CAREER: An Integrated Approach to Predictive Protein-Protein Docking

基因组和蛋白质组项目产生了大量的生物学数据：全序列基因组、基因和蛋白质序列、蛋白质结构、蛋白质-蛋白质相互作用和基因表达模式。生物信息学的下一个重大挑战是确定所有基因和蛋白质的生物功能和细胞角色，以及它们之间的相互作用，以形成一个功能细胞。这项研究的目的是通过计算预测蛋白质复合体的三维结构。该项目将为相互作用蛋白质的功能提供洞察力，并可以指导设计新的试剂来调节蛋白质相互作用网络。研究工作将与生物信息学教育相结合。波士顿大学是首批设立生物信息学研究生学位课程的大学之一，该学位课程由NSF IGERT拨款资助，并且正在迅速增长。该计划的课程开发是这一计划的关键要素。该项目旨在建立一个预测蛋白质对接的通用系统：从未结合的组分蛋白质的坐标开始，通过计算获得复杂结构的模型，包括缔合时的构象变化。研究团队开发了一种初始阶段的对接算法，该算法与目标相似的最好算法进行了竞争性比较。提出的工作包括以下几个部分：(1)对现有的对接算法进行改进，以初步筛选为目标。(2)开发了一种新的算法来细化由初始阶段算法产生的结构。(3)开发新的对接算法目标函数。(4)广泛使用基准测试集和社区范围的盲测试来测试算法。(5)建立网络可访问、全自动化、用户友好的对接服务器。在这个项目中产生的结果和蛋白质复合体数据库也将在服务器上提供。该项目的成果将被整合到两门研究生课程(生物医学应用的计算基因组学和蛋白质工程)和一个生物信息学教师研讨会中。课程和研讨会还将包括结构和功能蛋白质组学和生物信息学领域的其他最新成果。

Genomic and proteomic projects have produced a vast amount of biological data: fully sequenced genomes, gene and protein sequences, protein structures, protein-protein interactions, and gene expression patterns. The next grand challenge in Bioinformatics is to determine the biological functions and cellular roles of all genes and proteins, and their interactions with one another to form a functional cell. The research problem aims to computationally predict the 3-dimensional structure of a protein complex from its components. This project will provide insights into the functions of the interacting proteins and can guide the design of novel reagents to regulate protein interaction networks. The research efforts will be integrated with Bioinformatics education. Boston University is among the first to establish a graduate degree program in Bioinformatics, funded by an NSF IGERT grant and growing rapidly. The curriculum development of the Program is a key element of this proposal.This project seeks to build a general system for predictive protein docking: to start with the coordinates of the unbound component proteins and to obtain computationally a model for the complex structure, including the conformational changes upon association. The research team has developed an initial-stage docking algorithm, which compares competitively with the best algorithms with a similar goal. The proposed work includes the following components: (1) To improve the existing docking algorithm, with the goal of initial screening. (2) To develop a new algorithm to refine the structures produced by the initial-stage algorithm. (3) To develop novel target functions for docking algorithms. (4) To test algorithms extensively using a benchmark set and community wide blind tests. (5) To establish a web accessible, fully automated, and user-friendly docking server. Results and the protein complex database produced in this project will also be available at the server. The results of this project will be integrated into two graduate courses (Computational Genomics and Protein Engineering for Biomedical Applications) and one Bioinformatics Teacher's workshop. The courses and workshop will also include other state-of-the-art results in the fields of structural and functional Proteomics and Bioinformatics.