Avaliação de persistência na lactação de vacas Guzerá utilizando modelos de regressão aleatória

Abstract

A total of 8,276 records from milk productions in test-day, between 6th and 305th days of lactation, had been considered of 1,021 first lactations Guzerát cows in five different herd. The aim of this work was to define the genetic parameters for persistency in the lactation and milk production in the test-day. The (co)variance components of the random regression coefficients had been esteem by the maximum restricted probability method in the software REMLF90, using two different functions (Ali and Schaeffer and Legendre polynomials). The models had included the contemporary group effect, cow class ages at calving, direct additive genetic, permanent environment and residual effect (that it was considered constant). The comparison to verify the better model adjustment was carried through the likelihood function (- 2ln L) and the Akaike Information Criterion (AIC). The function of Ali and Schaeffer were the one that better adjusted, but it showed convergence problems. The herdability for test day between 6th to 305th days varied from 0.19 to 0.40 for the function of Ali and Schaeffer and from 0.19 to 0.46 for the polynomial of Legendre. The correlations between milk production records tended to decrease with the increase of the days between each test-day, and reached high values between successive controls. The herdability estimate values for six persistence in lactation measures had varied from 0.20 to 0.42 for the function of Ali and Schaeffer and from 0.20 to 0.39 for the polynomial of Legendre. The correlations between the measures of persistence and milk production, in 305 days, had been low for the two models, denoting the possibility of modification in the shape of lactation curve without modifying the total milk production.