Evaluating low-intensity unknown signals in quantitative proton NMR mixture analysis

Area of Science:

• Analytical Chemistry
• Metabolomics
• Biochemistry

Background:

• Complex mixtures like biofluids and foods often contain unknown compounds at low concentrations.
• Nuclear Magnetic Resonance (NMR) spectral analysis is crucial for identifying these compounds.
• Conventional methods like spectral binning and high-resolution analysis have limitations in detecting low-intensity signals.

Purpose of the Study:

• To compare conventional NMR spectral analysis methods with a novel library-based method, targeted profiling of unknowns (TPU).
• To evaluate the effectiveness of these methods in identifying low-intensity unknown peaks (LIUPs) in complex biological samples.
• To assess the significance of LIUPs in multivariate statistical analysis for distinguishing between different animal groups.

Main Methods:

• Proton NMR spectral data from ultrafiltered mouse serum were analyzed.
• Conventional methods (spectral binning, high-resolution analysis) and TPU were applied.
• LIUPs were assessed for their significance in multivariate statistical analysis.
• NMR signal linearity was determined by titrating metabolites into serum.
• Carbon-13 decoupling was used to eliminate isotope-satellite peaks.

Main Results:

• The TPU method successfully identified 25 LIUPs.
• Four specific LIUPs were significant in separating arthritic from diseased animals.
• Conventional methods struggled with baseline noise and overlapping signals, failing to identify LIUPs reliably.
• NMR signal linearity was confirmed for low incremental concentration changes (< 10 microM).

Conclusions:

• The TPU method is recommended for analyzing peaks with low signal-to-noise ratios or when high-fidelity spectral data compression is needed.
• Conventional methods, especially high-resolution analysis, are suitable for peaks with moderate signal-to-noise.
• TPU enhances the ability to integrate NMR data into cross-platform studies and improves metabolomics research.