Confirmation and fine-mapping of a major QTL for resistance to infectious pancreatic necrosis in Atlantic salmon (Salmo salar): population-level associations between markers and trait

Abstract

Background: Infectious pancreatic necrosis (IPN) is one of the most prevalent and economically devastating diseases in Atlantic salmon (Salmo salar) farming worldwide. The disease causes large mortalities at both the fry- and post-smolt stages. Family selection for increased IPN resistance is performed through the use of controlled challenge tests, where survival rates of sib-groups are recorded. However, since challenge-tested animals cannot be used as breeding candidates, within-family selection is not performed and only half of the genetic variation for IPN resistance is being exploited. DNA markers linked to quantitative trait loci (QTL) affecting IPN resistance would therefore be a powerful selection tool. The aim of this study was to identify and fine-map QTL for IPN-resistance in Atlantic salmon, for use in marker-assisted selection to increase the rate of genetic improvement for this trait.

Results: A genome scan was carried out using 10 large full-sib families of challenge-tested Atlantic salmon post-smolts and microsatellite markers distributed across the genome. One major QTL for IPN-resistance was detected, explaining 29% and 83% of the phenotypic and genetic variances, respectively. This QTL mapped to the same location as a QTL recently detected in a Scottish Atlantic salmon population. The QTL was found to be segregating in 10 out of 20 mapping parents, and subsequent fine-mapping with additional markers narrowed the QTL peak to a 4 cM region on linkage group 21. Challenge-tested fry were used to show that the QTL had the same effect on fry as on post-smolt, with the confidence interval for QTL position in fry overlapping the confidence interval found in post-smolts. A total of 178 parents were tested for segregation of the QTL, identifying 72 QTL-heterozygous parents. Genotypes at QTL-heterozygous parents were used to determine linkage phases between alleles at the underlying DNA polymorphism and alleles at single markers or multi-marker haplotypes. One four-marker haplotype was found to be the best predictor of QTL alleles, and was successfully used to deduce genotypes of the underlying polymorphism in 72% of the parents of the next generation within a breeding nucleus. A highly significant population-level correlation was found between deduced alleles at the underlying polymorphism and survival of offspring groups in the fry challenge test, parents with the three deduced genotypes (QQ, Qq, qq) having mean offspring mortality rates of 0.13, 0.32, and 0.49, respectively. The frequency of the high-resistance allele (Q) in the population was estimated to be 0.30. Apart from this major QTL, one other experiment-wise significant QTL for IPN-resistance was detected, located on linkage group 4.

Conclusion: The QTL confirmed in this study represents a case of a major gene explaining the bulk of genetic variation for a presumed complex trait. QTL genotypes were deduced within most parents of the 2005 generation of a major breeding company, providing a solid framework for linkage-based MAS within the whole population in subsequent generations. Since haplotype-trait associations valid at the population level were found, there is also a potential for MAS based on linkage disequilibrium (LD). However, in order to use MAS across many generations without reassessment of linkage phases between markers and the underlying polymorphism, the QTL needs to be positioned with even greater accuracy. This will require higher marker densities than are currently available.

Figure 1

Mortality curves from post-smolt challenge test.

Figure 2

Fine mapping of major IPN-resistance using post-smolt and fry material. The DNA markers used are indicated at the top of the graph. The 95% bootstrap confidence intervals for position is indicated by horizontal bars.

Figure 3

Levels of inter-marker LD on linkage group 21 plotted by genetic distance. The measure of LD is the square of the correlation coefficient, r². All possible paired combinations of the 21 genotyped microsatellites were plotted. A 6^th degree polynomial was fitted to the graph.

Figure 4

Animals used in the study. The animals referred to as parents in this study were parents of the 2005 generation of the breeding nucleus of an Atlantic salmon breeding company, and also parents of offspring group that were challenge-tested for IPN at the fry and post-smolt stages. These parents originated from four separate populations (or more specifically, year-classes of separate populations). Each population (i.e. year class) is denoted according to year of hatching. In the figure, the numbers refer to number of parents, and every box located within another box is a subset of the larger box. The numbers at the arrows indicate the progress of the study.

Figure 5

Distribution of challenge test mortality rates among offspring of QQ, Qq, and qq parents. Mortality rates of half-sib groups in the fry challenge test, classified according to the deduced genotype of the common parent at the underlying polymorphism. Each half-sib group contains two full-sib groups, and half-sib mortality rate is the average of the two full-sib mortality rates.