Possibilities to improve the genetic evaluation of a rare breed using limitedgenomic information and multivariate BLUP

摘要

The use of molecular genetic information in the evaluation of livestock has become more common. This study looks at the efficacy of using such information to improve the genetic evaluation of a rare breed of dual-purpose cattle. Data were available in the form of pedigree information on the Gloucester cattle breed in the United Kingdom and recorded milk and beef performance on a small number of animals. In addition, molecular genetic information in the form of multi-marker, multiple regression results converted to a 1 to 10 score (Igenity scores) and 123 single nucleotide polymorphism (SNP) genotypes for 199 non-recorded animals were available. Appropriate mixed-animal models were explored for the recorded traits and these were used to calculate estimated breeding values (EBV), and their accuracies, for 6527 animals in the breed’s pedigree file. Various ways to improve the accuracy of these EBV were explored. This involved using multivariate BLUP analyses, genomic estimated breeding values (GEBV) and combining Igenity scores with recorded traits in a series of bivariate genetic analyses. Using the milk recording traits as an example, the accuracy of a number of traits could be improved using multivariate analyses by up to 14%, depending on the combination of traits used. The level of increase in accuracy largely corresponded to the absolute difference between the genetic and residual correlations between two traits, but this was not always symmetrical. The use of GEBV didnot increase the accuracy of milk trait EBV owing to the low proportion of varianceexplained by the 101 SNPs used. Using Igenity scores in bivariate analyses with therecorded data was more successful in increasing EBV accuracy. The largest increases werefound in genotyped animals with no recorded performance (e.g. a 58% increase in fat weightin milk); however, the size of the increase depended on the level of the geneticcorrelation between the recorded trait and the Igenity score for that trait. Lower levelsof improvements in accuracy were seen in animals that were recoded but not genotyped, andancestors which were neither genotyped nor recorded. This study demonstrated that it waspossible to improve the accuracy of EBV estimation by including Igenity score informationin genetic analyses but it also concluded that increasing the level of performancerecording in the breed would be beneficial.