253 full-sib families from 33 males and 23 females of European seabass were produced in a partly factorial mating design. All fish were reared in the same tank during 14 months, then 7000 of them were dispatched in four farms to different locations (France, Israel, Italy, Portugal) representing a wide variety of environmental conditions. Around 400 g mean weight, 1177 to 1667 fish in each site were weighed and length was measured. Condition factor (K) was calculated. Pedigrees were redrawn a posteriori using microsatellites markers: parental origin could be retraced for 99.2% of fish. Due to a high incidence of deformities, the useful sample size was reduced to 491-670 fish per site.Maternal effects were small. Using a simple animal model, heritability of weight ranged from 0.38 ± 0.14 to 0.44 ± 0.14 in the different sites. Length was highly correlated to weight, with similar heritabilities. GxE interaction, estimated through genetic correlations of weight across the different environments ranged from 0.70 ± 0.10 to 0.99 ± 0.05. Genetic correlations between weight or length and K were not similar in the different sites.

Malformations in farmed fish are a prime research topic as it has a major impact on fish welfare and on the economical value of the product. In this study, the genetic component of spine deformities was estimated in European sea bass, a major species of Mediterranean aquaculture, with the use of a posteriori reconstruction of pedigree with microsatellites. Our population exhibited a very high malformation rate (81%), probably due to intense swimming of fish during pre-growing. The heritability of global spine deformities was evaluated to be 0.21 ± 0.04 on the underlying scale (0.33 ± 0.06 for lordosis, 0.13 ± 0.04 for scoliosis). Deformities scored by an internal or external examination turned out to be genetically the same trait (genetic correlations > 0.9), although phenotypic correlations were moderate to high (0.45 ± 0.04 to 0.92 ± 0.01) due to difficulties in scoring the less severe deformities. Along time (from 100 to 800 g mean weight), lordosis had a constant heritability, with very high genetic correlations between stages (> 0.9). Its incidence increased from 39 to 53% indicating that the malformation was probably present from the beginning of the rearing period but that its expression increased over time. A positive linkage between spine deformities and growth rate was demonstrated both by the higher early weight of affected fish and by the positive genetic correlation between weight and malformation (from 0.21 ± 0.15 to 0.40 ± 0.14), although deformities have a negative phenotypic impact on growth. However, heritabilities of malformation traits calculated here are only estimates that may indirectly describe the sensitivity of fish to the particular pre-growing environment as a function of different body weight genotypes. A quantification of realised heritability would therefore be necessary before concluding on the existence of a genetic basis for spinal malformations such as lordosis and scoliosis. Nevertheless and by precaution, selection for fast growth rate needs careful monitoring of deformities, and selection against malformations may counteract impacts of selection for fast growth

Selective breeding of European sea bass (Dicentrarchus labrax) receives a growing interest, as the estimated heritability of growth is medium to high. In this study, we compared the offspring of four groups of sea bass sires, mated with the same wild dams: wild (W), first generation of domestication (D), first generation of mass selection for length (M), first generation of PROSPER-like selection for length (P). The comparison was done both in replicated tanks (separate rearing) and in mixed tanks (mixed rearing) where sire origins were recovered by genotyping of eight microsatellite markers. Weight, length and growth rate were measured from day 238 post-fertilization (69 g mean weight) to day 611 post-fertilization (390 g mean weight). Both in mixed and separate tanks, both selected groups (P, M) were larger than unselected groups (W, D). No difference was seen at any time between W and D, nor between M and P. The selection response estimate on weight was larger in mixed tanks when compared to separate tanks (+ 42% in mixed tanks, + 23% in separate tanks at day 611), yielding realized heritability estimates of 0.60 and 0.34, respectively, and confirming the excellent potential of the species for growth improvement through selective breeding. Both selection response and the amplification effect between mixed and separate tanks decreased as rearing density increased. Our hypothesis is that selection response is magnified by competition in mixed tanks, while sub-optimal rearing conditions lower the observed selection response, more in separate tanks (where selected thus larger fish are at a higher density than unselected ones) than in mixed tanks (where all fish experience the same density effects).

Viral Nervous Necrosis (VNN) is a major threat for the European sea bass (Dicentrarchus labrax) aquaculture industry. The improvement of disease resistance through selective breeding is a promising option to reduce outbreaks. With the development of high-throughput genotyping technologies, identification of genomic regions involved in the resistance could improve the efficiency of selective breeding. The aim of this study was to identify quantitative trait loci (QTL) involved in VNN resistance and to quantify their effect.Four experimental backcross families comprising 378, 454, 291 and 211 individuals and two commercial populations A and B comprising 1027 and 1042 individuals obtained from partial factorial crosses (59♂ x 20♀ for pop A; 39♂ x 14♀ for pop B) were submitted to a redspotted grouper nervous necrosis virus (RGNNV) challenge by bath. A high-density single nucleotide polymorphism (SNP) chip panel was designed to develop the ThermoFisher Axiom™ 57 k SNP DlabChip, which was used for genotyping all individuals and building a high quality linkage map. In the backcross families, composite interval mapping was performed on 30,917, 23,592, 30,656 and 31,490 markers, respectively. In the commercial populations, 40,263 markers in pop A and 41,166 markers in pop B were used to perform genome-wide association studies (GWAS) using a GBLUP and a BayesCπ approach.Please note that this is an author-produced PDF of an article accepted for publication following peer review. The definitive publisher-authenticated version is available on the publisher Web site.One QTL was identified on chromosome LG12 in three of the four experimental backcross families, and one additional QTL on LG8 was detected in only one family. In commercial populations, QTL mapping revealed a total of seven QTLs, among which the previously mentioned QTL on LG12 was detected in both. This QTL, which was mapped to an interval of 3.45 cM, explained 9.21% of the total genetic variance in pop A, while other identified QTLs individually explained less than 1% of the total genetic variance.The identification of QTL regions involved in VNN resistance in European sea bass, with one having a strong effect, should have a great impact on the aquaculture industry. Future work could focus on the fine mapping of the causal mutation present on LG12 using whole genome sequencing. Highlights► Viral Nervous Necrosis is a major disease for European sea bass. ► A novel SNP array for European sea bass was designed. ► A total of nine QTL were detected. ► One QTL, shared by five over six of the data sets and located on the LG12 explained 9.2% of the total genetic variance.