Accuracy of Genomic Breeding Values in Small Genotyped Populations-A Simulation Study

In the present study two genetically connected small and large populations were simulated and the effect of different sources of information from foreign populations on the accuracy of predicted genomic breeding values of young animals of the small population was investigated. A large population consist of 200000 animals over 15 generations and a small population consist of 5000 animals over 3 generations were generated with QMSim simulation software in a such way that the small population was connected to the large one. Three scenarios were defined for estimating the accuracy of genomic evaluations based on various sources of available information. In the first scenario, the accuracy of genomic breeding values was estimated based on the genomic breeding values of individuals in the small population. In the second scenario, the accuracy of genomic breeding values of animals was estimated using the information of individuals in the small population accompanied with the genomic breeding values of male individuals from the large population. In the third scenario, phenotypic, genotypic and pedigree data of individuals from the large population were integrated to estimate the genomic breeding values of individuals in the small population. The averages for accuracy of estimated genomic breeding values were 0.34, 0.40 and 0.50 under the first, second and third defined scenarios, respectively. Furthermore, the averages for regression coefficient of prediction for genomic breeding values were 0.73, 0.83 and 0.93 under the first, second and third defined scenarios, respectively. The obtained results revealed that the integration of phenotypic, genotypic and pedigree information of both large and small populations had the most advantage for estimating the genomic breeding values of selected candidates in the small populations.