Using single-step methodology for estimating genomic values of economic interest characteristics in Nellore cattle

Abstract

Reproductive, meat and carcass traits are crucial to the economic efficiency of beef cattle production system and should therefore be considered as objective selection in animal breeding programs . In traditional breeding , the breeding value estimation has been performed using only the phenotypic observations for each animal and pedigree, without any knowledge of the genetic architecture of traits . By identifying genes that act on these features and the selection of animals that carry desirable genes, a greater genetic gain can be achieved in a shorter time. As compared to traditional genomic selection based on BLUP ( Best Linear Unbiased Prediction), provides greater accuracy for the prediction of breeding values, mainly for the characteristics that are expressed in one sex and / or that have a low heritability, such as for reproductive traits. Currently, genomic reviews using complex methods being implemented by a procedure which involves several steps, as shown, for example, by VanRaden (2008) and Hayes et al. (2009). In dairy cattle, a typical evaluation is performed in three steps : a) traditional evaluation with an animal, b ) model estimation of genomic effects for a relatively small number of animals and genotyped, and c ) estimation of genomic values for a selection index . The use of this procedure in the genetic evaluation requires the use of several parameters and assumptions , resulting in a complex methodology with greater risks of bias. Misztal et al. (2009 ), to simplify the procedure for several steps, the method proposed for the single -step in which the genomic information is incorporated in step a) , wherein the relationship matrix based on pedigree (A) is combined with a matrix genomic relatedness (G ) based on information from molecular markers, then becoming a proceeding one step. Thus, the matrix A can be modified in a matrix H, which includes both kinship based on pedigree as differences based on genomic information ( AA ) : H = A + AA, where AA is a matrix that contains the variances due to information genomics. Aguilar et al. (2010 ) used the methodology of single -step in genetic evaluation and compared its performance with the multi-step procedures . These authors concluded that based on genomic information and pedigree reviews using the procedure of the single-step were as good as those using several stages in terms of accuracy and bias, and that the main advantages of the single -step are its simplicity and the automatic derivation weights for combining the multiple sources of information in the estimation of genomic breeding values. In Brazil, the availability of phenotypic data on reproductive traits is relatively large, since they are covered in some genetic evaluation programs. With the data being collected in the thematic project , it will be possible also analyze data from meat and carcass characteristics. The objectives of the present work are: to predict genomic values using single-step method and to compare the results with multi-step approaches; compare the models using cross-validation methods; to study biologically relevant pathways associated to economic traits. The results of this project will provide tools that will assist in improving these characteristics, which increase the efficiency of beef production and calf production systems . This should lead to greater economic return, with a decrease of the production cycle, benefiting the entire production chain.