Non-uniform sampling of NMR relaxation data

Area of Science:

• Nuclear Magnetic Resonance (NMR) Spectroscopy
• Biophysical Chemistry
• Structural Biology

Background:

• Non-uniform sampling (NUS) in NMR spectroscopy offers substantial reductions in data acquisition time.
• However, its application in quantitative experiments, such as spin relaxation rate measurements, is limited due to potential inaccuracies in peak intensities and subsequent errors in dynamic parameter extraction.

Purpose of the Study:

• To evaluate the performance of multi-dimensional decomposition and iterative re-weighted least-squares algorithms for reconstructing NMR spectra from non-uniformly sampled data.
• To determine the minimum data coverage required for accurate peak intensity measurements in (15)N Carr-Purcell-Meiboom-Gill (CPMG) relaxation dispersion experiments.

Main Methods:

• Systematic reduction of Nyquist grid coverage for (15)N CPMG relaxation dispersion datasets from four different proteins.
• Comparative analysis of multi-dimensional decomposition and iterative re-weighted least-squares algorithms on non-uniform sampled (NUS) data.
• In silico reduction of spectrum quality from a fully sampled spectrum to estimate reliable NUS coverage levels.

Main Results:

• Multi-dimensional decomposition accurately reconstructs spectra from NUS data when at least one fully sampled spectrum is present in the dataset.
• For some datasets, 20% data coverage yielded results comparable to fully sampled data.
• However, such low coverage is not universally sufficient for reliable quantitative analysis, necessitating careful estimation of coverage levels.

Conclusions:

• Incorporating a single fully sampled spectrum enables accurate spectral reconstruction from NUS data using multi-dimensional decomposition.
• The study provides a method to estimate adequate NUS coverage by analyzing in silico reduced-quality spectra.
• This approach facilitates faster quantitative NMR experiments without compromising data accuracy.