Supplementary data to publication “An approximate Bayesian significance test for genomic evaluations” (Biom J)

A simulation study has been conducted to analyse the association between genetic and phenotypic variation in livestock. Following the density and distribution of single nucleotide polymorphisms (SNPs) on the Illumina BovineSNP50 chip, 52,773 SNPs were simulated on the cattle genome of 30 Morgan length. Several generations of random mating were executed in which random recombination events according to the genetic distance between SNPs and random mutation of SNP alleles were considered. In the most recent generations, 50 sires were mated to 20 dams in order to generate multiple half-sib families. The data were split into training (n=2,000) and validation/testing set (n=2,000). Twenty-three SNPs were randomly preselected to be the causative variants, and additive, dominance and epistatic effects were simulated. Two different traits were achieved by adding different residual error terms to the sum of genetic effects, such that the total genetic variation contributed either 30% or 50% to the phenotypic variation. Then, 5,227 SNPs (every 10-th SNP including the causative variants) were selected. The simulation was repeated 100 times. More details can be found in Wittenburg et al. (2011) Including non-additive genetic effects in Bayesian methods for the prediction of genetic values based on genome-wide markers. BMC Genetics 12:74, https://doi.org/10.1186/1471-2156-12-74

本研究开展了一项模拟实验，以分析畜禽遗传变异与表型变异之间的关联。基于Illumina BovineSNP50芯片的单核苷酸多态性（single nucleotide polymorphisms, SNPs）密度与分布特征，在长度为30摩根（Morgan）的牛基因组上模拟得到52773个SNPs。随后进行多代随机交配，模拟过程中纳入了基于SNPs间遗传距离的随机重组事件，以及SNP等位基因的随机突变。在最后几代实验中，选取50头公畜与20头母畜交配，以构建多个半同胞家系。将数据集划分为训练集（n=2000）与验证/测试集（n=2000）。随机预先选取23个SNPs作为致病变异，并模拟加性、显性及上位性遗传效应。通过向遗传效应总和添加不同的残差误差项，构建两种不同的表型性状：其中总遗传变异对表型变异的贡献率分别为30%与50%。随后筛选出5227个SNPs（每10个SNPs选取1个，包含致病变异）。该模拟实验重复执行100次。更多细节可参见Wittenburg等人（2011）发表于《BMC Genetics》的研究："Including non-additive genetic effects in Bayesian methods for the prediction of genetic values based on genome-wide markers"，DOI: 10.1186/1471-2156-12-74