Seleção de codornas de corte para produção de ovos utilizando modelos de regressão aleatória

The objective of this study to compare random regression models with different orders in Legendre polynomials, to evaluate the best fit to promote the production of eggs and thus obtain the heritability, variances and correlations. Data were obtained from a genetic group of quails, comprising 644 females UFV1. The characteristic studied was the weekly production of eggs, individually evaluated by the number of eggs obtained from 6th to 57th week of life of quail. The genetic analysis of egg production the animal modelused in random regression using the software Wombat (Meyer, 2007). To model the fixed path and the random trajectories were used Legendre polynomials. To test the difference between the models studied comparisons were made between models using some criteria, but the model with four K = 3 for the fixed effects, K = 4 for additive direct genetic effects and Kc = 4 for the permanent environmental effect animal, provided better fit and the heritabilities found were low. The model with a Legendre polynomial of order three for the fixed effects and order four for the random effects, can be used to describe the trajectory of egg production and the highest values of heritability are found near the peak of production.