ABI Innovation: Synergistic application of cheminformatics and computational geometry approaches for predicting protein-protein interactions

位于查佩尔山的北卡罗来纳州大学获得了一个奖项，以开发和交付最终以用户友好的软件形式的分析蛋白质-蛋白质相互作用（PPI）的新方法，该方法在结构生物信息学，计算几何学和化学信息学的界面上采用了独特的概念组合。蛋白质-蛋白质相互作用在活细胞中发生的所有主要信号事件中起着核心作用。基于先前开发的SNAPP方法，将开发一个新的PPI化学信息学描述符家族，该化学信息学描述符来自PPI界面的Delaunay镶嵌，称为SNAP 3（蛋白质-蛋白质包装的简单邻域分析）。这些描述符将被用作新方法的一部分，以（1）识别蛋白质表面的热点区域和（2）使用本项目开发的新GridDock对接方法预测蛋白质-蛋白质（或蛋白质-肽）复合物的结构。在后基因组时代，PPI的分析和预测被认为是阐明许多（如果不是大多数）蛋白质功能的关键，以更好地理解细胞生物学，细胞网络和人类疾病。该项目将为实验和计算界提供分析和预测PPI的新的和强大的方法和软件，并实现几个重要的应用，包括：（1）使用化学计量学方法对PPI进行独特的描述和比较;（2）识别蛋白质表面上未知的PPI热点;和（3）合理设计对靶蛋白具有所需结合特异性的新肽。它将产生广泛的影响，使所有PPI数据集，计算工具和模型开发通过一个新的模块在公开的框架，ChemBench门户网站访问。

An award is made to the University of North Carolina at Chapel Hill to develop and deliver, ultimately in the form of user-friendly software, novel approaches to analyzing protein-protein interactions (PPIs) that employ a unique combination of concepts at the interface of structural bioinformatics, computational geometry, and cheminformatics. Protein-protein interactions play a central role in all major signaling events that occur in living cells. A new family of cheminformatics descriptors of PPIs derived from Delaunay tessellation of PPI interface termed SNAP3 (Simplicial Neighborhood Analysis of Protein-Protein Packing), based on the previously-developed SNAPP approach, will be developed. These descriptors will be employed as part of novel approaches to (1) identify hot spot regions on protein surfaces and (2) predict the structure of protein-protein (or protein-peptide) complexes using novel GridDock docking method to be developed in this project. In the post-genomic era, analysis and prediction of PPIs is considered critical for elucidating of many if not most protein functions to better our understanding of cell biology, cellular networks and human diseases. This project will provide both the experimental and computational communities with novel and powerful methodologies and software for analyzing and predicting PPIs and enable several important applications including: (1) unique description and comparison of PPIs using chemometric approaches; (2) the identification of unknown PPI hot spots on protein surfaces; and (3) rational design of novel peptides with desired binding specificity against target proteins. It will have a broad impact by making all PPI datasets, computational tools, and models developed accessible via a new module within the publicly available framework, the ChemBench web portal.