AF: Small: An Integrated Approach to Characterizing Conformational Changes of Large Proteins

理解生理过程和新药物设计的主要障碍之一是对分子三维结构和灵活性的理解，这对它们的功能和相互作用有关键的了解。在许多生物学途径中，已知构象灵活性在驱动功能机制中起着至关重要的作用，并且一直是计算和实验研究的主题。现代实验方法，如氢/氘交换耦合质谱（HDX-MS）可以探测大蛋白质的灵活性，但这些发现是在可用的（静态）晶体结构上映射的，从而失去了相当一部分动态信息。本提案将研究计算方法在多大程度上可以帮助发现和可视化与实验一致的大蛋白质的构象变化。由于在本提议的背景下选择研究的特定蛋白质，提议的工作具有巨大的社会影响：它有助于促进对涉及各种免疫，炎症，神经退行性，缺血性和年龄相关疾病的过程的理解。该计划将在莱斯大学和宾夕法尼亚大学就跨学科课题培训本科生、研究生和博士后。它还将为女学生提供培训机会。所有工作不仅将通过在国家会议上的技术介绍、在同行评议的期刊和书籍章节中发表的出版物，而且还将通过访问和与学术界和生物技术工业的研究人员会谈，以及通过开放源码开发和分发为广泛使用而开发的算法来传播。该建议的核心部分涉及到对蛋白质构象空间进行采样的计算框架的发展。它依赖于蛋白质和蛋白质复合物的自由度的新表示以及在不同分辨率下采样原语的定义。该框架借鉴了受几何启发的采样技术，这些技术可以有效地探索大型搜索空间，同时跟踪搜索空间所覆盖的范围。同时，将进行HDX-MS实验，并将其用作构象变异性的探针。实验结果将与计算框架的结果相关联，以选择最接近实验的构象。这项工作的一个显著特点是与实验的联系，以及两者相互促进的方式。这一建议是算法、计算几何、降维、数据挖掘和可视化方法的交叉，并与现代实验方法联系在一起，这些方法正在积极发展，对理解蛋白质的结构和功能有很大的希望。

One of the major hurdles in the understanding of physiological processes and the design of new pharmaceutical drugs has been the understanding of the three-dimensional structure and flexibility of molecules, which hold key insight into their function and interactions. In many biological pathways conformational flexibility is known to play an essential role in driving functional mechanisms and has been the topic of both computational and experimental investigations. Modern experimental methods such as Hydrogen/Deuterium Exchange coupled with Mass Spectroscopy (HDX-MS) can probe the flexibility of large proteins, but the findings are mapped on the available (static) crystal structures, thereby losing a considerable part of the dynamic information. This proposal will investigate the extent to which computational methods can help discover and visualize conformational changes of large proteins that are in agreement with experiment. Because of the specific proteins selected for study in the context of this proposal, the proposed work has a great societal impact: it helps advance the understanding of processes that are involved in a variety of immune, inflammatory, neurodegenerative, ischemic, and age-related diseases. This proposal will train undergraduate, graduate, and postdoctoral students at Rice and the University of Pennsylvania on interdisciplinary topics. It will also provide training opportunities to women students. All work will be disseminated not only through technical presentations at national meetings, publications in peer reviewed journals and book chapters, but also through visits and talks with researchers in academia and the biotechnology industry, as well as with open source development and distribution of the algorithms developed for widespread use.A central part of this proposal concerns the development of a computational framework that samples the conformational space of proteins. It relies on a new representation of the degrees of freedom of proteins and protein complexes and on the definition of sampling primitives at various resolutions. The framework draws from geometry-inspired sampling techniques that efficiently explore large search spaces while keeping track of the extent that the search space is covered. In parallel, HDX-MS experiments will be carried out and used as probes for conformational variability. The experimental results will be correlated with the results of the computational framework to select the conformations that most closely match experiments. A distinguishing feature of this work is the connection with experiment and the ways in which each can feed the other. This proposal lies at the intersection of algorithms, computational geometry, dimensionality reduction, data mining, and visualization methods and connects with modern experimental methods which are actively under development and hold great promise for the understanding of protein structure and function.