Abstract
With the advent of beef cattle breeding programs, the search for alternative traits to be evaluated in order to increase the productive efficiency of the system, as well as new methodologies that are increasingly efficient in predicting the genetic merit of animals are of great importance. In this regard, the objective of this project is to test different approaches for genomic relationship matrix (G) used in ssGBLUP and two marker densities, in order to verify the accuracy of genomic predictions, estimates of genetic parameters and classification of animals in different methodologies. Two databases will be used, one simulated and one real, consisting of pure and synthetic animals (Montana), and three traits. These were the probability of pregnancy at 14 months (PP14), Age at First Calving (AFC), and weight at 550 days (P550). The matrix G will be approached in three different ways, being: 1 - the traditional approach, program default ssGBLUP (G1); 2 - the G matrix will be centered on the specific allelic frequencies of each population (G2); 3 - will be centered and scaled to the specific allelic frequencies of each population (G3). In addition, the densities of 30 and 777K markers will be tested to ascertain the need for imputation or genotyping of the animals in high-density panels. The genetic parameters, Estimated Breeding Values (EBV), the accuracy of the models and three different validation methods will be estimated to test the predictive ability of the models, for each population, trait, and model studied. The result of this study should provide subsidies for the use of the ssGBLUP methodology adequately in multibreed populations in order to increase the accuracy of genetic evaluations and genomic selection. (AU)
|