Modelos de regressão aleatória para teores de gordura do leite de vacas da raça holandesa

Abstract

In the current economic globalization, it becomes necessary to discuss national dairy farming and aspects that can provide its development, with a view to competitiveness between domestic and international markets. Thus, this study aimed to estimate the variance components and heritability for milk fat content from Holstein cows of second and third orders of lactation, to see at what stage of lactation there is a greater influence of genetic factors. This information can be used to select genetically superior individuals of interest to both producers and dairy industry. The percentages of fat were taken from a database refer to Holstein cows belonging to 33 herds. The consistency of the database was performed with the aid of SAS and Visual FoxPro software. Random regression models were employed, once the database has repeated measurements of animals within the same lactation. The age of the cow on the analysis of milk and days in milk (stage) were used as covariates. Components of genetic variance, permanent environmental and residual were estimated throughout lactation. Used to describe the trajectory of the lactation curve of the animals, the random regression models differ by the degree of the Legendre polynomial. Large variations in heritability estimates for most measures of fat content, using models involving adjustments of Legendre polynomials were observed. Heritability estimates ranged from 0.20 to 0.50 for fat content for 2nd lactation order, and 0.06 to 0.30 for the 3rd lactation order. The orthogonal polynomial of third degree was the one that best fit the data of 2nd lactation order and fourth-degree polynomial to 3rd lactation order, providing more accurate estimates of genetic parameters, indicating that selection must be done in the end of lactation for second order of lactation and between 215th to 248th to animal from third order.