Efficiency of genomic selection using Bayesian multi-marker models for traits selected to reflect a wide range of heritabilities and frequencies of detected quantitative traits loci in mice
Efficiency of genomic selection using Bayesian multi-marker models for traits selected to reflect a wide range of heritabilities and frequencies of detected quantitative traits loci in mice
复制标题
DOI：
10.1186/1471-2156-13-42
复制
发表时间：
2012-05-31
期刊：
影响因子：
2.9
通讯作者：
Roehe, Rainer
中科院分区：
文献类型：
--
作者：
Kapell, Dagmar N. R. G.;Sorensen, Daniel;Roehe, Rainer
背景：基因组选择使用密集单核苷酸多态(SNP)标记来预测育种值，而传统的评估是基于表型记录和系谱信息来估计多基因效应。这项研究的目的是比较多基因、基因组和多基因-基因组组合模式，包括混合模式(根据被认为有实质性影响的基因SNP标记的百分比进行标记，范围从2.5%到100%)。数据包括2,188只小鼠的表型和SNP基因类型(10,946个SNPs)。选择不同的生长、行为和生理性状进行分析，以反映广泛的遗传力(0.10到0.74)和检测到的影响这些性状的数量性状座位(QTL)数量(1到20个)。结果：与传统的多基因选择相比，基因组选择具有较高的预测能力，特别是对于遗传力中等的性状和家系间的交叉验证。尽管使用基因组模型的性状的基因组变异比例比使用多基因模型的小22%到33%，但这种情况还是发生了。使用2.5%的混合基因组模型，基因组变异的比例比多基因模型小79%。尽管当假定较少数量的SNPs对性状有显著影响时，标记所解释的方差比例进一步降低，但大多数性状的基因组选择的PA几乎没有受到影响。这些低混合百分比导致了对单一SNP效应的更好估计。针对QTL较少的性状建立的基因组模型比多基因模型的PA更低。结论：基因组选择总体上优于传统的多基因选择，尤其是在家系间交叉验证的情况下。减少被认为影响性状的标记的数量并没有显著改变大多数性状的PA，特别是在家庭内交叉验证的情况下，但增加了与QTL相关的标记的数量。影响性状的QTL数量对PA有影响，与多基因模型相比，使用基因组模型时QTL数量较少，导致PA较低。
Background: Genomic selection uses dense single nucleotide polymorphisms (SNP) markers to predict breeding values, as compared to conventional evaluations which estimate polygenic effects based on phenotypic records and pedigree information. The objective of this study was to compare polygenic, genomic and combined polygenic-genomic models, including mixture models (labelled according to the percentage of genotyped SNP markers considered to have a substantial effect, ranging from 2.5% to 100%). The data consisted of phenotypes and SNP genotypes (10,946 SNPs) of 2,188 mice. Various growth, behavioural and physiological traits were selected for the analysis to reflect a wide range of heritabilities (0.10 to 0.74) and numbers of detected quantitative traits loci (QTL) (1 to 20) affecting those traits. The analysis included estimation of variance components and cross-validation within and between families.Results: Genomic selection showed a high predictive ability (PA) in comparison to traditional polygenic selection, especially for traits of moderate heritability and when cross-validation was between families. This occurred although the proportion of genomic variance of traits using genomic models was 22 to 33% smaller than using polygenic models. Using a 2.5% mixture genomic model, the proportion of genomic variance was 79% smaller relative to the polygenic model. Although the proportion of variance explained by the markers was reduced further when a smaller number of SNPs was assumed to have a substantial effect on the trait, PA of genomic selection for most traits was little affected. These low mixture percentages resulted in improved estimates of single SNP effects. Genomic models implemented for traits with fewer QTLs showed even lower PA than the polygenic models.Conclusions: Genomic selection generally performed better than traditional polygenic selection, especially in the context of between family cross-validation. Reducing the number of markers considered to affect the trait did not significantly change PA for most traits, particularly in the case of within family cross-validation, but increased the number of markers found to be associated with QTLs. The underlying number of QTLs affecting the trait has an effect on PA, with a smaller number of QTLs resulting in lower PA using the genomic model compared to the polygenic model.