Genomic Selection for Economic Traits in Inner Mongolia Cashmere Goats by Integrating GWAS Prior Information.

The accuracy of genomic selection has a significant impact on the selection of superior individuals in livestock. Studies have reported that integrating GWAS information can improve the accuracy of genomic prediction. In this study, phenotypic data, systematic environmental data, and genotypic data of important economic traits (cashmere yield, cashmere diameter, body weight, and cashmere length) of Inner Mongolia cashmere goats were utilized. Based on the results of a previous genome-wide association study that considered additive and dominance effects, the top 5%, top 10%, top 15%, and top 20% of loci were extracted as prior marker information. The genomic breeding values for each trait were estimated using the GBLUP-GA method based on GWAS prior information, and the accuracy of genomic prediction was further evaluated using a five-fold cross-validation method. The results showed that the contribution of significant loci to the genetic variance of each trait gradually increased with an increase of the number of integrated loci. The genetic variance contribution rates of significant loci to cashmere yield, cashmere diameter, body weight, and cashmere length were 64-71%, 47-57%, 76-82%, and 66-80%, respectively. The additive heritability estimates for cashmere yield, cashmere diameter, body weight, and cashmere length using GWAS prior information were 0.252-0.266, 0.297-0.580, 0.305-0.330, and 0.107-0.117, respectively. These values were higher than those obtained using the traditional G matrix constructed from all loci, with increases of 0.052-0.066, 0.007-0.29, 0.134-0.159, and 0.015-0.025, respectively. The results of genomic prediction accuracy showed that when 5% of the GWAS prior information was integrated, the highest genomic prediction accuracy was achieved for cashmere yield (0.8156), body weight (0.8361), and cashmere length (0.7571). When 20% of the GWAS prior information was integrated, the genomic prediction accuracy for cashmere diameter was 0.8074, which was significantly higher than that at other levels. Additionally, it was found that the dominance heritability for cashmere diameter, body weight, and cashmere length was very small and could be ignored when integrating GWAS prior information. Therefore, when integrating prior information for genomic selection of these traits, the influence of dominance effects can be disregarded.