Modeling phenotypic transitions in gene expression state space

描述（由申请人提供）：现代生物医学研究中最大的挑战之一是理解基因型和表型之间的联系，更重要的是，创建模型，使我们能够预测生物系统对可能影响它们的各种扰动的反应。尽管基因组技术带来了数据的爆炸式增长，并且不断努力开发捕捉生物系统复杂性的模型，但我们尚未在实现这些目标方面取得重大进展。在这里，我们建议开发新的现象学方法，利用经验数据和来自其他来源的信息，如生物医学文献，来创建可以进一步测试和验证的预测模型。为了了解转录网络的全局特性并预测最有可能影响表型变化的元素，我们将我们的模型映射到热力学自旋系统，这使我们能够识别稳健的基因模块和相关的顺序参数。我们假设这些顺序参数以类似于水在固体、液体和气体之间依赖压力和温度的相变的方式驱动表型转变。为了开发和测试这种方法，我们将使用目前通过nhgri资助的基因组科学卓越中心（CEGS）项目生成的数据，在该项目中，我们正在研究致癌病毒衍生基因干扰细胞网络以驱动细胞转化的机制。我们确定的模块将用于对转化至关重要的关键参数进行具体预测，这些预测将通过使用RNAi干扰细胞网络进行测试，并通过基因表达分析和表型分析评估反应。与我的导师John Quackenbush博士以及共同导师Karl Munger博士和Giovanni Parmigiani博士合作，这个项目将使我能够制定一个研究计划，这将有助于我从物理学过渡到生物学，并帮助我获得一个独立的研究职位。为了进一步支持我的职业发展，我和我的导师制定了一个严格的培训计划，将为我提供统计学和计算生物学方面的正式培训，以及现代实验室分子生物学方面的实践培训。

DESCRIPTION (provided by applicant): One of the greatest challenges in modern biomedical research is to understand the link between genotype and phenotype and, more importantly, to create models that allow us to predict the response of biological systems to various perturbations that can affect them. Despite an explosion of data that has emerged from genomic technologies, and continued efforts to develop models that capture the complexity of biological systems, we have not yet made significant progress in achieving these goals. Here we propose to develop new, phenomenological approaches that leverage empirical data together with information from other sources such as the biomedical literature, to create predictive models that can be further tested and validated. To understand the global properties of transcriptional networks and to predict the elements with the greatest potential to effect phenotypic changes, we will map our models to thermodynamic spin systems which allow us to identify robust gene modules and associated order parameters. We hypothesize that these order parameters drive phenotypic transitions in a manner analogous to the pressure and temperature dependent phase transitions of water between solid, liquid and gas. To develop and test this approach, we will use data currently being generated through the NHGRI-funded Center for Excellence in Genome Science (CEGS) program in which we are investigating the mechanisms by which oncovirus-derived genes perturb cellular networks to drive cellular transformation. The modules we identify will be used to make concrete predictions about key parameters essential for transformation and these predictions will be tested by using RNAi to perturb the cellular networks and evaluating the response by gene expression analysis and phenotypic assays. Working with my mentor, Dr. John Quackenbush, and co-mentors, Dr. Karl Munger and Dr. Giovanni Parmigiani, this project will allow me to develop a research program that will facilitate my transition from physics to biology and help me to secure an independent research position. To further support my career development, my mentors and I have developed a rigorous training program that will provide me with formal training in statistics and computational biology, and hands-on training in modern laboratory molecular biology. PUBLIC HEALTH RELEVANCE: Knowing how genes interact in cellular networks to influence the way cells change from one state to another is essential for understanding a wide variety of processes ranging from development to human disease. Our project will use data on viral transformation of human cells, together with other publicly-available information, to develop computational models of gene-gene interactions and their role in this process; these models will then be mapped to equivalent systems borrowed from theoretical physics to identify crucial parameters governing phenotypic transitions. This general framework will provide insight into why particular cell states are robust and the best ways to perturb them to alter their phenotype.