USDA, ARS, US Meat Animal Research Center, PO Box 166, Clay Center, NE 68933, USA.
J Anim Sci. 2011 Jun;89(6):1731-41. doi: 10.2527/jas.2010-3526. Epub 2011 Feb 4.
The effects of individual SNP and the variation explained by sets of SNP associated with DMI, metabolic midtest BW, BW gain, and feed efficiency, expressed as phenotypic and genetic residual feed intake, were estimated from BW and the individual feed intake of 1,159 steers on dry lot offered a 3.0 Mcal/kg ration for at least 119 d before slaughter. Parents of these F(1) × F(1) (F(1)(2)) steers were AI-sired F(1) progeny of Angus, Charolais, Gelbvieh, Hereford, Limousin, Red Angus, and Simmental bulls mated to US Meat Animal Research Center Angus, Hereford, and MARC III composite females. Steers were genotyped with the BovineSNP50 BeadChip assay (Illumina Inc., San Diego, CA). Effects of 44,163 SNP having minor allele frequencies >0.05 in the F(1)(2) generation were estimated with a mixed model that included genotype, breed composition, heterosis, age of dam, and slaughter date contemporary groups as fixed effects, and a random additive genetic effect with recorded pedigree relationships among animals. Variance in this population attributable to sets of SNP was estimated with models that partitioned the additive genetic effect into a polygenic component attributable to pedigree relationships and a genotypic component attributable to genotypic relationships. The sets of SNP evaluated were the full set of 44,163 SNP and subsets containing 6 to 40,000 SNP selected according to association with phenotype. Ninety SNP were strongly associated (P < 0.0001) with at least 1 efficiency or component trait; these 90 accounted for 28 to 46% of the total additive genetic variance of each trait. Trait-specific sets containing 96 SNP having the strongest associations with each trait explained 50 to 87% of additive variance for that trait. Expected accuracy of steer breeding values predicted with pedigree and genotypic relationships exceeded the accuracy of their sires predicted without genotypic information, although gains in accuracy were not sufficient to encourage that performance testing be replaced by genotyping and genomic evaluations.
从 1159 头育肥牛的体重和个体采食量估计了与 DMI、代谢中期 BW、BW 增重和饲料效率相关的个体 SNP 以及 SNP 组的效应，这些效应以表型和遗传残差采食量表示。这些 F(1)×F(1)(F(1)(2)) 育肥牛的父母是 AI 配种的 F(1)后代，其亲本是 Angus、Charolais、Gelbvieh、Hereford、Limousin、Red Angus 和 Simmental 公牛，与美国肉类动物研究中心 Angus、Hereford 和 MARC III 综合母牛交配。这些育肥牛使用 BovineSNP50 BeadChip 分析（Illumina Inc.，圣地亚哥，加利福尼亚州）进行基因分型。在 F(1)(2)代中，频率>0.05 的 44163 个 SNP 的效应通过混合模型进行估计，该模型包括基因型、品种组成、杂种优势、母畜年龄和屠宰日期的当代群体作为固定效应，以及具有动物间记录谱系关系的随机加性遗传效应。通过将加性遗传效应分为与谱系关系有关的多基因成分和与基因型关系有关的基因型成分来估计该群体中归因于 SNP 组的方差。评估的 SNP 组包括包含 6 到 40000 个 SNP 的子集，这些 SNP 是根据与表型的关联选择的。90 个 SNP 与至少 1 个效率或组成性状强烈相关（P<0.0001）；这些 SNP 占每个性状总加性遗传方差的 28%到 46%。与每个性状具有最强关联的包含 96 个 SNP 的性状特异性集合解释了该性状加性方差的 50%到 87%。使用谱系和基因型关系预测育肥牛的育种值的预期准确性超过了没有基因型信息预测其父亲的准确性，尽管准确性的提高不足以鼓励用基因分型和基因组评估代替性能测试。
