Estimation of Genetic Parameters by Restricted Maximum Likelihood under Multi-trait Animal Models in Selected Layer Lines.

Abstract

Genetic parameters for body weight at 256 days (BW), age at first egg (SM), egg weight at 252 days (EW) and egg production efficiencies between 169-280 days (EPA) and between first egg and 280 days (EPB) were estimated in five layer chicken lines selected at the National Livestock Improvement Center, Okazaki Station, between 1986 and 1994. The lines were three White Leghorn lines (E1, E4 and S46), one recessive White Rock line (L17) and one Rhode Island Red line (Y8). Parameters were estimated by restricted maximum likelihood procedure under the expectation maximization algorithm using multiple trait animal models. The model included year and house as fixed effects, inbreeding as a covariate and random additive genetic effects.Estimates of heritability differed slightly between lines and, for BW, SM, EW, EPA and EPB were between 0.51-0.65, 0.35-0.47, 0.51-0.60, 0.13-0.19 and 0.20-0.27 respectively with weighted average heritabilities of 0.58, 0.44, 0.56, 0.17 and 0.23 respectively. Genetic correlations between EPA and EPB were expectedly high at between 0.80 and 0.93 while the genetic correlations between these two traits and the other traits were low and negative except with BW where the correlations were low but positive. Estimated heritabilities and, especially, the correlations were different among lines and from those assumed at the station for making the desired gain indices which indicates the importance of estimating variance and covariance components in each line for use in making selection decisions.