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-kappaB 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.