Statistical Discovery of S. Cerevisiae Genetic Networks

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (Applicant's Abstract): We plan to elucidate mechanisms underlying S. cerevisiae transcriptional regulation through the analysis of high-density DNA array (microarray) data. We will create models of S. cerevisiae transcriptional regulation that can be statistically validated against microarray data in comparison with one another. In particular, we will focus on modeling the effects of gene specific factors (activators and repressors), general transcription factors, and a complex of proteins called the Srb/mediator complex. In investigating these mechanisms, we will study how combinatorial control is implemented, and we will work towards building a single, integrated model of interactions that will permit us to predict the consequences of environmental stimulation or genetic disruptions. We are specifically interested in how the relatively small number of regulatory genes in S. cerevisiae can combine to control the expression of 6000 genes. We hypothesize that a second layer of control is contained in S. cerevisiae beyond gene-specific regulators, and we seek to characterize how specific components of the general transcriptional apparatus can be used to control expression. To approach the problem soundly, we will accomplish our objective of elucidating transcriptional control by achieving three specific aims: 1) we will computationally elucidate the combinatorial control that is exercised by gene specific and general transcription factors in S. cerevisiae, 2) we will validate different models of transcriptional control using statistical metrics, and 3) we will implement automated methods for model elaboration, model discovery, and optimal suggestion of biological experiments. Because a substantial amount of machinery underlying transcriptional regulation appears to be conserved across eukaryotes, the ability to construct a predictive model of complex eukaryotic cell behavior has ultimate application in the development of therapeutics for improving human health. Moreover, our proposed methods are general and thus could also be used to analyze human microarray data directly.