Collaborative Research and RUI: Stochastic Dynamic Network Models of Gene Regulation under Environmental Stress

本项目由一位生物学家和一位数学家及其本科生合作完成，研究环境压力下基因调控网络的随机动态建模。所有生物都必须对环境的变化和压力作出反应才能生存和繁殖。这些环境压力包括养分或氧气可用性的变化，渗透压或pH值的变化，活性氧或其他破坏性物质的存在，以及温度的突然或巨大变化。生物体通过基因表达的特征性程序对环境胁迫作出反应。在理解这种环境应激反应中，最有趣和最具挑战性的问题是细胞内基因调控网络的动态行为。仔细调节这些网络是生物体的一项基本活动。该项目的目标是：(1)通过对公开数据的挖掘、酵母缺失菌株的遗传筛选和内部微阵列数据的分析，以及基于我们模型的贝叶斯网络重建方法，确定出芽酵母（Saccharomyces cerevisiae）对冷休克反应的转录因子网络；(2)对该模型进行分析，将其与科学界正在开发和使用的确定性化学动力学和动态贝叶斯网络模型进行比较；(3)建立多种温度变化的额外外源扰动及其对生长速率的影响模型，并将其纳入随机动态网络模型；(4)利用qRT-PCR和DNA微阵列对野生型和基因缺失菌株的总RNA和转录活性mRNA进行实验验证模型预测，并通过连续的模拟和实验对模型进行改进；(5)开发和分析适用于研究各种基因调控网络的通用数学建模框架。高通量基因组学方法的发明改变了21世纪的生物学。一次一个基因？整个系统的分析方法。贝克?酵母（Saccharomyces cerevisiae）是一种理想的研究模式生物，因为它生长迅速，而且在一次实验中可以测量所有6000个基因的表达。虽然这种一次测量所有基因表达的能力是理解基本细胞过程和导致疾病的缺陷的重要的第一步，但它仍然只是一种方法。零件清单吗?。正如列出建造房子所需的木板、钉子、砖块和灰泥的数量和种类并不能解释房子是如何组装在一起的一样，简单地测量所有的基因并不能解释细胞是如何运作和对环境压力作出反应的。相反，为了了解细胞功能，我们需要了解哪些基因在什么情况下表达以及这些基因如何相互作用的规则。简而言之，我们需要了解这个复杂的基因调控网络是如何随时间变化的。在这个项目中，我们将建立一个数学模型，可以用来对细胞功能进行可测试的预测。我们的方法提供了合成目前正在使用的许多看似不同的基因调控分析技术的潜力。该研究项目还将对酵母对冷休克反应的整体调节机制产生生物学上的见解，这一机制目前尚不清楚。我们的工作将确定所涉及的特定调节因素，环境应激反应途径重叠的程度，并将提供单个因素的直接和间接影响的测量。我们开发的数学技术应该为基因表达数据的动态建模和统计分析的集成提供一个富有成效的框架，并且应该广泛适用于复杂生物体的生物学。特别是，一个明确处理基因在调控网络中的间接影响的模型应该提供对复杂疾病（如癌症）的原因的洞察，其中涉及多个基因和环境影响。

This project is a collaboration between a biologist and a mathematician and their undergraduate students in stochastic dynamic modeling of gene regulatory networks under environmental stress. All organisms must respond to changes and stresses in their environment to survive and reproduce. Such environmental stresses include changes in nutrient or oxygen availability, changes in osmolarity or pH, the presence of reactive oxygen species or other damaging agents, and sudden or large changes in temperature. Organisms respond to environmental stresses through characteristic programs of gene expression. Among the most interesting and challenging problems in understanding this environmental stress response is the dynamic behavior of gene regulatory networks within the cell. The careful regulation of these networks is a fundamental activity of the organism. The objectives of this project are (1) to identify the network of transcription factors that regulate the response to cold shock in budding yeast, Saccharomyces cerevisiae, through a combination of mining of publicly available data, the genetic screening of systematic yeast deletion strains and the analysis of in-house microarray data, and a Bayesian approach to network reconstruction based on our model; (2) to analyze the model, comparing it to deterministic chemical kinetic and dynamic Bayesian network models in development and use in the research community; (3) to develop models of the additional exogenous perturbations of multiple temperature shifts and the resultant affect on growth rate for integration into the stochastic dynamic network model; (4) to test the model predictions experimentally using qRT-PCR and DNA microarrays on both total RNA and transcriptionally active mRNA, in both wild type and gene deletion strains, improving the model through successive rounds of simulation and experiment; and (5) to develop and analyze a general mathematical modeling framework suitable for studying a wide variety of gene regulatory networks. The invention of high-throughput genomics methods has transformed 21st century biology from a ?one gene at a time? approach to the analysis of whole systems. Baker?s yeast, Saccharomyces cerevisiae, is an ideal model organism to study because it grows quickly and the expression of all 6000 genes can be measured in a single experiment. While this ability to measure the expression of all the genes at once is a significant first step towards understanding fundamental cellular processes and the defects that lead to disease, it is still only a ?parts list?. Just as listing the numbers and kinds of boards, nails, bricks, and mortar it takes to build a house does not explain how the house is put together, simply measuring all of the genes does not explain how cells function and respond to environmental stresses. Instead, to understand cell function, we need to understand the rules that govern which genes are expressed under what circumstances and how the genes interact with each other. In short, we need to understand how this complex gene regulatory network changes over time. In this project, we will build a mathematical model that can be used to make testable predictions about cell function. Our approach offers the potential to synthesize a number of seemingly disparate techniques of gene regulatory analysis currently being used. The research program will also yield biological insight into the overall regulatory mechanism of the response to cold shock in yeast, which is poorly understood. Our work will determine the particular regulatory factors involved, the extent of environmental stress response pathway overlap, and will provide a measure of the direct and indirect effects of individual factors. The mathematical techniques we develop should provide a fruitful framework for the integration of dynamic modeling and statistical analysis of gene expression data and should be broadly applicable to the biology of complex organisms. In particular, a model that explicitly deals with the indirect effects of genes in a regulatory network should provide insight into the causes of complex diseases such as cancer where multiple genes and environmental effects are involved.