Abstract
Random regression models (RRM) differ in terms of the functions used to describe the shape of lactation curves. The aim was to compare random regression models under different functions to describe the lactation curves from Holstein cows in herds of the state of Minas Gerais. A database of 28,118 production records was analyzed using the test-day records of 4,230 first parity cows from five herds. The Wilmink, Ali & Schaeffer and Legendre polynomial (orders 4, 5 and 6) functions were adjusted in RRM to model the mean production trend (fixed) and genetic and permanent environmental (random) effects. The residual variances were assumed to be constant throughout lactation. Analyses were performed using the AIREMLF90 program. Except for the model with the polynomial function of order 5, all models converged. The Wilmink function showed lower values for criteria based on the -2log (L), AIC and BIC. The model with the Legendre polynomial of order 6 showed lower residual variance. Heritability estimates were similar between functions, ranging from 0.07 to 0.18 and were higher from 215 days of lactation. From 155 days of lactation, genetic and permanent environmental correlations between successive controls are of high magnitude. The Wilmink function is the most suitable for the study of milk yields from primiparous Holstein cows. The selection of animals is possible from 155 days of lactation on. Permanent environmental effects have greater influence on the milk production at the end of lactation of primiparous cows and should be considered since they are important and may be cumulative throughout lactation.
Highlights
The Holstein breed has the largest number of individuals in the world and has been, and still is, the target of an intensive genetic selection, which has turned it into an extremely productive animal.According to the Foreign Agricultural Service of the US Department of Agriculture (USDA) December 2012 report on milk yield, Brazil was ranked third in the world
Through Regression Models (RRM), the genetic merit of each individual can be obtained at any point of lactation
Data were provided by a company specialized in software for agricultural management, from Belo Horizonte, MG, Brazil
Summary
The Holstein breed has the largest number of individuals in the world and has been, and still is, the target of an intensive genetic selection, which has turned it into an extremely productive animal. Each country and breed association defines a method for evaluating phenotypic and genetic traits of economic interest that usually are selected simultaneously, but with different weights, within selection indices. Among the different milk yield analysis approaches throughout lactation, Random Regression Models (RRM) have numerous advantages in the genetic evaluation of dairy cattle. In RRM, by the use of functions that define the production on day “t” as a linear function of time or of a logarithmic or exponential function of time, an adjustment of one global mean curve for the population is sought.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call