The genomic context for management of eastern oysters Crassostrea virginica in the Gulf of Mexico includes population structure and deep underlying divergence

Objective Eastern oysters Crassostrea virginica have been harvested commercially for well over 200 years in the United States, but worldwide declines in oyster populations have led to an increased focus on conservation of oyster species in general. Among the risks to extant oyster populations, genetic containment associated with human-mediated mariculture and stock enhancement is an important consideration due to the risk posed by introducing potentially exotic or invasive genotypes to wild populations. Knowledge of wild population structure can improve strategies for ensuring genetic compatibility of cultured and wild stocks. Methods Here, we examined genetic diversity and divergence among eastern oyster populations sampled from 386 individuals in 14 major bay systems throughout the Gulf of Mexico (Gulf), from southern Texas to southern Florida. We used a reduced representation genomic library consisting of 3,173 single nucleotide polymorphisms to make inferences about population structure and connectivity throughout the sampled range. Results The major findings of this work were (1) deep genomic divergence among four population clusters principally centered around the western Gulf, northern Gulf, northern Florida, and southern Florida, (2) a zone of admixture between highly diverged clusters in north and south Texas, and (3) high rates of intermediate genotypes observed in oysters sampled from northern Florida compared with the northern Gulf and southern Florida. Regarding this last finding, inbreeding coefficients (FIS) in the northern Florida population were sufficiently high and heterozygosity sufficiently low that contemporary admixture seemed an unlikely explanation for the observed pattern. Conclusions Taken as a whole, these results imply the presence of multiple discrete populations of eastern oysters in the Gulf of Mexico, and the geographic boundaries that separate them are roughly consistent with previously described studies. The presence of structured populations has implications for the selection of broodstock, and the knowledge of population boundaries will allow for more accurate assessment of risk when cultured stocks might harbor germplasm that is genomically inconsistent with wild populations in the area of culture.