The sensitivity of genetic values for body weight of meat type quails predicted at 21 and 35 days of age under diets with different crude protein levels was evaluated. Data from subjects belonging to two strains (EV1 and EV2) were used to fit a random regression model under heterogeneity of residual variance assumption. The random regression coefficients for intercept (bo) and slope (b1) were positively correlated in all analyses results, but the correlation was higher in the EV2 data analyses for both ages. Results indicated that additive genetic variance and heritability change as a function of the environment gradient for both genetic strains and ages. The reaction norms for EV1 strain suggest there is genotype by environment interaction (G x E) for both ages as there were remarkable changes in the ranking of body weight breeding values for different crude protein levels. Furthermore, changes in the magnitude of the genetic effects dispersion as a function of protein level of diet indicates there is G x E in EV1 and EV2 strains. Therefore, the prediction of breeding values for body weight of quails under a specific level of crude protein in the diet does not hold for different environments regarding the level of this nutrient.
genotype by environment interaction; random regression; genetic evaluation; meat type quails
