Utility of indirect phenotype measurement system on genomic prediction in Japanese Black cattle: a simulation study using public data

Abstract The demand for genetic improvement of novel traits such as meat quality, including fatty acid composition, in Japanese Black cattle has been growing. The simple indirect measurement system is an alternative to traits for which direct measurement of multi-sample data is costly. However, the heritability estimated using the data obtained by the simple indirect measurement system is often lower than that by the direct measurement system. Here, the performance of genomic prediction using the simple indirect measurement system was assessed by simulating phenotypic data with the publicly available information on the genomic relationship matrix for 9,850 Japanese Black fattened animals. First, a theoretical method was developed to simulate genomic breeding value from the genomic relationship matrix based on eigenvalue decomposition. Next, genomic breeding value prediction was performed by simulating phenotypic measurement sets with heritabilities of 0.7 and 0.4 for the direct and indirect measurement systems, and genetic and phenotypic correlations of 0.95 and 0.8 between the two measurement systems, respectively. Variance components estimated via two-trait model analysis using records of all 9,850 individuals were in good agreement with their respective setting values, indicating that the proposed simulation method works well. The efficiency of genetic selection based on the results of genomic prediction using the indirect measurement system seems to depend on the genetic correlation between direct and indirect measurement systems, especially when the size of the training population increases. The results provide fundamental information for the efficient, cost-effective genetic improvement of novel traits using the simple indirect measurement system.