The goal of selective breeding is to maximize the genetic gain per generation, while maintaining control of inbreeding. In 2012, Trygve Gjedrem published a review in the journal Aquaculture, alongside Nick Robinson and Morton Rye, highlighting the importance of selective breeding in enabling aquaculture to meet such demands. They argue that “aquaculture generally trails far behind plant and farm animal industries in utilizing selective breeding as a tool to improve the biological efficiency of production.”
The same argument can be made in the application of genomic tools to existing selective breeding programs. The use of genomic selection, for example, has had dramatic impacts in both the swine and dairy industries over the last 10 years, yet it has only recently started in aquaculture. It is clear, based on the success in terrestrial livestock, that genomic selection can elevate breeding strategies – and therefore, productivity – by increasing genetic gain per generation. Breeding programs using genomic selection will maximize the impacts of selective breeding and be better positioned to help fulfill aquaculture’s future as a major supplier of high-quality animal protein.
The concept of genomic selection was first published by Meuwissen et al. (2001) in their seminal paper, “Prediction of total genetic value using genome-wide dense marker maps.” They describe several methods to estimate the effects of segments of an animal’s genome (i.e. haplotypes) on a simulated phenotype and generate “genetic” estimated breeding values (GEBVs). In the other words, they devised a method to identify genome-wide genetic signatures that correlate with performance for a given trait. At that time, the use of gene chips – microarrays designed to genotype genetic markers such as single-nucleotide polymorphisms (SNPs) – was in its infancy and the authors noted that eventually they could render high-density genotyping cost-effective.
Fast forward to 2015, around one million Holstein cows are routinely being genotyped using the Bovine50K SNP chip and the Meuwissen et al. (2001) article has been cited ~150 times per year since 2003. During the first seven years of the application of genomic selection in the U.S., gains of 50 to 100 per cent in yield of milk, protein and fat were recorded in dairy cows.
In genomic selection, the key concept is that by using dense marker information one can achieve more accurate selection of elite breeders than when using only phenotypes and pedigree information, i.e. the estimated genetic superiority of an animal is more highly correlated with its true genetic merit when genomic selection is used.
There are a number of publications that demonstrate producers of aquatic species would also benefit from the use of genomic selection (e.g. Castillo-Juárez et al., 2015; Nielsen et al., 2009; Ragavendran and Muir, 2011; Sonesson and Meuwissen, 2009). For example, Nielsen et al. (2009) conducted a simulation of the application of genomic selection to an aquaculture breeding program and predicted that the accuracy of selection could be increased by 38 per cent when genomic selection is used compared to traditional methods for estimating breeding values in family-based breeding programs. Further, genomic selection is particularly powerful when working to improve traits that cannot be measured on candidate breeders, such as disease resistance and fillet quality. When assessing/measuring traits of interest, often siblings of the candidate breeders are evaluated for presence or absence of a trait, and the selection program assumes that the performance of a family member is predictive of that of another member. This approach has reduced accuracy because every family has its outliers. In other words, even if one family is observed to demonstrate higher performance relative to the general population, not all the members of the family will perform equally. The use of genomic selection allows an accurate identification of family members that are more likely to perform as well as those that are recognized as having the most desirable traits when tested.
Lastly, controlling inbreeding is also a key element of successful genetic improvement programs. It prevents strong selection pressure from reducing genetic diversity leading to production of animals that are not equipped to succeed when faced with new challenges or environments. Genomic selection allows for control of inbreeding at the genome level (Sonesson et al., 2012) by considering the actual similarity of the genome of family members rather than the expected similarity based on a pedigree.
View the embedded image gallery online at:
Genomic selection can help producers running established selective breeding programs accelerate genetic gains and become more sustainable and more productive. There are two main technical roadblocks preventing the wide use of genomic selection in aquaculture: lack of genomic resources and cost. Potential solutions for both issues lie, in part, in the technological advances in sequencing and genotyping technologies.
Sequencing genomes, and consequently developing genome-wide catalogs of SNPs, is less expensive than it used to be when the effort to sequence the Atlantic salmon (Salmo salar) genome began (Figure 1). The decreasing costs, together with the advent of long read technologies such as PacBio and 10X genomics, have allowed many new genomes from aquatic species to be sequenced; e.g. yellowtail, amberjack, coho salmon, American oyster and Pacific white-legged shrimp among others. With respect to genotyping – screening samples from individuals to characterize thousands of genetic markers – the use of SNP chip technology is now more accessible in terms of cost. However, advances that have been made in genotyping-by-sequencing (GBS) technology leave this tool poised to become a game changer for the development of commercially feasible applications of genomic selection in aquaculture.
GBS can be used to replace SNP chips as the source for high-density (HD) genotypic data (Gorjanc et al., 2015) or, it can be used in combination with SNP arrays for imputation-based genomic generation of HD genotypes (Hori et al., 2018). Imputation, the estimation of missing genotypes of a HD panel (50,000 SNPs) from a lower density panel (e.g. 2,000 SNPs), is a promising approach to reduce the cost of genomic selection for use in aquaculture.
At the Center for Aquaculture Technologies (CAT), we are currently investigating the use of targeted GBS panels combined with imputation to generate HD genotypes. The approach lowers the cost per sample and increases the number of selection candidates that can be genotyped. The research team has found that the accuracy of imputation from a ~2,000 SNP panel to a ~50,000 SNP panel is higher than 90 per cent in Atlantic salmon (Hori et al., 2018).
The use of such an approach can reduce the cost of genotyping for a genomic selection program by ~40 per cent when compared to using SNP chips alone to produce HD genotypes.
The future of genomic selection in aquaculture is bright and the Atlantic salmon industry is leading the way in its adoption. This is not surprising, given that Atlantic salmon is the commercial aquatic species with the most genomic resources and research available. However, to enable a wider application of genomic selection in aquaculture, alternative approaches that further reduce cost such as imputation of lower density panels to GBS HD data (Gorjanc et al., 2015), within-family estimation of GEBVs (Lillihammer et al., 2013), and pooling (Sonesson et al., 2010) must be evaluated. It is unlikely that a single approach will be the best fit for all aquatic species given that the economics of production are not equal for every industry.
The knowledge generated from the application of genomic selection in aquaculture has grown significantly in the last few years; a search for aquaculture and genomic selection in Pubmed returns an average of 40 publications per year since 2014. Still, despite the weight of evidence generated by research, the commercial application of genomic selection in the broader aquaculture industry still lags.
Dr. Tiago S. Hori is the director of genomics at the Center for Aquaculture Technologies. Hori holds a B.Sc in Biology, a M.Sc in Genetics and Biochemistry, and a Ph.D in Biology. His work focuses on understanding the architecture of traits relevant to the aquaculture industry. His research has included the use of both functional and structural genomics to investigate the molecular mechanisms driving such traits. Hori has published over 25 scientific articles in high-impact, peer-reviewed journals such as BMC Genomics, Physiological Genomics, Aquaculture and Molecular Immunology.