Identifying gene regulatory modules of heat shock response in yeast

Wei-Sheng Wu,Wen-Hsiung Li

doi:10.1186/1471-2164-9-439

Abstract

BackgroundA gene regulatory module (GRM) is a set of genes that is regulated by the same set of transcription factors (TFs). By organizing the genome into GRMs, a living cell can coordinate the activities of many genes in response to various internal and external stimuli. Therefore, identifying GRMs is helpful for understanding gene regulation.ResultsIntegrating transcription factor binding site (TFBS), mutant, ChIP-chip, and heat shock time series gene expression data, we develop a method, called Heat-Inducible Module Identification Algorithm (HIMIA), for reconstructing GRMs of yeast heat shock response. Unlike previous module inference tools which are static statistics-based methods, HIMIA is a dynamic system model-based method that utilizes the dynamic nature of time series gene expression data. HIMIA identifies 29 GRMs, which in total contain 182 heat-inducible genes regulated by 12 heat-responsive TFs. Using various types of published data, we validate the biological relevance of the identified GRMs. Our analysis suggests that different combinations of a fairly small number of heat-responsive TFs regulate a large number of genes involved in heat shock response and that there may exist crosstalk between heat shock response and other cellular processes. Using HIMIA, we identify 68 uncharacterized genes that may be involved in heat shock response and we also identify their plausible heat-responsive regulators. Furthermore, HIMIA is capable of assigning the regulatory roles of the TFs that regulate GRMs and Cst6, Hsf1, Msn2, Msn4, and Yap1 are found to be activators of several GRMs. In addition, HIMIA refines two clusters of genes involved in heat shock response and provides a better understanding of how the complex expression program of heat shock response is regulated. Finally, we show that HIMIA outperforms four current module inference tools (GRAM, MOFA, ReMoDisvovery, and SAMBA), and we conduct two randomization tests to show that the output of HIMIA is statistically meaningful.ConclusionHIMIA is effective for reconstructing GRMs of yeast heat shock response. Indeed, many of the reconstructed GRMs are in agreement with previous studies. Further, HIMIA predicts several interesting new modules and novel TF combinations. Our study shows that integrating multiple types of data is a powerful approach to studying complex biological systems.

Highlights

A gene regulatory module (GRM) is a set of genes that is regulated by the same set of transcription factors (TFs)
By integrating transcription factor binding site (TFBS), mutant, ChIP-chip, and gene expression data, Heat-Inducible Module Identification Algorithm (HIMIA) identified 29 GRMs, which in total contain 182 heat-inducible genes regulated by 12 heatresponsive TFs
A TF name is colored blue if it is known to be involved in heat shock response but black otherwise

Summary

Introduction

A gene regulatory module (GRM) is a set of genes that is regulated by the same set of transcription factors (TFs). Lemmens et al [12] developed ReMoDiscovery to identify GRMs of the yeast cell cycle and yeast stress response by combining motif information, ChIP-chip and gene expression data. Tanay et al [13] applied a graph theoretic approach and developed SAMBA to reveal the modular organization of the yeast regulatory system by combining protein interactions, growth phenotype data, ChIP-chip, and gene expression data. All these module inference algorithms are statistics-based methods, treating dynamic time series gene expression data the same way as static steady state gene expression data. The aim of this study is to develop a module inference method that suits this need

Objectives

Methods

Results

Discussion

Conclusion