Checks for the presence of heterogeneity of variances on milk production in first
lactation of females of Brown Swiss and, impact on genetic evaluation of breeding, using the
Bayesian inference using Gibbs sampling we used 2981 records on milk yield and age at
calving in first lactation of Holstein Brown Swiss, distributed in 62 herds. The records were
from the service of dairy control of the Brazilian Association of Breeders of Brown Swiss
cattle, with births occurring between the years 1980 to 2002. Established two classes of
phenotypic standard deviation for milk production. Subsequently, the data were analyzed
ignoring and considering the classes of standard deviation. The means and standard
deviations for milk production in the classes of high and low standard deviation and analysis
were generally equal to 5802.02 ± 1929.96, 4844.37 ± 1592.99, 5373.47 ± 1849.13
respectively. The averages for the later components of variance were higher in the high
standard deviation. The heritability obtained in the high standard deviation was close to the
value observed in the overall analysis and less than the value found in the low phenotypic
standard deviation. The genetic correlation for milk production between the classes of
standard deviation was equal to 0.48. Pearson and Spearman correlation coefficients and
coefficient between breeding values for milk production obtained in the overall analysis, with
the values obtained for each class of standard deviation were all higher than 0.80, when
considering all breeding. However, refining the sample of players shows that the correlations
decrease in magnitude. If there is a greater variability present in flocks in the high standard
deviation, and the impact of this heterogeneity of variance on genetic evaluation of breeding
is small because the main source of this heterogeneity is due to genetic factors confirming the
presence of heterogeneity of variances.
|
