Integrated resource for reproducibility in macromolecular crystallography

大分子晶体学重现性的综合资源

• 批准号: 9069902
• 负责人: WLADEK MINOR
• 金额: $ 47.19万
• 依托单位: UNIVERSITY OF VIRGINIA
• 依托单位国家: 美国
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-06-01 至 2018-05-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9069902
• 关键词: 3-Dimensional; Address; Algorithmic Software; Algorithms; Archives; Biological; Calibration; Collection; Communities; Computer software; Crystallography; Data; Data Analyses; Data Set; Databases; Deposition; Detection; Development; Diffuse; Discipline; Disease; Educational workshop; Ensure; Fraud; Goals; Health; Ice; Image; Image Analysis; Infection; Lead; Learning; Libraries; Life; Ligands; Link; Location; Metadata; Methods; Microscopic; Modeling; Molecular; Morphologic artifacts; Network-based; Online Systems; Procedures; Process; Proteins; Protocols documentation; Reproducibility; Research; Research Personnel; Resolution; Resources; Roentgen Rays; Semantics; Site; Software Tools; Structural Biologist; Structure; Synchrotrons; System; Technology; Technology Transfer; Testing; Training; Twin Multiple Birth; Validation; Vendor; Work; X-Ray Crystallography; beamline; beneficiary; cell dimension; data mining; data reduction; data wrangling; density; design; detector; electron density; geometric methodologies; heuristics; improved; member; prevent; programs; repository; research study; statistics; structural biology; structural genomics; syntax; tool; tool development; working group

 DESCRIPTION (provided by applicant): We propose the development of a collection of data wrangling tools to store, parse, manipulate, validate, curate, analyze, and disseminate macromolecular diffraction images together with all associated relevant metadata. The proposed system will have several benefits, by (1) creating a means to improve existing structures as technology for processing diffraction image advances, (2) detecting errors (and potentially, fraud) in existing structures to ensure structure quality and reproducibility, (3) preventing the loss of data collected by structural genomics and other programs that have closed or will close, (4) providing data for analysis of diffuse diffraction effects, and (5) buildng a "training set" for new diffraction analysis algorithms and hardware. Biologists, bioinformaticians, and software and hardware developers will all be beneficiaries of these tools. The proposed research is designed for semantic rather than syntactic analysis of diffraction images, and has several specific goals. First, we will develop tools for automatically extracting and curating diffraction images and associated metadata, as well as producing descriptions of all data needed for reprocessing when methods for structure determination improve. Second, we will create a web-based system for organizing, searching, analyzing, and data mining of appropriate subsets of diffraction images and associated metadata in machine- readable formats. This will include a comprehensive API for programmatic access, the ability to link multiple instances into a distributed federation, and state-of-the-art compression and transfer technologies. Third, we will develop tools to automatically validate, preprocess, and score diffraction images, and to detect potential issues and errors. These tools will make use of new and existing programs for image and data analysis, contain heuristics to identify possible errors, and provide statistics to correlate errors with specific metadata. Fourth, we will create a mechanism to discover diffraction data that have not yielded X-ray structures with currently available methods. Fifth, we will set up a pilot resource incorporating all the developed tools, and collect a test data set for the development of new tools for validation and error detection. We will work closely with multiple collaborators. Most important is the RCSB Protein Data Bank (PDB), who will help us ensure the accuracy and completeness of the diffraction metadata. Other partners will include the diffuse X-ray scattering community, detector vendors, synchrotron beamline managers, members of the IUCr Diffraction Data Deposition Working Group (DDDWG) and the crystallographic community in general. Together with the RCSB PDB, we will organize workshop(s) with these communities in order to (a) improve metadata extraction and (b) better define subsets of diffraction images. By addressing the currently common, irreversible and unnecessary loss of raw diffraction data during the data reduction process, our project helps ensure that the discipline of macromolecular crystallography is capable of continuous self-improvement.