Production environment is determinant in the genetic relationship between growth traits of Nile tilapia through a reaction norm model

Genotype by environment interaction (G x E) has been a matter of interest for breeders and fish producers around the world, mainly because the selection performed in one environment may not have the desired response in other environments. Most studies conducted on G x E in aquaculture species have used a single trait approach; however, breeding programs have selected for multiple traits simultaneously. Thus, it was of interest in this study to use a multiple trait reaction norm model to analyze the genetic relationship between growth traits of Nile tilapia in the presence of G x E in a (sub)tropical environment. A multiple trait reaction norm model was applied to data from 12,192 animals evaluated for body weight (BW), trunk length (TL) and head percentage (HP). The adopted environmental descriptor was the solution of contemporary groups for BW. The ratio between slope and additive genetic intercept of the reaction norms ranged from 0.59 to 0.70, which demonstrated the presence of G x E for BW, TL, and HP. The mean of the heritability estimates showed variation of up to 300% along the environmental descriptor. Thus, the selection responses for BW, TL, and HP should be different depending on the production environment. The genetic correlations within each trait across different levels of the environmental descriptor were high and positive between closer environments and reached values close to or below zero be-tween distant environments (-0.34,-0.31, and 0.07 for BW, TL, and HP, respectively). Therefore, the best individuals in one environment will not necessarily be the same in a different environment. Given genetic correlation estimates greater than 0.85 between BW and TL across the environmental descriptor, selection for increased BW or TL will lead to a favorable reciprocal correlated response regardless of the environment. The genetic correlations between BW and HP and between TL and HP across environmental descriptor reached negative values of-0.63 and -0.77 in better environments, respectively. Thus, selection to increase BW or TL contributes to a more expressive reduction in HP in favorable production environments compared to more restrictive ones.