LUMIN: A novel algorithm for automated mixture quantification using 1D 1H NMR spectra.

BACKGROUND: One-dimensional proton nuclear magnetic resonance (1D 1H NMR) spectroscopy is a non-destructive, non-targeted analytical technique providing both qualitative and quantitative insights, particularly beneficial for mixture analysis. Quantitative 1H Nuclear Magnetic Resonance (1H qNMR) spectroscopy is a powerful tool for analyzing mixtures, offering simple sample preparation and non-destructive measurement. Despite its advantages, a critical bottleneck exists: the manual selection of appropriate quantification peaks. This process, which requires identifying signals with high signal-to-noise ratios and no spectral overlap, is time-consuming, subjective, and heavily reliant on analyst expertise. This challenge is particularly acute for complex mixtures and high-throughput analyses, significantly impeding the efficiency and broader application of the qNMR technique.

RESULTS: We construct a library-utilized tool for automated qNMR analysis, named LUMIN. The tool automatically identifies optimal quantification peaks without requiring analysts' pre-defined choices in the reference library and provides two operating modes to balance efficiency with analytical depth. Performance was evaluated in terms of peak-selection reliability and concentration accuracy. In synthetic model mixtures, LUMIN correctly identified over 95% of quantification peaks, fully consistent with expert assignments and without false positives. Quantitative accuracy, benchmarked against manual integration, showed excellent linear correlation with relative errors consistently within 10%. Application to real-world mixtures of both plant and animal origin further demonstrated its robustness and practical applicability.

SIGNIFICANCE: LUMIN translates expert decision-making logic into a robust, automated workflow for 1D 1H qNMR. By eliminating tedious and subjective manual peak selection, it significantly reduces analysis time while enhancing objectivity and reproducibility. This user-friendly tool supports both high-throughput screening and detailed spectral analysis, lowering the barrier for the broader application of qNMR in large-scale studies of complex mixtures.