Development of ‘Targeted’ Genotyping-by-Sequencing in Atlantic salmon (Salmo salar)

Selective breeding is a fundamental component of sustainable animal and plant production1. Traditionally, the aquaculture industry has lagged behind its agricultural counterparts with respect to utilisation of genomic technologies for animal breeding1. However, this gap has narrowed in recent years due to the advancement of next generation sequencing (NGS), which is often applied through reduced representation approaches to minimise costs1. Genotyping-by-sequencing2 (GBS), a popular reduced representation technology, employs restriction enzymes to sample a small proportion of the genome for variant discovery and genotyping. As a high-throughput approach, GBS represents a valuable tool to identify and concurrently genotype large numbers of genetic markers in thousands of samples; providing a multipurpose resource for association analyses, genetic diversity studies, pedigree reconstruction and genomic selection1. Consequently GBS and related technologies have underpinned the development of genomic resources for major aquaculture species1. Atlantic salmon (Salmo salar) are one of the most highly studied aquaculture species, their high economic value has fuelled large-scale genotyping studies to derive the genetic architecture underlying production-relevant traits1. A classic example is the identification of a single gene explaining 80-100% of genetic variation in resistance to infectious pancreatic necrosis (IPN), a viral disease causing devastating losses in the salmon farming industry3,4. Since the discovery of this quantitative trait locus (QTL), there has been a rapid decline in IPN outbreaks in European salmon farms, largely due to implementation of marker assisted selection (MAS) to select resistant animals for breeding5. Here we employed the GBS workflow in farmed Atlantic salmon, with the purpose of demonstrating the utility of low-depth GBS for association analyses and QTL mapping. By applying a statistical framework to account for genotype uncertainty6, GBS data with an average read depth of 4.15 was successfully used to identify the QTL for resistance to IPN disease. For traits of economic importance that are controlled by a small number of large-effect QTL, it is often desirable to employ targeted, deep sequencing of the associated genetic markers to ensure a high level of confidence in genotype calls. In addition to the restriction enzyme-based GBS workflow, we develop a targeted-GBS assay to specifically genotype key variants associated with Atlantic salmon production traits, based on the Genotyping-in-Thousands by sequencing methodology7 (GT-seq). Genotyping-in-Thousands by sequencing enables high throughout genotyping of relatively small panels of polymorphisms in thousands of samples by harnessing the power of NGS. A total of 12 genetic markers associated with resistance to IPN, age at maturity, and sex determination, traits that are all highly valuable to the salmon breeding industry, were included in the targeted assay. Validation of this assay was carried out in 376 Atlantic salmon breeding candidates. The average number of reads achieved per locus was 519 and all 12 variants were successfully genotyped in 98.7% of animals. We demonstrate the use of GT-seq in combination with GBS, effectively supplementing GBS data with specific large effect SNPs, representing a comprehensive workflow to obtain genomic information for the production of superior Atlantic salmon