Expression-based monitoring of transcription factor activity: the TELiS database
Expression-based monitoring of transcription factor activity: the TELiS database
复制标题
DOI：
10.1093/bioinformatics/bti038
复制
发表时间：
2005-03-15
期刊：
影响因子：
5.8
通讯作者：
Zack, JA
中科院分区：
文献类型：
--
作者：
Cole, SW;Yan, W;Zack, JA
动机：在微阵列研究中，识别介导所观察到的基因表达变化的上游转录控制通路通常是令人感兴趣的。转录元件监听系统（TELiS）将基于序列的基因调控区域分析与统计流行度分析相结合，以识别在上调或下调基因的启动子中过度表达的转录因子结合基序（TFBMs）。通过将问题分解为两个步骤来实现效率最大化：（1）预先编制流行度矩阵，该矩阵指定给定微阵列所检测的所有基因的启动子中各种转录因子的假定结合位点数量；（2）对预先编制的流行度矩阵进行实时统计分析，以识别在差异表达基因的启动子中过度表达或表达不足的TFBMs。相互关联的JAVA应用程序，即启动子扫描（PromoterScan）和启动子统计（PromoterStats）执行这些任务，并共同构成用于转录因子活性反向推断的TELiS数据库。
结果：在两项验证研究中，TELiS准确地检测到HIV - 1感染体内激活的核因子 - κB（NF - κB）和I型干扰素系统，以及外周血单核细胞中糖皮质激素受体的药理激活。TELiS所依据的基于群体的统计推断在分析灵敏度上优于传统的统计检验，参数研究表明从少至20个差异表达基因就能准确识别转录因子活性。因此，TELiS为识别介导所观察到的基因表达动态的转录控制通路提供了一种简单、快速且灵敏的工具。
Motivation: In microarray studies it is often of interest to identify upstream transcription control pathways mediating observed changes in gene expression. The Transcription Element Listening System (TELiS) combines sequence-based analysis of gene regulatory regions with statistical prevalence analyses to identify transcription-factor binding motifs (TFBMs) that are over-represented among the promoters of up- or down-regulated genes. Efficiency is maximized by decomposing the problem into two steps: (1) a priori compilation of prevalence matrices specifying the number of putative binding sites for a variety of transcription factors in promoters from all genes assayed by a given microarray, and (2) real-time statistical analysis of pre-compiled prevalence matrices to identify TFBMs that are over- or under-represented in promoters of differentially expressed genes. The interlocking JAVA applications namely, PromoterScan and PromoterStats carry out these tasks, and together constitute the TELiS database for reverse inference of transcription factor activity.Results: In two validation studies, TELiS accurately detected in vivo activation of NF-kappa B and the Type I interferon system by HIV-1 infection and pharmacologic activation of the glucocorticoid receptor in peripheral blood mononuclear cells. The population-based statistical inference underlying TELiS out-performed conventional statistical tests in analytic sensitivity, with parametric studies demonstrating accurate identification of transcription factor activity from as few as 20 differentially expressed genes. TELiS thus provides a simple, rapid and sensitive tool for identifying transcription control pathways mediating observed gene expression dynamics.