐genome sequencing

Selection can create complex patterns of adaptive differentiation among populations in the wild that may be relevant to management. Atlantic cod in the Northwest Atlantic are at a fraction of their historical abundance and a lack of recovery within the Gulf of Maine has created concern regarding the misalignment of fisheries management structures with biological population structure. To address this and investigate genome‐wide patterns of variation, we used low‐coverage sequencing to perform a region‐wide, whole‐genome analysis of fine‐scale population structure. We sequenced 306 individuals from 20 sampling locations in U.S. and Canadian waters, including the major spawning aggregations in the Gulf of Maine in addition to spawning aggregations from Georges Bank, southern New England, the eastern Scotian Shelf, and St. Pierre Bank. With genotype likelihoods estimated at almost 11 million loci, we found large differences in haplotype frequencies of previously described chromosomal inversions between Canadian and U.S. sampling locations and also among U.S. sampling locations. Our whole‐genome resolution also revealed novel outlier peaks, some of which showed significant genetic differentiation among sampling locations. Comparisons between allochronic winter‐ and spring‐spawning populations revealed highly elevated relative (F(ST)) and absolute (d(xy)) genetic differentiation near genes involved in reproduction, particularly genes associated with the brain‐pituitary‐gonadal axis, which likely control timing of spawning, contributing to prezygotic isolation. We also found genetic differentiation associated with heat shock proteins and other genes of functional relevance, with complex patterns that may point to multifaceted selection pressures and local adaptation among spawning populations. We provide a high‐resolution picture of U.S. Atlantic cod population structure, revealing greater complexity than is currently recognized in management. Our genome‐scan approach likely underestimates the full suite of adaptive differentiation among sampling locations. Nevertheless, it should inform the revision of stock boundaries to preserve adaptive genetic diversity and evolutionary potential of cod populations.