Novel Algorithms to expedite experimental protein structure determination

尽管大分子结构在揭示生命的生化基础方面具有重要意义，但实验结构的确定仍然是一项长期、困难和昂贵的努力。人们对蛋白质结构的兴趣源于一个普遍的信条，即蛋白质的功能是由蛋白质结构决定的。这表明，蛋白质结构的知识不仅有助于分析野生型（或天然）蛋白质，也有助于分析与疾病过程有关的突变蛋白。因此，蛋白质结构的知识已经成为药物发现和开发的关键资源。在未来，大分子结构的测定将是一个自动化的交钥匙操作，应该是快速、廉价和高产的。我建议的研究重点是开发准确和高效的算法，以加快实验蛋白质结构测定的过程。拟议的研究计划将结构生物学的知识与机器学习，机器视觉和计算搜索的算法相结合。我的跨学科研究项目的长期目标是开发和应用计算机科学和工程技术来解决结构生物学（蛋白质结构和功能的研究）中的问题。该提案中强调的研究代表了我实验室一系列项目的一部分，包括用于实验结构确定的算法开发、蛋白质设计、基于结构的药物设计和化学合成计划。在本提案中，我们描述了几种新的计算方法来自动化x射线晶体学和电子低温显微镜中的当前实验瓶颈。随着有用的算法里程碑的实现，我们将打包我们的软件，并使其免费提供给研究社区。

Despite the importance of macromolecular structure in unraveling the biochemical basis of life, experimental structure determination remains a long, difficult, and expensive endeavour. The interest in protein structures is motivated by the general mantra that protein function is dictated by protein structure. This suggests that knowledge of protein structure is helpful not only in analysis of wildtype (or natural) proteins but also for mutant proteins such as those implicated in disease processes. Consequently, knowledge of protein structure has emerged as a critical resource for drug discovery and development. In the future, macromolecular structure determination will be an automated turn-key operation, and should be fast, inexpensive, and high-yield. My proposed research focuses on developing accurate and efficient algorithms to expedite the process of experimental protein structure determination. The proposed research program integrates knowledge from structural biology with algorithms in machine learning, machine vision, and computational search. The long-term goals of my interdisciplinary research program are to develop and apply techniques from Computer Science and Engineering to problems in Structural Biology (the study of protein structure and function). The research highlighted in this proposal represents part of a family of projects in my lab including algorithm development for experimental structure determination, protein design, structure-based drug design, and chemical synthesis planning. In this proposal, we describe several novel computational approaches for automating current experimental bottlenecks in X-Ray Crystallography and Electron CryoMicroscopy. As useful algorithmic milestones are achieved, we will package our software and make it freely available to the research community.