Random regression models in the selection of meat‑type quails for egg production

Abstract

The objective of this work was to evaluate random regression models with different orders of Legendre polynomials, as to the best fit for egg production of meat‑type quails. UFV1 and UFV2 genetic groups of meat‑type quails, from different origins, of the genetic breeding program of the Universidade Federal de Viçosa, Brazil, were evaluated. Egg production was determined weekly for 1,294 quails, from their 6^th to their 57^th week of age, of which 644 were from the genetic group UFV1, and 650 from UFV2. The animal model was used in random regression by Wombat software, and, for modeling variable trajectories in time, the Legendre’s polynomial functions were applied. Comparisons were made by the Akaike’s information criterion, the Bayesian information criterion of Schwarz, the logarithm of the likelihood function, and the likelihood ratio test. The model with K = 3, for fixed effects, K_a = 4, for genetic effects, and K_c = 4, for the effect of environment provides the best fit, do not causes major changes in the variance components, and provides better estimates of heritability.