Detecting differential expressions in GeneChip microarray studies: A quantile approach
Detecting differential expressions in GeneChip microarray studies: A quantile approach
复制标题
DOI：
10.1198/016214506000001220
复制
发表时间：
2007-03-01
影响因子：
3.7
通讯作者：
He, Xuming
中科院分区：
文献类型：
--
作者：
Wang, Huixia;He, Xuming
在本文中，我们考虑通过对基因芯片研究中基于探针水平测量的基因表达分位数进行建模和分析，来检验差异表达基因。通过针对具有随机效应的线性分位数模型开发一种稳健的秩得分检验，我们提出了一种可靠的检验方法，用于检测强度分布的某些分位数的差异。通过对检验统计量进行全基因组调整以考虑阵列内相关性，我们证明了即使阵列数量较少，所提出的秩得分检验也是非常有效的。我们对实际实验数据的实证研究表明，检测探针水平数据的四分位数差异是对基于高斯似然的常规混合模型分析的一种有价值的补充。本文提出的方法是开发混合模型中分位数回归推断工具的首次尝试。
In this article we consider testing for differentially expressed genes in GeneChip studies by modeling and analyzing the quantiles of gene expression through probe level measurements. By developing a robust rank score test for linear quantile models with a random effect, we propose a reliable test for detecting differences in certain quantiles of the intensity distributions. By using a genomewide adjustment to the test statistic to account for within-array correlation, we demonstrate that the proposed rank score test is highly effective even when the number of arrays is small. Our empirical studies with real experimental data show that detecting differences in the quartiles for the probe level data is a valuable complement to the usual mixed model analysis based on Gaussian likelihood. The methodology proposed in this article is a first attempt to develop inferential tools for quantile regression in mixed models.