A Computational Capability for Fast and Reliable Characterization of Protein Complexes

蛋白质-蛋白质相互作用是生物活动的核心。它们构成了许多生物过程的基本组成部分，如信号转导、细胞周期控制、代谢和一般细胞机器。在后基因组测序时代，系统地表征细胞中蛋白质复合物中的蛋白质-蛋白质相互作用是功能基因组学和蛋白质组学的一个极具挑战性和重要的问题。通过（a）分析质谱数据和化学交联信息以及（B）在源自这些实验数据的几何约束下的蛋白质对接预测，将开发用于表征细胞中蛋白质复合物的综合计算能力。最初，这种能力将在一组选定的酵母蛋白质复合物上进行测试和验证，作为原理的证明。当完全开发后，这种能力将用于酵母（和其他基因组）中蛋白质复合物的基因组规模编目。这一拟议项目的具体目标是：（一）开发改进的计算方法，用于根据质谱数据定位和识别交联，特别是分子间交联;（二）开发新的计算方法，通过分析交联和其他实验数据，识别蛋白质复合物及其组成蛋白质;（iii）发展两种蛋白质的数据约束对接的新计算方法，这将在预测准确性和应用通用性方面显著改善现有对接方法;（iv）发展数据约束多方蛋白质对接的新计算方法;以及（v）将所开发的方法应用于选定的一组蛋白质复合物，以表征其复合物结构。 所提出的计算能力将显著提高以一种以前不可能的方式询问蛋白质-蛋白质相互作用的能力，为功能基因组学和蛋白质组学的新兴领域打开了新的大门。研究这些用于表征蛋白质复合物的计算能力将为有兴趣进入新兴的计算基因组学领域的研究生和博士后学员提供许多机会。生物学/生物信息学从其他学科。将聘请一些博士后学员来实施本研究的一些关键组成部分，并获得博士学位学生将被招募到这个项目的工作作为他们的论文项目的一部分。此外，在2002年秋季学期，将为ORNL/UTK联合基因组科学和技术研究生方案的研究生提供计算生物学课程。本专题中遇到的一些定义明确但具有挑战性的问题将作为本课程的学生学期专题。本项目所要解决的具有挑战性的科学问题将使研究生和博士后学员接触到生物信息学的核心问题，即，智能和有意义的解释大量的生物数据和生物结构/过程的计算建模，提供了一个很好的机会动手培训。通过实施该项目获得的新的科学发现将通过期刊出版物和会议介绍向公众开放。该项目开发的计算机软件将通过互联网免费提供给学术和政府机构。该项目产生的所有生物学数据都将作为数据库整理，并通过互联网免费公开提供。

Protein-protein interactions are at the heart of biological activities. They constitute the basic components of many biological processes such as signal transduction, cell-cycle control, metabolism, and general cellular machines. Characterization of protein-protein interactions in protein complexes in a cell, in a systematic manner, represents a highly challenging and important problem to functional genomics and proteomics in the post-genome sequencing era. An integrated computational capability for characterization of protein complexes in a cell will be developed, through (a) analyzing mass spectrometry data and chemical cross-linking information and (b) protein docking prediction under geometric constraints derived from these experimental data. Initially this capability will be tested and validated on a selected set of protein complexes from yeast, as a proof of principle. When fully developed, this capability will be used for genome-scale cataloging of protein complexes in yeast (and other genomes in general). The specific aims of this proposed project are: (i) development of improved computational methods for locating and identifying cross-links, particularly inter-molecular ones, from mass spectrometry data; (ii) development of new computational methods for identification of protein complexes and their component proteins, through analysis of cross-linking and other experimental data; (iii) development of new computational methods for data-constrained docking of two proteins that will significantly improve the existing docking methods in both prediction accuracy and application generality; (iv) development of new computational methods for data-constrained multi-party protein docking; and (v) applications of the developed methods to a selected set of protein complexes for their complex structure characterization. The proposed computational capability will significantly improve the ability to interrogate protein-protein interactions in a way not possible before, and open new doors for the emerging fields of functional genomics and proteomics.Investigation of these computational capabilities for characterization of protein complexes will provide a number of opportunities for graduate students and postdoctoral trainees who are interested in moving into the emerging field of computational biology/bioinformatics from other disciplines. A number of postdoctoral trainees will be hired to implement some of the key components of this research, and Ph.D. students will be recruited to work on this project as part of their thesis projects. In addition, a computational biology course will be offered to the graduate students of the joint ORNL/UTK Graduate Program in Genome Science and Technology, in the Fall semester of 2002. Some well-defined but challenging issues encountered in this project will be used as student term projects through this course. The challenging scientific problems to be solved in this project will expose the graduate students and postdoctoral trainees to the core issues of bioinformatics, i.e., intelligent and meaningful interpretation of massive amount of biological data and computational modeling of biological structures/processes, providing an excellent opportunity for hands-on training. New scientific findings through the implementation of this project will be made publicly accessible through journal publications and conference presentations. Computer software to be developed in this project will be made freely available to the academic and government organizations through the Internet. All the biological data generated in this project will be organized as databases, and also be made freely and publicly available through the Internet.