Semi-parametric statistical models for complex traits analysis using genomic data from Nelore cattle

Abstract

A large amount of genomic data is now available for identification and selection of genetically superior individuals, with potential to increase the accuracy of prediction of breeding values and, thus, the efficiency of animal breeding programs. Several statistical methods have been proposed for use in genomic data. Although several studies are being conducted, is still very limited number of studies determining the ability of prediction with statistical parametric and semi- parametric models with a set of real genomic data. Thus, in the proposed project, the prediction ability of genomic values with parametric models BLUP (Best Linear Unbiased Predictor), ssGBLUP (Single Step Genomic Best Linear Unbiased Predictor), BayesianLasso and the semi-parametric model RKHS (Reproducing Kernel Hilbert Spaces regression) will be studied using a set of simulated and real genomic data for growth characteristics, reproductive and carcass and meat in Nellore cattle. With the results, we hope to obtain important and relevant information in the genetic evaluation of commercial cattle populations. (AU)