Genetic parameters and genotype by environment interaction for production traits and organ weights of gilthead seabream (Sparus aurata) reared in sea cages

Gilthead seabream is a key fish species for farming in the Mediterranean region and is farmed in a large geographical area under various production circumstances. However, more than 80% of the genetically improved fingerlings originate from a single country, Greece, which poses a potential risk for genotype by environment interaction (GxE). Therefore, the objective of this study was to quantify GxE for several traits of gilthead seabream in two distinct commercial production sites, one in south of Greece (Galaxidi Marine Farm) and another in southeast of Spain (Cudomar). For this GxE experiment, a population of juveniles was produced by mass spawning of 33 males and 20 females on a single day. These juveniles were stocked in sea cages in both locations when they reached stocking size (~3 g) and grown under commercial conditions. Management conditions during the grow-out period were kept the same between the production sites, while the fish were subject to naturally occurring differences such as water temperature, dissolved oxygen, and salinity. Phenotypes were recorded when the fish reached commercial harvest size (~400 g). Genetic parameters were estimated by using a genomic relationship matrix that was built by using ~30 k SNP. All traits studied had higher genetic variation and heritabilities in Cudomar. For instance, the heritability of harvest weight was 0.37 ± 0.05 in Galaxidi and 0.55 ± 0.05 in Cudomar. GxE was estimated as genetic correlations between the same trait measured on different fish in the two environments. Moderate GxE was found for harvest weight (0.45 ± 0.11), growth (0.43 ± 0.11), fillet weight (0.49 ± 0.12), liver weight (0.61 ± 0.11), and viscera weight (0.62 ± 0.10). Weak GxE was found for fillet fat (0.87 ± 0.06), heart weight (0.76 ± 0.11), cardiosomatic index (0.93 ± 0.14), viscerosomatic index (0.90 ± 0.05), and hepatosomatic index (0.79 ± 0.09). In conclusion, moderate GxE estimates for growth traits indicate that with a single breeding program, performance data from both environments should be included, or that two separate breeding programs may be needed for the two environments. The higher genetic variances observed in Cudomar suggest that this environment is a more suitable test environment for selective breeding.