Computational Methods Towards the Spatio-Temporal Organization of the Proteome

描述(由申请人提供)：蛋白质组学和基因组学项目已经产生了细胞中蛋白质的详细清单。然而，人们对这些蛋白质在细胞内的空间和时间排列方式知之甚少。这些知识对于理解蛋白质如何对活细胞的结构和功能做出贡献是必不可少的。就像单词必须被组合成句子、段落和章节才有意义一样，重要的细胞功能是由蛋白质的结构化组装(或复合体)执行的，而不是单个分子。通常，这些复合体由数十或数百种蛋白质组成。这项提议描述了一套计算方法，将利用和整合几种新出现的实验数据来揭示蛋白质复合体的结构和动力学，最终目标是实现对细胞的机械理解。特别是，我们将通过以下项目在三个水平上表征蛋白质组组织。(1)我们将开发一种有效的计算方法来确定单个蛋白质复合体的结构，方法是将多个组分同时拟合到冷冻电子显微镜图上。(2)我们将开发先进的模式挖掘方法来发现和定位全细胞冷冻电子断层图像中的未知蛋白质复合体-这是全面视觉蛋白质组学的前提。(3)我们将开发模拟方法，在生物相关的时间尺度和现实的细胞环境中研究蛋白质相互作用网络的系统动力学。有了这些工具，我们准备好对蛋白质组的空间和时间组织进行建模。我们将免费向科学界提供所有方法的软件包和源代码。

DESCRIPTION (provided by applicant): Proteomics and genomics projects have yielded a detailed inventory of proteins in a cell. However, little is known about how these proteins are spatially and temporally arranged within the cell. Such knowledge is essential to understanding how proteins contribute to the structure and function of a living cell. Just as words must be assembled into sentences, paragraphs, and chapters to make sense, vital cellular functions are performed by structured assemblies (or complexes) of proteins rather than individual molecules. Often, these complexes comprise tens or hundreds of proteins. This proposal describes a set of computational methods that will exploit and integrate several kinds of emerging experimental data to uncover the structure and dynamics of protein complexes, toward the ultimate goal of achieving a mechanistic understanding of the cell. In particular, we will characterize proteome organization on three levels through the following projects. (1) We will develop an efficient computational method to determine the structure of individual protein complexes by simultaneously fitting multiple components to cryo-electron microscopy maps. (2) We will develop advanced pattern mining methods to discover and localize unknown protein complexes in whole-cell cryoelectron tomograms - a prerequisite towards comprehensive visual proteomics. (3) We will develop simulation methods to study the systems dynamics of protein interaction networks on biologically relevant time scales and within realistic cellular environments. With these tools, we are poised to model the spatial and temporal organizations of proteome. We will freely provide software packages and source code for all methods to the scientific community.