Improving the resistance of Atlantic salmon to amoebic gill disease through quantitative genetics and genomics

Amoebic gill disease (AGD) is caused by the amoeba Paramoeba perurans, which colonises gill tissue of several fish species. In salmon farming, AGD has been a major problem in Tasmania for decades, and is currently an emerging issue in Northern Europe. The treatment consist of repeated baths of freshwater or H2O2, which is expensive, logistically challenging and invasive. In Tasmania, selective breeding for lower gill-score in filed test has successfully increased the interval between the treatments over generations. The aim of this study was to estimate genetic parameters for AGD resistance of Atlantic salmon (Salmo salar) under controlled and field conditions in Northern Europe. Smolt from 150 families from the Marine Harvest nucleus was infected with P. perurans in a controlled challenge test (n=20 fish per family) while another sample of fish from the same families (n=40 fish per family)were placed at a high-risk location in Ireland where AGD-infections occurred naturally. Gill scores were obtained after two subsequent infections both in the controlled test and in the field, and mortality was recorded after the second treatment in the field. Heritability of gill-score ranged from 0.10 to 0.20. Genetic correlation between first and second gill-score was close to zero and non-significant in the challenge test, and 0.70±0.10 in the field test. Genetic correlations between gill-scores from challenge test and field-test ranged from -0.07 and 0.38. The fish from the field test suffered >20 % mortality shortly after the second freshwater treatment which was associated to poor gill health. This mortality had heritability of 0.06±0.01, and was positively genetically correlated to first (but not second) gill score in the challenge test. Genetic correlations between mortality and gill score in field was close to zero and non-significant. The challenge test hence had low power to predict field gill score, but some power to predict field mortality. The first gill score in the field was a good predictor for second gill score in the field. A random sample of 1,141 fish (offspring of 64 sires and 119 females) from field test in Ireland were genotyped using an Affymetrix chip for Atlantic salmon with 57,184 SNPs. We searched over all these SNPs for SNPs with greater variation in DNA base para (A / T and C / G) in fish with high resistance (law gill score) than in fish with low resistance (high gill score) against AGD. We found three SNP variants in two chromosoms with significant effect on gill scores recorded during the first AGD score in the test (one with significance level P<8.74e-07, and two with significance level P<2.62e-05). The three SNP variants explain 33% of the total genetic variation in gill score. Three candidate gens were located within the 20 Kb of the detected SNP variant on one of the chromosome. One of the candidate genes is known to play a role in the innate immune system, and may have an effect on resistance to AGD in salmon. The accuracy of breeding values with genomic information was compared to what can be attained from pedigree information for amoebic gill disease (AGD) for 1,130 animals, the offspring of 74 sires and 117 dams. A data set of about 10 offspring per sire family (n=760) were used to estimate the SNP-effects, while the remaining animals (n=370) were used to validate these effects. Accuracy of breeding values increased by 9-15% when using genomic information as compared with pedigree information. Use of genomic breeding values makes it also possible to select within families among candidates with no gill-score. We therefore conclude that use of genomic information can increase genetic gain for resistance to AGD as compared with classical selection. We first compared the transcriptome profiles of naïve individuals with fish during the 1st and 2nd infections in a challenge test. Through comparative analyses, we identified different components of the immune system that are differentially regulated through subsequent infections. These findings suggest that the 1st infection should be treated as a separate phenotype compared to later infections and provide further support for the low genetic correlation previously reported from the challenge experiments. We next examined gene regulatory variations between individuals with higher scores of AGD compared to those with the lower scores. In particular, we found genes involved in glycolysis (breaking down the glucose and releasing energy) and osmoregulation to have modified their expressions. We are currently investigating the functional properties of candidate genes and genetic pathway, important for AGD infection identified through transcriptome analysis. We also aim to integrate these results with the findings from our earlier genome wide association analysis performed on a field data.

Amoebic gill disease (AGD) is caused by the marine amoeba Neoparamoeba perurans, and is a major challenge for salmon production in Australia. In northern Europe, the parasite has been occasionally observed, but has not been documented to cause increased m ortalities until a few years ago. Since 2011, AGD has however been the major cause of salmon mortality in Ireland and Scotland, and has been observed in Norway, probably because of increased water temperature. AGD is treated by freshwater or H2O2 baths, w hich adds production costs and stress to the fish. Selective breeding for AGD resistance is currently performed in Australia and is to date the only known method to reduce the AGD problem over time. To breed for AGD resistance in Norway, Scotland and Irel and it is necessary to obtain genetic parameters for AGD resistance in north east Atlantic salmon, which will be obtained in this project. Through testing the same families for AGD resistance both in a controlled challenge test and under field conditions in Ireland will show whether the challenge test gives a good prediction of field resistance. Breeding for AGD resistance in Norway is dependent on such a controlled test, as the natural outbreaks occur with irregular intervals. Further, 1000 fish from the se trials will be SNP genotyped for 2 purposes; estimation of genomic breeding values and QTL search. Including genomic information in the breeding value estimation, either by detecting major QTL or by genomic selection could increase the efficiency of th e breeding program through within family selection. To gain insight about the mechanisms involved in AGD resistance, 60 fish will be RNA sequenced and included in a gene expression study. The results from this project will give the involved partners knowl edge on how to efficiently test and implement selection for increased AGD resistance in a breeding program for Atlantic salmon.