Estimativas de parâmetros genéticos em pesos corporais de codornas de corte através de análises unicaracterísticas com inferência bayesiana.

Abstract

This study aimed to estimate the genetic parameters of body weight from 15 sucessive generations of meat-type quails, using the unicharacteristic model. A total of 8130 records obtained from the Federal University of Pelotas quail improment program were used. The body weight trait was evaluated at hatch, 7, 14, 21, 28, 35 and 42 days of age. The MTGSAM computer program, which allows Bayesian interference, using Gibbs sampling, applied to an animal model to estimate componentes of covariance and heritabilities was used. By implementing the Gibbs sampling, 50000 interactions were used, due to the previous knowledge of estimates, obtained by the REML methodology, by firstly eliminating 20000 of the period of burn in of Gibbs chain and the interval of removing 100 interactions, generating a total of 300 samples of components of variance. It was observed that as the quails aged, all variances have shown a tendency to grow.. The estimated additive genetic variance incrased markedly from birth to 42 days of age, reaching 398.837 (BW42). The phenotypic and environrnental variances have shown similar behaviour, with the behaviour growing along the growing curve, obtaining in BW42 1250.59 and 851.75, respectively. The heritabilities tended to decrease from hatch to 42 days of age, varying from 0.49 (BW1) to 0.31 (BW14), slightly increasing in the BW21, being mantained at BW28, increasing again at BW35 and decreasing once again at 42 days (0.32). The highest estimate was found at hatch (0.49), indicating the the selection would have been successful if performed at 7 or 35 days of age. The presence of genetic variability indicates the the body weight trait may be used as selection criteria. These results suggest that the Bayesian inference suits the objective of obtaining estimates of meat-type quail body weight.