Estimation of Variances and Covariances for Direct and Maternal Effects from Preweaning Growth Records in Japanese Black Cattle

As a preliminary study for estimating variances and covariances for direct and maternal effects from preweaning growth records by restricted maximum likelihood (REML), the effect of stages of the population, base generations and models on the estimates of variance and covariance were studied in Japanese Black Cattle. Records were obtained from the Tottori National Live-stock Breeding Station covering the period from 1960 to 1985. Traits analyzed were birth weight and DG from birth to 2-months old. Three stages of the population were considered. The first stage was from 1960 to 1967, the second stage (midterm) from 1968 to 1977 and the last stage from 1978 to 1985. The following three periods, from 1952 to 1959, 1960 to 1967 and 1968 to 1977, were assumed to be the base generations. Two different models were considered. Model 1 includes direct additive genetic effect and maternal additive genetic effect as random effects. While model 2 includes non-additive maternal effect as additional random effects. The computation was done using the Derivative-Free Restricted Maximum Likelihood program (DFREML) written by MEYER. No consistent change of the estimates was observed with the stages. The estimates obtained assuming the different base generations seemed to be the same, and maternal heritability was increased under the model that did not account for the non additive maternal effect. It is concluded from these results that the generation assumed to be the base have little effect on the estimates and the model accounting for non-additive maternal effect is preferable in explaining preweaning growth in Japanese Black Cattle.