Modelos estatísticos para análise de características longitudinais em bovinos de corte sob coleta seletiva de dados

Abstract

The aim of the current study was to verify which statistical model is the most suitable for analysis of longitudinal traits in herds that use selective data collection. Weights at 330, 385, 440, 495 and 550 days of age of Nellore males were used. The single-trait (STM), multi-trait (MTM) and random regression models with linear polynomial spline (RRM) models were compared. The selective collection was simulated through sequential samplings of the 85 and 70% heavier animals over time. The posteriori distributions of the genetic parameters were obtained by the Gibbs sampler, with the three models mentioned above. The variances for weights obtained for the selective collection file analysis were smaller than those obtained for the complete file when the STM model was used. On the other hand, there was no significative difference between variances, heritabilities and correlations for weight in all files evaluated, when the MTM and RRM models were used. Multi-trait and random regression models are suitable for the analysis of longitudinal traits in herds that perform selective data collection