Abstract: 【Background】The contribution of genetic breeding is highest in improving the efficiency of animal production. Through breeding, animal husbandry enterprises can improve production efficiency and obtain maximum economic benefits. Genome selection has been widely used in plant and animal breeding. Genomic selection can estimate breeding values (EBV) by using high density markers covering the whole genome. Compared with pedigree information, the average relationship between individuals obtained by using these markers is more accurate, so that breeding values can be more accurately estimated and individuals can be selected. In practical breeding program, all individuals are not genotyped, especially for pigs, whose economic values are not high enough, hence the application of genomic selection is limited in pig breeding. The single-step genomic best linear unbiased prediction (ssGBLUP） can utilize both pedigree and genotypes information, allowing part of individuals are genotyped, thus greatly reducing genotyping costs while maintaining high prediction accuracy. At present, many studies have shown that the use of genomic selection in pig breeding can improve the accuracy of prediction, but in actual breeding, breeding cost is also an important issue in livestock enterprises to consider. Therefore, how to implement breeding program economically and effectively is of great research value.【Objective】The study on the effect of one-step genome selection on the population evaluation of Duroc provided the basis for genome selection breeding program.【Methods】In this study, three important economic traits of duroc pig born from 2009 to 2018 in a pig farm in Fujian province were studied. We compared the accuracy of BLUP, GBLUP, and ssGBLUP in calculating the estimated breeding value on reproductive and growth traits of Duroc pigs. We explored the impacts of genotyped individuals with different proportions in the reference population on the ssGBLUP prediction abilities. We also studied the influence of different chip density on GBLUP prediction abilities. 【Result】The results showed as follows: 1) The heritability of the age at 100kg, backfat and eye muscle area was 0.257, 0.250 and 0.399, respectively; 2) Compared with BLUP, the accuracy of ssGBLUP was improved by 14.7%~51.1%; Compared with GBLUP, the accuracy increased by 13.4%~45.7%; 3) When 10%~30% of individuals was genotyped, the prediction accuracy of ssGBLUP can exceed that of BLUP; The prediction accuracy reaches a plateau when 40-60% of individuals genotyped. 【Conclusion】Based on the above results, it was concluded that: 1) Compared with BLUP, ssGBLUP can improve the accuracy and reliability of EBV for each trait; Compared with the GBLUP, ssGBLUP was slightly lower than the GBLUP with only pedigree information of those ungenotyped, but ssGBLUP performed better than the GBLUP method after the addition of phenotypes of the ungenotyped individuals. 2) As the proportion of genotyped individuals in the reference population increases, no matter which selection method is used to determine genotyped individuals (random selection and selection the key individuals), the prediction ability of ssGBLUP is gradually improved. The results showed that the ssGBLUP can improve the prediction ability of individual breeding value even if only a proportion of individuals were genotyped when the breeding budget was limited.

Key words: duroc, genomic selection, ssGBLUP, the proportion of individuals genotyped