OPTIMIZATION OF MULTIDIMENSIONAL NMR DATA ANALYSIS

多维核磁共振数据分析的优化

• 批准号: 3308869
• 负责人: RAY LIU
• 金额: $ 9.61万
• 依托单位: UNIV OF MARYLAND, COLLEGE PARK
• 依托单位国家: 美国
• 财政年份: 1993
• 资助国家: 美国
• 起止时间: 1993-08-01 至 1996-07-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/3308869
• 关键词: artificial intelligence; bioengineering /biomedical engineering; computer program /software; computer system design /evaluation; mathematical model; method development; model design /development; nuclear magnetic resonance spectroscopy; parallel processing; protein structure

The advent of efficient methods for isotopic enrichment of proteins with 13C and 15N, together with the development of three- and four-dimensional (3D and 4D) NMR methods, now makes it possible to study in detail the structure and dynamics of proteins in the 15-30 kD molecular weight range, as was demonstrated most recently for the proteins interleukin- 1beta, calmodulin complexed with a fragment of myosin light chain kinase and interferon-gamma. However, the present procedures for data collection and analysis are extremely time consuming, and the spectral resolution obtainable frequently presents a limiting factor. The goal of this proposal is the development of improved and new procedures for the analysis of multi-dimensional NMR spectra. Significant improvements over present methodology are expected to be possible considering that most of the methodology presently does not utilize the advances which have been made in the last decade in the area of signal processing. Improved and new procedures will focus on but not limit to the following: (1) Improvement of the linear prediction methods: improving the stability and computation efficiency of the 1-D and 2-D linear prediction algorithms. (2) Optimization of a priori information: Use of a priori information in NMR data to improve the spectral resolution. (3) Non-linear spectral enhancement: Development of the 3-D maximum entropy method and improving the convergence rate of methods based on iterative deconvolution. (4) Development of fast algorithms: Development of fast routines that make possible the use of algorithms in multi-dimensional space. (5) signal subspace approach: Separation of the signal and the noise subspace to obtain high NMR spectral resolution. (6) Higher-order spectra estimation: Use of high- order statistics to gain a better estimate of dense NMR spectra. (7) Multi-resolution techniques: New signal representation for NMR data to achieve optimal interpretation and processing. (8) Data compression: Compression of multi-dimensional NMR data by removing redundant information for processing, networking, and archiving. (9) Parallel implementation: Development of parallel routines for high-performance parallel computing. Emphasis will be on the development of code which is easily transportable and which will be made freely available to the NMR community. The complexities of the task of developing algorithms and software for advanced processing of multi-dimensional NMR data are substantial and require effective integration of expertise in the areas of NMR and signal processing. This proposal is designed to foster such integration.

有效方法的出现，用于蛋白质的同位素富集 13c和15n，以及三维和四维的发展 （3D和4D）NMR方法，现在可以详细研究 15-30 kD分子量中蛋白质的结构和动力学 范围，正如最近在蛋白质白介素中所证明的那样 1beta，钙调蛋白与肌球蛋白轻链激酶的片段复合 和干扰素 - 伽马。 但是，目前的数据程序 收集和分析非常耗时，光谱 可获得的分辨率经常提出一个限制因素。 该提议的目的是发展改进和新的 分析多维NMR光谱的程序。 对当前方法的显着改善有望是 考虑到目前的大多数方法都不是可能的 利用该地区过去十年中已取得的进步 信号处理。 改进，新程序将重点关注但不关注 限制以下限制：（1）改进线性预测 方法：提高1-D的稳定性和计算效率 和2-D线性预测算法。 （2）优先化 信息：在NMR数据中使用先验信息来改进 光谱分辨率。 （3）非线性光谱增强：开发 3-D最大熵方法的收敛速率 基于迭代反卷积的方法。 （4）快速发展 算法：开发快速例行程序，使其成为可能的使用 多维空间中的算法。 （5）信号子空间方法： 信号和噪声子空间的分离以获得高NMR 光谱分辨率。 （6）高阶光谱估计：使用高级 命令统计数据以更好地估计密集的NMR光谱。 （7） 多分辨率技术：NMR数据的新信号表示形式 实现最佳的解释和处理。 （8）数据压缩： 通过删除冗余来压缩多维NMR数据 用于处理，网络和归档的信息。 （9）平行 实施：开发高性能的平行程序 并行计算。 重点将是易于运输的代码的开发 并将免费提供给NMR社区。 为开发算法和软件的任务的复杂性 多维NMR数据的高级处理是实质的，并且 需要在NMR和信号领域有效整合专业知识 加工。 该建议旨在促进这种整合。