Estimation of variance components for some production traits of
Iranian Holstein dairy cattle using
Bayesian and AI-REML methods
Sadegh Alijani*, Mehdi Jasouri, Nasrolah Pirany and Hossein Daghigh
Kia
Department of Animal Sciences, Faculty of Agriculture,
University of Tabriz, Tabriz,
5166614766, Iran
*Corresponding author: sad-ali@tabrizu.ac.ir
Abstract
The used data set included the records of 131990 Iranian Holstein dairy cattle
for first three lactations that were collected from 1981 to 2008 time period by
Animal Breeding Center, Iran.
The traits which were considered for 305 days of lactation included milk, fat
and protein yield and percentages of milk fat and protein. Variance components
were estimated using average information restricted maximum likelihood (AI-REML)
algorithm using AIREMLF90 software under single trait and repeatability models
and Bayesian method by using a Gibbs sampling technique (BAGS) and by MTGSAM and
GIBBS3F90 software by same models. The linear statistical models of the analyses
included herd-year-season and lactations as fixed effects, age at calving as
covariate and animal and permanent environment as random effects. The ranges
of heritabilities estimates for lactations 1 to 3 by animal single and
repeatability models using AI-REML and BAGS methods were
0.19 to 0.29, 0.17 to 0.26, 0.20 to 0.25, 0.21 to 0.25 and 0.19 to 0.35 for
milk, fat and protein yield and percentage of milk fat and protein respectively.
Repeatability
estimates by using
BAGS method were 0.44, 0.35, 0.43 and by AI-REML method the values
were 0.43, 0.34, 0.39 for milk, fat and protein yield, respectively. The results
showed that estimated genetic parameter values by AI-REML analyses for all
traits and lactations in both models were smaller than BAGS method. In addition,
estimated heritability values for later lactations were lower in comparison with
the first lactation.