Valafar Homayoun, Prestegard James H
Southeast Collaboratory for Structural Genomics, Department of Biochemistry and Molecular Biology, University of Georgia, GA 30602, USA.
J Magn Reson. 2004 Apr;167(2):228-41. doi: 10.1016/j.jmr.2003.12.012.
Recent advancements in the utilization of residual dipolar couplings (RDCs) as a means of structure validation and elucidation have demonstrated the need for, not only a more user friendly, but also a more powerful RDC analysis tool. In this paper, we introduce a software package named REsidual Dipolar Coupling Analysis Tool (REDCAT) designed to address the above issues. REDCAT is a user-friendly program with its graphical-user-interface developed in Tcl/Tk, which is highly portable. Furthermore, the computational engine behind this GUI is written in C/C++ and its computational performance is therefore excellent. The modular implementation of REDCAT's algorithms, with separation of the computational engine from the graphical engine allows for flexible and easy command line interaction. This feature can be utilized for the design of automated data analysis sessions. Furthermore, this software package is portable to Linux clusters for high throughput applications. In addition to basic utilities to solve for order tensors and back calculate couplings from a given order tensor and proposed structure, a number of improved algorithms have been incorporated. These include the proper sampling of the Null-space (when the system of linear equations is under-determined), more sophisticated filters for invalid order-tensor identification, error analysis for the identification of the problematic measurements and simulation of the effects of dynamic averaging processes.
利用剩余偶极耦合(RDC)作为结构验证和阐释手段的最新进展表明,不仅需要一个更用户友好的,而且需要一个更强大的RDC分析工具。在本文中,我们介绍了一个名为剩余偶极耦合分析工具(REDCAT)的软件包,旨在解决上述问题。REDCAT是一个用户友好的程序,其图形用户界面是用Tcl/Tk开发的,具有很高的可移植性。此外,这个图形用户界面背后的计算引擎是用C/C++编写的,因此其计算性能非常出色。REDCAT算法的模块化实现,将计算引擎与图形引擎分离,允许灵活且容易的命令行交互。这个特性可用于设计自动化数据分析会话。此外,这个软件包可移植到Linux集群上用于高通量应用。除了用于求解序张量以及从给定序张量和提议结构反算耦合的基本实用工具外,还纳入了许多改进算法。这些算法包括零空间的适当采样(当线性方程组欠定时)、用于识别无效序张量的更复杂滤波器、用于识别问题测量值的误差分析以及动态平均过程效果的模拟。