Macromolecular Docking

描述（由申请人提供）：可靠和有效的蛋白质相互作用几何预测仍然是一个未解决的问题。然而，它是未来结构生物学的关键技术。我们的长期目标是开发可靠的蛋白质对接预测程序，以促进我们对生物体中发生的复杂调节和代谢过程的结构理解，并使设计药物能够阻断或改变这些相互作用。目前的对接技术主要有两个瓶颈。首先，两种蛋白质在结合时会发生诱导的构象变化，而目前的方法无法充分处理这种灵活性。其次，对接过程的初始阶段，即所谓的刚体对接，速度太慢，无法考虑多种受体和配体构象。我们的主要目标是克服这些瓶颈，并开发新的方法来预测发生骨干变形的蛋白质对接。

DESCRIPTION (provided by applicant): Reliable and efficient predictions of protein interaction geometry remains an unsolved problem. Yet it is a key technology for the structural biology of the future. Our long term goal is to develop reliable protein docking prediction procedures to facilitate our structural understanding of the complex regulatory and metabolic processes that occur in living organisms, and to enable design of drugs for blocking or modifying these interactions. There are two principal bottlenecks in the current docking technology. Firstly, two proteins undergo induced conformational changes upon association, while the current methods cannot adequately deal with such flexibility. Secondly, the initial phase of the docking procedure, so called rigid body docking, is too slow to consider multiple receptor and ligand conformations. Our main goal is to overcome these bottlenecks and develop new methods to predict docking of proteins undergoing backbone deformations. The results of the first Critical Assessment of PRotein Interaction predictions (CAPRI) revealed some modest but hopeful signs in dealing with the induced fit. Here we propose to further develop the docking methodology by incorporating recent density matching algorithms, the latest achievements in small molecule docking, as well as by finding an efficient way to deal with both side-chain and backbone flexibility. To speed up the rigid body docking we will adapt a new Five Dimensional Fast Fourier Transform method for simultaneous determination of rotation and translation to protein docking. This will lead to a quick generation of trial solutions and will allow exploring multiple conformations of the interacting partners. Secondly, new trial solution re-scoring schemes and better grid potential maps used by the 5DFFT method will be developed and optimized on a benchmark of known complexes. Thirdly, better ways to treat explicit protein side-chain flexibility will be developed. Finally, we will study the phenomenon of backbone rearrangements upon association and predict multiple conformations for rigid body docking, as well as incorporate the rearrangements into refinement procedure. The emerging procedures will be applied to biological problems and tested in the CAPRI docking competition.