Interpolating and extrapolating with hmsIST: seeking a t_max for optimal sensitivity, resolution and frequency...

Area of Science:

• Structural Biology
• Biophysical Chemistry
• Nuclear Magnetic Resonance Spectroscopy

Background:

• High-field Nuclear Magnetic Resonance (NMR) instruments offer optimal resolution, but traditional uniform sampling in 3D/4D experiments is limited by long measurement times.
• Extrapolating uniformly sampled data to higher resolution using methods like linear prediction or iterative soft thresholding (IST) can compromise peak position accuracy.

Purpose of the Study:

• To investigate the optimal experimental maximum evolution time (t_max) for non-uniform sampling (NUS) in NMR to maximize sensitivity and peak accuracy.
• To compare the performance of NUS with uniform sampling (US) for high-resolution NMR data acquisition.

Main Methods:

• Development of time-equivalent NMR experiments to systematically evaluate NUS strategies.
• Utilizing PoissonGap sampling and the hmsIST reconstruction method.
• Assessing sensitivity via signal-to-maximum-noise ratio and evaluating peak position accuracy.

Main Results:

• Non-uniform sampling, specifically PoissonGap sampling to a t_max of approximately ½*T₂^*, significantly improves both sensitivity and resolution compared to traditional uniform sampling.
• Extrapolation of data to 2*T₂^* further enhances digital resolution without requiring experimental data points beyond ½*T₂^*.
• NUS followed by appropriate reconstruction (hmsIST) largely improves peak position accuracy, a limitation of US extrapolation.

Conclusions:

• Non-uniform sampling is a superior strategy for exploiting the full resolution potential of high-field NMR instruments, especially for complex 3D and 4D experiments.
• Optimizing the sampling duration (t_max) in NUS is crucial for achieving the best balance between sensitivity, resolution, and accuracy.
• PoissonGap sampling to ½*T₂^* followed by extrapolation represents an effective approach for high-resolution NMR data processing.