Application of blind source separation to real-time dissolution dynamic nuclear polarization.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Biophysics

Background:

• Hyperpolarized Nuclear Magnetic Resonance (NMR) enables rapid, high-sensitivity measurements of dynamic processes.
• Dissolution Dynamic Nuclear Polarization (D-DNP) significantly enhances NMR signal strength for in vivo and in vitro studies.
• Time-series NMR data from D-DNP experiments often contain complex mixtures of signals.

Purpose of the Study:

• To demonstrate the application of Blind Source Separation (BSS) for analyzing time-series NMR data.
• To evaluate BSS as a method for component identification and kinetic analysis in D-DNP experiments.
• To assess the potential for automated analysis of complex NMR datasets.

Main Methods:

• Application of a Blind Source Separation (BSS) algorithm to time-series NMR spectral data.
• Utilizing BSS to decompose mixed NMR signals into individual source spectra without requiring reference spectra.
• Analyzing the time evolution of separated signal components.

Main Results:

• BSS successfully separated signal components in a time-resolved enzymatic conversion study (oxaloacetate to malate).
• Computed source spectra closely matched individual compound spectra, with improved signal-to-noise ratio.
• Reconstructed time evolutions accurately reflected integrated signal intensities, correlating with kinetic processes.

Conclusions:

• BSS is an effective, automated approach for analyzing complex NMR data from D-DNP experiments.
• The method facilitates efficient identification of chemical or metabolic components and estimation of reaction kinetics.
• BSS offers a powerful tool for accelerating the analysis of hyperpolarized NMR time-series data.