Agriculture and Food Systems, University of Melbourne, Parkville, VIC 3010, Australia.
Hum Mol Genet. 2012 Oct 15;21(R1):R45-51. doi: 10.1093/hmg/dds332. Epub 2012 Aug 16.
The genetic architecture of complex traits in cattle includes very large numbers of loci affecting any given trait. Most of these loci have small effects but occasionally there are loci with moderate-to-large effects segregating due to recent selection for the mutant allele. Genomic markers capture most but not all of the additive genetic variance for traits, probably because there are causal mutations with low allele frequency and therefore in incomplete linkage disequilibrium with the markers. The prediction of genetic value from genomic markers can achieve high accuracy by using statistical models that include all markers and assuming that marker effects are random variables drawn from a specified prior distribution. Recent effective population size is in the order of 100 within cattle breeds and ≈ 2500 animals with genotypes and phenotypes are sufficient to predict the genetic value of animals with an accuracy of 0.65. Recent effective population size for humans is much larger, in the order of 10,000-15,000, and more than 145,000 records would be required to reach a similar accuracy for people. However, our calculations assume that genomic markers capture all the genetic variance. This may be possible in the future as causal polymorphisms are genotyped using genome sequence data.
牛复杂性状的遗传结构包括影响任何给定性状的大量基因座。这些基因座中的大多数都有较小的影响，但偶尔也会有因对突变等位基因的选择而发生中等至较大效应分离的基因座。基因组标记捕获了大多数但不是全部性状的加性遗传变异，这可能是因为存在具有低等位基因频率的因果突变，因此与标记不完全连锁不平衡。通过使用包括所有标记并假设标记效应是从指定先验分布中抽取的随机变量的统计模型，可以从基因组标记中实现遗传值的高精度预测。牛群的近期有效种群大小约为 100，而具有基因型和表型的 ≈ 2500 只动物足以预测动物的遗传值，准确率为 0.65。人类的近期有效种群大小要大得多，约为 10000-15000，而要达到类似的人类准确性，则需要超过 145000 条记录。然而，我们的计算假设基因组标记捕获了所有的遗传变异。随着使用基因组序列数据对因果多态性进行基因分型，这在未来可能成为可能。
