Herdabilidade de características de crescimento em bovinos da raça Nelore utilizando métodos da Máxima Verossimilhança Restrita e Inferência Bayesiana.

Animal breeding is a predictive science which is carried out on the basis of the genetic variability in the herd and uses selection as an important tool, whose efficiency is assessed by the heritability of the traits of interest, which is a specific parameter of each herd. Due to its good ability to indicate the consequences of the genetic management in production systems, heritability has attracted attention of researchers and received contribution of the evolution of computational resources and statistical methods for the estimation of variance components. This has resulted in higher precision estimates, initially by the Least Squares Method, later by Maximum Likelihood and Bayesian statistics. The Brazilian Nellore herd has undergone these stages in its genetic improvement process, which has been guided by the selection of growth traits (weights at standard ages) as phenotypes of economic interest, which showed moderate to high heritability. Heritability has been estimated by different approaches, but the estimation of variance components by Restricted Maximum Likelihood (REML) has been the favorite. However, limitations such as computational demand and big amount of data have favored the use of Bayesian Inference. Based on these assumptions, the objective of this review is to discuss heritability of growth traits in Nellore cattle, using Maximum Likelihood and Bayesian Inference. [ABSTRACT FROM AUTHOR]