Integration of polygenic and individual SNP effects in genome-wide association analyses

The lack of consideration of polygenic effects in genome-wide association studies (GWAS) may bias the results in complex traits controlled by multiple genes. The goal of this study is to develop a composite-GWAS model that identifies individual SNPs while adjusting for polygenic effects. The complex trait residual feed intake (RFI), an indicator of the feed efficiency based on maintenance and growth, was modeled. RFI and genotypic data (5,910 SNPs from chromosomes 3, 11 and 24) from 1,387 steers from different breeds and receiving different diets were analyzed, with and without the additive polygenic effect. The model included the fixed effects of days of feed, diet, breed and interaction between diet and breed, and the random effects of contemporary group and additive polygenic effect. A total of 69 and 141 SNPs were detected (P-value < 0.01) with the model including and excluding polygenic effects, respectively. The higher number of SNPs identified by the second model confirms that ignoring polygenic effects in GWAS of multi-gene traits can lead to false positives due to linkage disequilibrium. Seven SNPs (P-value < 0.001), four in chromosomes 3, two in chromosome 11 and one in chromosome 24, were detected using the polygenic model. Two SNPs, one from chromosome 3 and one from 11 are located within coding gene regions. Our results demonstrate the need to use composite-GWAS that include polygenic effects in complex multi-gene traits. These results indicated that the genetic improvement of feed efficiency in beef cattle may be accelerated by the incorporation of these markers in genomic selection strategies.