Flexible Macromolecular Crystallography

柔性高分子晶体学

• 批准号: 10506287
• 负责人: James M Holton
• 金额: $ 46.54万
• 依托单位: UNIVERSITY OF CALIF-LAWRENC BERKELEY LAB
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2027-08-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10506287
• 关键词: Acoustics; Active Sites; Age; Artificial Intelligence; Attention; Binding; Biological; Biological Assay; Biological Sciences; Blood capillaries; Chemicals; Collaborations; Collection; Communities; Complement; Computer software; Cryoelectron Microscopy; Crystallization; Crystallography; Data; Data Analyses; Data Collection; Data Set; Decision Making; Devices; Diagnostic; Diagnostic Procedure; Diagnostic tests; Dimethyl Sulfoxide; Disease; Drops; Elements; Epiphysial cartilage; Familiarity; Feedback; Fostering; Geography; Goals; Growth; Harvest; Health; Image; In Situ; Individual; Libraries; Ligands; Liquid substance; Location; Logistics; Machine Learning; Mathematics; Measurement; Measures; Metals; Methods; Modeling; Molecular; Molecular Conformation; Motion; Noise; Optics; Phase; Preparation; Procedures; Process; Property; Proteins; Protocols documentation; Research Project Grants; Resistance; Resolution; Resources; Robot; Robotics; Roentgen Rays; Sampling; Services; Shapes; Signal Transduction; Source; Structure; Synchrotrons; System; Techniques; Technology; Temperature; Tertiary Protein Structure; Testing; Training; Variant; base; beamline; coronavirus disease; data access; data quality; design; experimental study; flexibility; improved; insight; multiple datasets; operation; predictive tools; screening; small molecule; structural biology; success; tool

Project Summary/Abstract - Core 3 – Flexible Macromolecular Crystallography This 3rd Technology Operations Core (TOC3) complements TOC1 by diversifying the ALS-ENABLE technology base to maximize flexibility in this now transformative era for structural biology. Artificial intelligence (AI) has revolutionized solving the phase problem, and we will not only make these new structure prediction tools accessible to our User community, but also other AIs that benefit our workflows, such as object location of sample loops and crystals, diffraction image interpreters or variational auto encoders for modelling protein domain motions. These will be put to use once they are proven effective. For example, we expect to enable efficient yet unattended in-situ serial data collection direct from crystallization trays by training now mature and off-the-shelf AI technology to locate diffraction-quality crystals in their growth drops. If successful, even a modest improvement in hit rate will revolutionize serial data collection using our in-situ goniometer. This in-situ capability also completes a chain of diagnostic tests of the sample preparation process, allowing our Users to understand the origins of poor diffraction and focus their efforts appropriately. This diagnostics chain leverages the capabilities of TOC1 micro-focus, TOC2 solution stability, and TOC4 mapping molecular interfaces. Our uniquely accommodating robotics solution with broad pin compatibility will get a capacity upgrade to help ease the transitions our User community will have to make from synchrotron to synchrotron as APS and then ALS undergo long shutdowns for major upgrades. We will upgrade our X-ray optics to match the properties of the ALS-U source. We will also upgrade robotics to provide remote access data collection at non-cryo temperatures, ranging from -20C to +50C, making these valuable multi-temperature tools accessible to a geographically diverse User community. Functional studies at these temperatures will be assisted by rolling out state-of-the-art difference-data analysis software, such as PanDDA, as part of beamline workflows. By explicitly supporting difference data analysis our users will have access to state-of-the-art technology for visualizing weak yet critical difference features, such as low-occupancy ligands and functionally-relevant conformational shifts. And because fragment screening is a critical tool for the bioscience community to quickly respond to an emerging health crisis, we will support as well as document best practices such as DMSO tolerance testing in our ALS-ENABLE protocols as well as foster collaborations between user groups with access to advanced yet shareable sample preparation tools such as fragment libraries and acoustic drop liquid handlers. Rather than leave users to their own devices to organize and analyze their data, we will deploy ISPyB/SynchWeb, the world’s most heavily used LIMS for structural biology data. Tools for merging multi-crystal data for improved data quality that performed well in our global challenge data set competition will be deployed under this framework. This will not only make cross-synchrotron data analysis available in one place, but maintain a level of familiarity to ease the transition of our Users to and from other synchrotrons. We group our aims by the mathematical operations they entail: adding data together to improve signal (Aim1), modulation of the sample to induce a change (Aim2), and subtraction of data to reveal the result (Aim3).

项目摘要/摘要 - 核心3 - 柔性大分子晶体学 通过多样化ALS-Enable，这个第三技术运营核心（TOC3）完成TOC1 在现在的结构生物学时代，技术基础是最大化灵活性。人工智能 （AI）彻底改变了解决阶段问题的革命性，我们不仅会做出这些新的结构预测 我们的用户社区可以访问的工具，以及其他受益于我们工作流的AI，例如对象位置 样品循环和晶体，衍射图像解释器或用于建模蛋白质的变异自动编码器 域运动。这些被证明有效后，将使用这些。例如，我们希望启用 通过训练现在成熟和 现成的AI技术可在其生长降低中找到衍射质量的晶体。如果成功，甚至 命中率的适度提高将使用我们的原位性角仪彻底改变串行数据收集。这是原地 功能还完成了样本准备过程的一系列诊断测试，使我们的用户可以 了解不良衍射的起源，并适当地集中精力。此诊断链的利用 TOC1微焦点，TOC2溶液稳定性和TOC4映射分子界面的功能。 我们具有广泛销钉兼容性的独特容纳机器人解决方案将获得能力升级以帮助 简化我们的用户社区必须从同步器到同步器作为AP的过渡，然后 ALS进行了长时间的关闭，以进行重大升级。我们将升级X射线光学元件以匹配 ALS-U来源。我们还将升级机器人技术，以在非cryo上提供远程访问数据收集 温度从-20C到 +50C，使这些有价值的多温度工具可访问 地理上不同的用户社区。这些温度下的功能研究将通过推出来帮助 最新的差异数据分析软件，例如PANDDA，作为光束线工作流程的一部分。通过明确 支持差异数据分析我们的用户将可以使用最先进的技术来可视化 弱但关键的差异特征，例如低占用配体和功能相关的构象 转移。而且由于碎片筛查是生物科学社区快速回应的关键工具 新兴的健康危机，我们将支持并记录最佳实践，例如DMSO耐受性测试 在我们的ALS启用协议以及用户组之间的促进协作中，可以访问高级 然而，可共享样品制备工具，例如碎片库和声学滴注液体处理程序。相当 除了将用户留给自己的设备来组织和分析其数据，我们将部署ISPYB/SYNCHWEB， 世界上最严重的LIM用于结构生物学数据。合并多晶数据以改进的工具 在我们的全球挑战数据集竞争中表现良好的数据质量将在此下部署 框架。这不仅可以使跨同步数据分析在一个地方可用，还可以保持一个级别 熟悉的是减轻用户与其他同步基因的过渡。我们将目标分组为 他们需要的数学操作：将数据一起添加以改善信号（AIM1），调制样本 诱导更改（AIM2）和数据减法以揭示结果（AIM3）。