Signal processing software for biomolecular NMR

University Of Connecticut Sch Of Med/Dnt, Farmington CT

Investigators

Linked publications & trials

Abstract

The development of very high-field magnets, cryogenic probes, and relaxation-optimized pulse sequences have dramatically increased the size and complexity of biomacromolecules amenable to study by NMR spectroscopy. In contrast, the computational methods most commonly used to analyze NMR data have remained essentially unchanged for more than two decades. These methods have well-documented shortcomings, for example requiring a nearly four-fold increase in the amount of data that must be collected in a three-dimensional experiment at 900 MHz in order to achieve the same frequency resolution as an experiment at 500 MHz. In practice this requirement is rarely met, limiting the potential resolution of high-field experiments. A number of modern methods of spectrum analysis have been developed that avoid these shortcomings, but they have not found routine application in NMR. Obstacles to their use include non- intuitive adjustable parameters and the complexity of the available software. In addition, the methods are all nonlinear, and thus prudent application requires careful error analysis. We propose to develop software to enable new applications of maximum entropy reconstruction in biomolecular NMR. The applications include nonuniform sampling in the time domain, to reduce data collection requirements while improving resolution, and deconvolution for performing "virtual decoupling". In addition to software to support multidimensional experiments and efficient computation using loosely-coupled clusters of computers, we will develop new tools for error analysis. We plan to develop a facile user interface, documentation, tutorials, and tools to support interoperability with other software packages, in order to make these advanced data processing and analysis capabilities accessible to non-experts in the broader biomolecular NMR community. The application of modern spectrum analysis in biomolecular NMR will enable the full potential of modern magnets, probes, and pulse sequences to be realized for solving challenging problems in structural biology.