Collaborative Research: ABI Development: The next stage in protein-protein docking

蛋白质-蛋白质相互作用是许多细胞控制机制的组成部分，包括蛋白质定位、变构和基因调控以及信号转导，因此，它们的表征是系统生物学实验和计算方法的重要任务。全基因组蛋白质组学研究提供了越来越多的推测蛋白质-蛋白质相互作用的清单，并证明了大多数蛋白质(如果不是所有的话)在细胞中都有相互作用的伙伴。然而，这些技术只能确定两种蛋白质是否可能相互作用。对蛋白质如何结合和形成复合体的全面理解只能来自高分辨率的三维结构，因为它们提供了理解相互作用如何发生以及如何实现高度特异性所必需的原子细节。虽然X射线结晶学提供了最完整的结构表征，但许多生物上重要的相互作用发生在微弱的、瞬时的复合体中，即使在两种蛋白质都可以分离并确定其结构的情况下，也无法进行直接的实验分析。该项目的目标是进一步改进蛋白质-蛋白质对接服务器ClusPro，该服务器从组分蛋白质的结构出发，试图以接近X射线结晶学提供的精度确定其复合体的结构。Https://cluspro.org的服务器已经拥有近10,000名注册用户，尽管它也可以在不注册的情况下使用。由ClusPro生成的对接结构已在600多篇研究论文中报道。公共服务器对于生物和化学科学家来说很重要，他们可能没有广泛的计算经验，但仍然能够使用最先进的对接方法。蛋白质复杂结构的增加将对生物学、生物化学和生物技术的许多领域产生重大影响。该项目有六个目标，重点是难度越来越大的问题。首先，流形快速傅立叶变换(FFT)相关算法的早期发展提供了一个新的进步机会，该算法将采样速度提高了数量级。该方法的另一个优点是增加相关函数项的数量在计算上是不昂贵的。因此，目标1是开发和实现复杂的能量函数，包括用于得分的距离相关势和三体势能。其次，最近开发的方法在利用进化协方差信息预测蛋白质残基之间的接触方面显示出相当大的前景。问题是，这些方法需要大量进化上相关的相互作用蛋白序列来强有力地评估残基共变异的程度。目标2是探索如何使用序列信息来提取蛋白质间的接触，以改善对接结果，即使在序列数量有点有限的情况下也是如此。第三，随着蛋白质复杂结构数量的增加，越来越需要集成直接对接和基于模板的预测方法，并将探索几种策略来实现这种集成。第四，对接方法将进行修改，以优化相互作用蛋白质的同源模型，而不是X射线结构。第五个目标是探索在结合时具有实质性骨架构象变化的蛋白质对接的新方法。第六个也是最终目标是在ClusPro服务器上实现新开发的算法。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Protein-protein interactions are integral to many mechanisms of cellular control, including protein localization, allosteric and gene regulation and signal transduction, and therefore their characterization is an important task for both experimental and computational approaches to systems biology. Genome-wide proteomics studies provide a growing list of putative protein-protein interactions, and demonstrate that most if not all proteins have interacting partners in the cell. However, these techniques can only identify whether two proteins possibly interact. A full comprehension of how proteins bind and form complexes can only come from high-resolution three-dimensional structures, since they provide the atomic details necessary to understand how the interactions occur and how the high degree of specificity can be achieved. While the most complete structural characterization is provided by X-ray crystallography, many biologically important interactions occur in weak, transient complexes that are not amenable to direct experimental analysis, even when both proteins can be isolated and their structures determined. The goal of this project is to further improve the protein-protein docking server ClusPro that, starting from the structures of component proteins, attempts to determine the structure of their complexes with accuracy close to that provided by X-ray crystallography. The server at https://cluspro.org already has close to 10,000 registered users, although it can also be used without registration. Docked structures generated by ClusPro have been reported in over 600 research papers. The public server is important for biological and chemical scientists who may not have extensive computational experience but still will be able to use state of the art docking methods. The increased availability of protein complex structures will have major impact in many areas of biology, biochemistry, and biotechnology. The project has six aims focusing on problems of increasing difficulty. First, a new opportunity to progress is provided by earlier development of the manifold fast Fourier transform (FFT) correlation algorithm, which increased the speed of sampling by orders of magnitude. An additional advantage of the method is that increasing the number of correlation function terms is computationally inexpensive. Therefore, Aim 1 is the development and implementation of complex energy functions, including distance-dependent and three-body potentials for scoring. Second, recently developed methods have shown considerable promise in predicting contacts between residues in proteins using evolutionary covariance information. The problem is that these methods require large numbers of evolutionarily related sequences of interacting proteins to robustly assess the extent of residue covariation. Aim 2 is exploring how sequence information can be used for extracting inter-protein contacts to improve docking results even with somewhat limited number of sequences. Third, with the increase in the number of protein complex structures, there is increasing need for integrating direct docking and template-based prediction methods, and several strategies will be explored to implement such integration. Fourth, the docking method will be modified to work optimally with homology models rather than X-ray structures of interacting proteins. The fifth aim is exploring new approaches to the docking of proteins with substantial backbone conformational changes upon binding. The sixth and final aim is implementing the newly developed algorithms in the ClusPro server.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.