Genomic selection for low salinity tolerance in the eastern oyster Crassostrea virginica in Louisiana and Chesapeake Bay populations

Authors: SCHWARTZ, LINDSEY - University Of Louisiana At Lafayette; Plough, Louis; LAVAUD, ROMAIN - Louisiana State University Agcenter; LA PEYRE, MEGAN - Us Geological Survey (USGS); PENDELIDES, ANN FAIRLY - University Of Louisiana At Lafayette; MCCARTY, ALEXANDRA - Cherrystone Aqua-Farms; SMALL, JESSICA - Aquaculture Genetics & Breeding Technology Center, William & Mary; STAUFFER, BETH - University Of Louisiana At Lafayette

Submitted to: Aquaculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/20/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary: Eastern oyster (Crassostrea virginica) aquaculture has grown rapidly over the last two decades but is negatively impacted by environmental stressors such as extreme low salinity, which can cause high mortality in coastal regions. Selective breeding for low salinity tolerance can improve survival among families and the novel application of genomic selection (GS), which involves creation of a model linking individual, high density genotype data to phenotype (low salinity survival) allows for more accurate selection within families to select oysters for breeding. However, the potential for genomic selection (GS) models to improve low salinity tolerance has not been evaluated. Here we examined the potential for GS models to predict low salinity survival in lab -based salinity challenges using oysters from two populations: Chesapeake Bay and Louisiana. Tissue was sampled from both mortalities and survivors of the lab-based salinity challenges (2 ppt, 28°C) and were genotyped on a high-density genotyping platform designed for the eastern oyster (66K single nucleotide polymorphism, SNP, array). Genome-wide association analysis identified a region on chromosome 1 in the Chesapeake animals that significantly affected survival, while no such genetic region was discovered in analysis of the Louisiana animals. Genomic selection model type (i.e. regression model type) had no significant impact on prediction accuracy in the Chesapeake animals, but a large impact in the Louisiana animals. Likewise, prediction accuracy remained consistent in the Chesapeake Bay challenge with low density of genotypes (as few as 5,000 SNP markers), while accuracy in the Louisiana challenge began to decline with fewer than 25,000 SNP markers. This likely reflects the higher relatedness and family structure in the Chesapeake individuals. Overall, with moderate prediction accuracies in both populations, GS has the potential to improve low salinity tolerance in eastern oysters. This will be beneficial to selective breeding programs and restoration efforts in regions where this trait improves oyster survival.

Technical Abstract: The use of genomic selection (GS) to improve traits of interest in aquaculture species is growing with the availability of genomic resources for those species. The eastern oyster, Crassostrea virginica, is a major aquaculture product and keystone species along the U.S. mid-Atlantic and Gulf coasts. Populations and production of C. virginica are threatened by several environmental stressors, and in certain coastal regions, episodic low salinity conditions have contributed to mass mortality events. Here we examined the potential for genomic selection to improve low salinity tolerance (< 5) in eastern oysters and report the evaluation of GS models for C. virginica from two regions; Louisiana and Chesapeake Bay. Tissue samples from both mortalities and survivors of the lab-based salinity challenges (2 ppt, 28°C) were genotyped on the 66K high-density C. virginica SNP array. We performed a genome wide association study (GWAS) for both challenges and then evaluated the effects of several different parameters on GS prediction accuracy, including model type (GBLUP, Bayesian, and weighted), data filtration, Bayesian modelling parameters, and trait encoding (Gaussian, ordinal, and censored). GWAS revealed a significant region on chromosome 1 in the Chesapeake animals while the Louisiana animals showed a more polygenic architecture with few significant quantitative trait nucleotides. Genomic selection model type had no significant impact on prediction accuracy in the Chesapeake animals, but a large impact in the Louisiana animals. Likewise, prediction accuracy remained consistent in the Chesapeake Bay challenge with as few as 5,000 markers, while accuracy in the Louisiana challenge began to decline with fewer than 25,000 markers. This likely reflects the higher relatedness in the Chesapeake individuals. Overall, with moderate prediction accuracies in both regions GS has the potential to improve low salinity tolerance in eastern oysters. This will be beneficial to selective breeding programs and restoration efforts in regions where this trait improves oyster survival.