家畜全基因组关联分析中的贝叶斯方法

发布时间：2018-02-10 17:38

本文关键词： 贝叶斯多元回归全基因组关联分析多重比较动物育种计算生物学　出处：《黑龙江畜牧兽医》2016年21期 　论文类型：期刊论文

【摘要】：为了研究贝叶斯统计在家畜全基因组关联分析(GWAS)中的应用,试验采用了贝叶斯多元回归(BMR)方法,主要讨论其理论基础、处理SNPs间局部高相关性和后验Ⅰ型错误率(PER)在控制GWAS中假阳性(FP)问题中的应用,并基于模拟数据来比较该方法的性质。结果表明:1)即使预测变量间存在由连锁不平衡(LD)导致的潜在相关性,BMR也可以成功地应用在全基因组预测问题中;2)在真阳性检测和低假阳性方面,本方法仍然具有较为优良的性质。说明贝叶斯多元回归方法可以得到优良的推断结果,但是QTL加性效应估计值可能出现与模拟数据"真"值正负号相反的情况。
[Abstract]:In order to study the application of Bayesian statistics in the whole genome association analysis of livestock, the Bayesian multivariate regression (BMR) method was used in the experiment, and its theoretical basis was discussed. The application of dealing with local high correlation between SNPs and type I error rate (per) in the control of false positive GWAS. The properties of this method are compared based on simulated data. The results show that even if there is a potential correlation between predictive variables caused by linkage disequilibrium (LDD), BMR can be successfully used in the whole genome prediction problem. Testing and low false positives, This method still has good properties. It shows that Bayesian multivariate regression method can obtain good inferences, but the QTL additive effect estimation may be opposite to the "true" value of simulated data.
【作者单位】：河套学院农学系;爱荷华州立大学动物科学系;河套学院土木工程系;
【基金】：国家自然科学基金项目(31460594) 国家留学基金委项目(201308155140) 河套学院教学研究项目(HTXYJZ14005)
【分类号】：S813.1

【相似文献】