Mapping QTL for multiple traits using Bayesian statistics

文献类型： 外文期刊

第一作者： Xu, Shizhong

作者： Xu, Shizhong;Xu, Chenwu;Wang, Xuefeng;Li, Zhikang;Li, Zhikang

作者机构：

期刊名称：GENETICS RESEARCH （ 影响因子：1.588； 五年影响因子：1.263 ）

ISSN：

年卷期：

页码：

收录情况： SCI

摘要： The value of a new crop species is usually judged by the overall performance of multiple traits. Therefore, in most quantitative trait locus (QTL) mapping experiments, researchers tend to collect phenotypic records for multiple traits. Some traits may vary continuously and others may vary in a discrete fashion. Although mapping QTLs jointly for multiple traits is more efficient than mapping QTLs separately for individual traits, the latter is still commonly practised in QTL mapping. This is primarily due to the lack of efficient statistical methods and computer software packages to implement the methods. Mapping multiple QTLs simultaneously in a single multivariate model has not been available, especially when categorical traits are involved. In the present study, we developed a Bayesian method to map QTLs of the entire genome for multiple traits with continuous, discrete or both types of phenotypic distribution. Instead of using the reversible jump Markov chain Monte Carlo (MCMC) for model selection, we adopt a parameter shrinkage approach to estimate the genetic effects of all marker intervals. We demonstrate the method by analysing a set of simulated data with both continuous and discrete traits. We also apply the method to mapping QTLs responsible for multiple disease resistances to the blast fungus of rice. A computer program written in SAS/IML that implements the method is freely available, on request, to academic researchers.

分类号： Q3`R394`Q3

  • 相关文献
作者其他论文 更多>>

微信小程序