PSI Structural Biology Knowledgebase

PSI | Structural Biology Knowledgebase
Header Icons

Related Themes

Membrane Proteome: Soft Sampling

SBKB [doi:10.1038/sbkb.2012.114]
Technical Highlight - December 2012
Short description: A novel method for processing non-uniformly sampled NMR data is orders of magnitude faster and permits very sparse sampling for challenging applications.

Outline of the IST procedure for processing NMR data containing signals of varying intensities. Figure courtesy of Gerhard Wagner and Sven Hyberts.

NMR spectroscopy can now tackle many challenging systems, including protein-protein complexes, nucleic acids, and membrane proteins. Such applications have relied on the development of higher magnetic field strengths and the resolution and sensitivity provided by these instruments. The sampling requirements at higher magnetic fields, however, have a drawback: optimal resolution cannot be achieved with current experimental schemes since, for typical measurement times, more data points would be required to fully characterize the range of frequencies in the spectrum. This is particularly evident in higher dimensional 3D and 4D data.

These developments have led to a renewed interest in non-uniform sampling (NUS) methods. In contrast to uniform sampling, NUS sparsely samples data points and, in principle, promises to deliver optimal resolution in a fraction of the time while using instruments to full potential. Whereas data processing for regular sampling schedules uses a simple Fourier transformation (FT), NUS data reconstruction typically requires alternative, and computationally intensive, digital signal processing methods.

Wagner and colleagues (PSI MPSbyNMR) have developed an algorithm based on iterative soft thresholding (IST), an approach previously used to reconstruct signals from the noisy data of MRI and other imaging methods. Using a Poisson-gap sampling scheme, the authors found that IST can be used in conjunction with iterative FT to gradually extract the signal from noise-like artifacts resulting from NUS. Their scheme, implemented in a convenient graphical user interface (GUI), has the significant advantage of using FT, dramatically reducing processing times.

The authors applied their method to several data sets, including a 4D spectrum of a 28 kDa protein-protein complex. The 4D spectrum was acquired with a sampling density of 14.5% and processed orders of magnitude faster, revealing many methyl-methyl interactions. A further experiment aimed at achieving fully optimal resolution using the same 4D experiment sampled at only 0.8% reproduced known assignments. An important requirement for any processing method is the faithful reproduction of relative signal intensities used to generate distance restraints for structure calculations. A comparison of intensities in two IST-processed spectra with equivalent uniformly sampled data demonstrated excellent fidelity in relative peak intensities.

The IST method will be particularly beneficial for α-helical membrane proteins where methyl-methyl contacts provide invaluable structural data. Combined with the convenient GUI and fast processing times, the IST method promises to be widely adopted.

Michael A. Durney


  1. S. G. Hyberts et al. Application of iterative soft thresholding for fast reconstruction of NMR data non-uniformly sampled with multidimensional Poisson gap scheduling.
    J Biomol NMR. 52, 315-327 (2012). doi:10.1007/s10858-012-9611-z

Structural Biology Knowledgebase ISSN: 1758-1338
Funded by a grant from the National Institute of General Medical Sciences of the National Institutes of Health