Analysis of control types in transcriptional regulation

基因表达的转录调控对包括生长、分化和发育在内的中枢生物过程是必不可少的。目前对复杂系统的高通量技术和生物信息学方法的可用性提供了一个前所未有的机会，可以就转录控制的组织逻辑提出一般性问题。在过去的几年中，独立的研究方向集中在分析(1)DNA拓扑对全球基因表达的影响和(2)转录调控网络在基因表达谱中的印记。从我们的角度来看，该系统可以在不同的实验条件下以不同的比例使用这两种类型的控制。此外，表达谱可以被认为是与新陈代谢需求相匹配的进化形成的模式。在这个项目中，我们试图从所有这些限制中提取一个一致的转录调控的统一控制理论。使用经典模式生物大肠杆菌，并在我们之前工作的基础上，我们将通过实验研究两种控制类型下的基因表达水平，并开发数学工具来定量评估每种控制类型在给定表达模式中的重要性。我们将通过实验研究结构蛋白突变的影响、西格玛因子的作用以及RNA聚合酶组成的变化。在理论方面，我们将使用有效网络的概念和适当的零模型来分析数据。通过将结合位点信息转化为局部DNA拓扑态的概率分布，我们的目的是预测在系统或环境确定的变化下基因表达的变化模式。

Transcriptional control of gene expression is essential for central biological processes including growth, differentiation and development. The current availability of high-throughput technology and bioinformatics approaches to complex systems provides an unprecedented opportunity to ask general questions on the organizational logic of transcriptional control. Independent lines of research over the last few years have focused on analyzing (1) the effect of DNA topology on global gene expression and (2) the imprint of the transcriptional regulatory network in gene expression profiles. From our perspective the system can employ these two types of control in different proportions under different experimental conditions. In addition, expression profiles can be thought of as evolutionarily shaped patterns matching the needs of metabolism. From all these constraints we attempt in this project to extract a consistent unifying control theory of transcriptional regulation. Using the classical model organism Escherichia coli and building up on our previous work we will experimentally study gene expression levels under variation of both types of control and develop mathematical tools for quantitatively assessing the importance of each control type in a given expression pattern. We will experimentally investigate the impact of mutations of structural proteins, the role of sigma factors and changes of RNA polymerase composition. On the theoretical side we will analyze the data using the concept of effective networks and appropriate null models. By translating binding site information into probability distributions of local DNA topological states we aim at predicting the patterns of changes of gene expression under defined changes of the system or the environment.