An approach to gene regulatory network inference.

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Recent advances in genomics technologies have provided the means to efficiently generate several types of global data that are informative on gene regulatory systems in human and mouse. In this project we will develop the methods in experimental design and computational analysis that are necessary for the successful use of these technologies in the study of complex gene regulatory networks in these species. Aim 1: Computational inference of gene regulatory network. We will combine the power of targeted gene perturbations, global functional genomic measurements, and computational inference and modeling, to develop an integrated, predictive approach for the study of gene regulatory networks in mammals. Although the combined analysis of gene expression data and transcription factor binding location data have been useful for the inference of gene regulatory network in simple organisms, such analyses have so far met with limited success in mammals. We will identify the key factors impeding progress and will design ways to ameliorate them. We will develop a "structure oriented" approach for the inference of a gene regulatory network from diverse data types, including data on the responses to gene perturbations. Our goal is to design and demonstrate an effective approach to infer complex gene regulatory networks responsible for maintaining stable cellular states in mammalian cells. Aim 2: Implementation our method to study gene regulatory systems in ESC The approach developed in aim 1 will be applied to study the gene regulatory network underlying pluripotency and self-renewal in mouse embryonic stem cells (ESC). We have previously shown that in this important cell type a major portion of the variation in global gene expression (65%) can be predicted by the binding patterns of a moderate number (12) of transcription factors. We will generate global gene expression data at selected time points after targeted gene perturbations on these and other regulators. Based on these data sets, together with the large amount of regulator binding location data already available for unperturbed ESC, we will use the methods developed in aim 1 to infer the gene regulatory network responsible for the maintenance of the undifferentiated state.