Wolc Anna, Arango Jesus, Jankowski Tomasz, Settar Petek, Fulton Janet E, O'Sullivan Neil P, Fernando Rohan, Garrick Dorian J, Dekkers Jack C M
Department of Genetic and Animal Breeding, Poznan University of Life Sciences, Wolynska Street 33, 60-637 Poznan, Poland.
Avian Dis. 2013 Jun;57(2 Suppl):395-400. doi: 10.1637/10409-100312-Reg.1.
A genome-wide association study (GWAS) using Bayesian variable selection was performed to determine genomic regions associated with mortality due to Marek's disease virus (MDV) infection in layers. Mortality (%) under experimental disease challenge (500 plaque-forming units of a very virulent plus MDV strain) was recorded for progeny groups (average 15.5 birds; range 3 to 30) of 253 genotyped sires from four generations of a brown-egg layer line. An additional generation of 43 sires with progeny data was used to validate results. Sires were genotyped with a 42K Illumina single-nucleotide polymorphism (SNP) chip. Methods BayesB (pi = 0.995) and BayesCpi, with or without weighting residuals by the size of progeny groups were applied. The proportion of genetic variance contributed by SNPs within each 1-megabase (Mb) genomic region was quantified. Average mortality was 33% but differed significantly between generations. Genetic markers explained about 11% of phenotypic variation in mortality. Correlations between genomic estimated breeding values and percentage of progeny mortality for the validation generation (sons of individuals in training) were 0.12, 0.17, 0.02, and 0.16 for BayesB, weighted BayesB, BayesCpi, and weighted BayesCpi, respectively, when using the whole genome, and 0.03, 0.20, -0.06, and 0.14, when using only SNP from the 10, 1-Mb regions, explaining the largest proportion of genetic variance according to each method. Results suggest that regions on chromosomes 2, 3, 4, 9, 15, 18, and 21 are associated with Marek's disease resistance and can be used for selection and that accounting for the size of progeny groups has a large impact on correct localization of such genomic regions.
开展了一项使用贝叶斯变量选择的全基因组关联研究（GWAS），以确定与蛋鸡马立克氏病病毒（MDV）感染所致死亡率相关的基因组区域。记录了来自一个褐壳蛋鸡品系四代的253头基因分型种公鸡后代组（平均每组15.5只鸡；范围为3至30只）在实验性疾病挑战（500个蚀斑形成单位的超强毒MDV毒株）下的死亡率（%）。另外一代43头有后代数据的种公鸡用于验证结果。使用Illumina 42K单核苷酸多态性（SNP）芯片对种公鸡进行基因分型。方法应用了BayesB（π = 0.995）和BayesCπ，对后代组大小加权或未加权残差。对每个1兆碱基（Mb）基因组区域内SNP所贡献的遗传方差比例进行了量化。平均死亡率为33%，但各代之间存在显著差异。遗传标记解释了死亡率表型变异的约11%。当使用全基因组时，验证代（训练个体的儿子）的基因组估计育种值与后代死亡率百分比之间的相关性，对于BayesB、加权BayesB、BayesCπ和加权BayesCπ分别为0.12、0.17、0.02和0.16；当仅使用根据每种方法解释最大比例遗传方差的10个1-Mb区域中的SNP时，相关性分别为0.03、0.20、-0.06和0.14。结果表明，2号、3号、4号、9号、15号、18号和21号染色体上的区域与马立克氏病抗性相关，可用于选择，并且考虑后代组大小对这类基因组区域的正确定位有很大影响。
