A Importância da qualidade da informação na predição de valores genéticos para características de crescimento em bovinos da raça nelore

Abstract

The objective of this study was to evaluate the influence of information quality in the prediction of genetic values for the Nellore cattle\'s growth traits. The information regarding the Nellore cattle from four farms participating in a program of the Brazilian National Association of Breeders and Researchers (ANCP, acronym in portuguese), named Nellore Brasil, is corresponding to the years 2012 and 2013. The selected farms are certified by the ANCP for quality of zootechnical information, having received the association\'s seal of approval, Global G1. Field data such as batch or management group were considered an assessment criteria, and the growth traits related to the body weight at 120 (W120), 210 (W210), 365 (W365) and 450 (W450) days were evaluated under different scenarios of information quality: inclusion of all management lot information and random inclusion of management lot information in 90%, 70%, 50%, 30% and absence of information on management lots. During the scrutiny of different scenarios of management lot information for the genetic evaluation of growth traits, it was found that there were changes in the file structure for genetic evaluation, after the statistical treatment of the data. The contemporary group experienced changes in the number of animals due to the poor quality of information. Changes were observed in all the (co)variance components and genetic parameters compared to those obtained by the scenario used as a reference. Concerning growth traits, there was an increase in the estimates of heritability and the additive genetic variance in as far as the number of information about management lots decreased. Due to the change in contemporary groups in the absence of information on management lots, the results obtained in genetic analyzes\' were biased, reaching overestimated breeding values, which can be misleading as to the choice of animals. With respect to the average of predicted breeding values, it was found that for the maternal additive genetic effects of the body weight at 120 and 210 days of age, the change in the information quality interfered negatively, leading to reduction of average breeding values of the herds evaluated. When evaluating correlations or Spearman rank breeding values obtained from data with information quality (scenario), in three groups of accuracy of breeding values compared to those obtained in different scenarios of the inclusion of information, in different proportions, it was found that there was a change in the animal classification for growth traits.