Genetic Analysis of Milk Yield in Holstein Friesian Cattle in Ethiopia Using Test-Day Models
No Thumbnail Available
Date
2015-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
A clear definition of breeding goal and accurate selection of breeding animals depends on well-organized performance recording and animal evaluation system. As Ethiopia is currently embarking on a national herd performance recording and animal evaluation system, genetic analysis for the estimation of genetic parameters for milk yield and identification of the modern and accurate animal evaluation models is essential. The objectives of this study were to select best test-day model(s) (TDM), estimate genetic and phenotypic parameters using the selected models, predict the genetic merit of animals and assess the relative performance of lactation average model (LAM) versus random regression test-day model (RRM) for milk yield in Holstein Friesian herds in Ethiopia. For this study, data from the first three lactations of Holstein Friesian cows that calved between 1997 and 2013 was used. The data comprised of 13,421 test-day milk yield records from 800 cows from two large dairy herds of Holstein Friesian. Variance components were estimated using the average information restricted maximum likelihood method fitting single trait first and multi-lactation animal models. The genetic merit of animals was evaluated using the best linear unbiased prediction method. In the first model comparison, TDM with intercept, linear, quadratic (RRM2) and cubic (RRM3) orders of Legendre polynomials fitted for additive genetic and permanent environmental effects were compared. The RRM3 model was favored by all model comparison criteria. However, the analysis of eigenvalues of additive genetic covariance matrix indicated that the addition of a third order Legendre polynomials explained only about 0.07 percent of the additive genetic variation. Moreover, the estimates of genetic parameters from RRM3 model were not logical and implausible biologically. On the other hand, the competing RRM2 model was less parameterized and gave biologically reasonable parameter estimates across lactation and was best fit to the variance structures of the data set. The next model selection was carried out for modeling the permanent environmental effects using RRMs of varying either intercept, first, second or third order of Legendre
xvii
polynomials whilst fitting with the same second order for additive genetic effects. The result suggested that at least a second or higher order Legendre polynomials fits are needed to model properly the permanent environmental variance structure for the accurate estimation of genetic parameters for test-day milk yield. The estimates of genetic parameters from the best selected RRM2 model for first lactation data showed that heritability estimates for test-day milk yield ranged from 0.17 to 0.29. It was lowest at the beginning and highest at and around 230 days in milk. The genetic and phenotypic correlations between milk yield at different test-days ranged from 0.37 to 0.99 and 0.29 to 0.71, respectively. Genetic correlations for adjacent days in milk were close unity and decreased as the distance between test-days increased. In multi-lactation RRM2 model analysis, the heritability estimates from first, second and third lactations ranged from 0.20 to 0.26, 0.15 to 0.27 and 0.17 to 0.28, respectively. Across lactation genetic correlations between first and second, second and third and first and third lactations on 305-d basis were 0.88, 0.83 and 0.70, respectively. These correlations were much lower than one. This implies that different lactations should be treated as different but correlated traits in a multi-lactation analysis. The estimated breeding values for milk yield from RRM2 model showed a higher standard deviation compared to LAM indicating that TDM makes efficient use of TD information. Correlations of breeding values between RRM2 and LAM models ranged from 0.90 to 0.96 for different groups of sires and cows and marked re-rankings were observed in top sires and cows in moving from the traditional LAM to RRM evaluations. In general, this study found that the less parameterized RRM2 model fits best the variance structure in the data. A multi-lactation analysis of test-day milk yield is the best strategy for the accurate genetic evaluation in Holstein Friesian herds for milk yield. The use of TDM than LAM will lead to more efficient utilization of available information resulting in more accurate estimates of breeding values. In view of the trait considered in this study, other important traits should be included in the herd performance recording practices and future estimates of genetic parameters should include several breeding goal traits.
Description
PhD Dissertation
Keywords
Breeding Value, Dairy Cattle, Genetic Parameters, Milk Yield