Determining the Relationships Between Genomic and Phenotypic Breeding Values

Abstract

The main aim of the animal breeding is to increase the yields of economic traits in future generations and increase the obtained production per animal. Genomic selection studies have become applicable as a result of rapid increase of both genetic and computer sciences. In this study, Bayes A, Bayes B, Bayes C, Bayes Cpi which are used in genomic selection and BLUP (best linear unbiased prediction) methods which are used traditional selection will be compared. For this aim genomic breeding value and phenotypic breeding value has been estimated for lactation milk yield of Holstein cows breeding in a private company in USA. Holstein cows (400 individuals) were genotyped with 54k SNPs. The marker input file was coded as - 10, 0, and 10 for marker genotypes AA, AB, and BB, respectively. A total of 50,000 iterations were used, with the first 5000 excluded as the burn-in. Bayes A, Bayes B, Bayes C and Bayes Cpi were performed using the software GenSel 4.55 and phenotypic breeding values for lactation milk yield in Holstein cows were estimated by restricted error maximum likelihood (REML), BLUP procedures using an MTDFREML (multiple trait derivative free restricted maximum likelihood) program employing an animal model. Correlations were examined between phenotypic and genomic breeding values with Spearman correlation. As a result, for estimating breeding values, while the highest correlation was found between BLUP with Bayes Cpi, the lowest correlation were found between BLUP with Bayes A.