LAMAIS: A library-aided approach for efficient 1D 1H NMR qualitative analysis in plant metabolomics

Xinyuan Xie,Jiayu Yang,Yuting Lu,Yingting Shi,Jianyang Pan,Haibin Qu
DOI: https://doi.org/10.1016/j.aca.2024.343100
2024-10-16
Abstract: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. However, the qualitative analysis of 1D 1H NMR spectra for mixture samples is laborious and time-consuming, involving extensive database searches and verification experiments like spiking. This process heavily relies on the analyst's expertise, leading to efficiency discrepancies. There is a pressing need for a reliable method to streamline operations and enhance the efficiency of qualitative analysis in complex mixtures. Results: We introduce a library-aided method for spectral profiling, named LAMAIS. This method achieves compound identification through similarity assessment between samples and template data, allowing rapid, automatic compound identification and full-spectrum peak assignment without the need for fitting. LAMAIS correctly identifies over 90 % of components in synthetic mixtures and more than 75 % in experimental mixtures, surpassing other representative methods with a higher F2 score. Our reference library, which currently includes 71 compounds, is tailored to capture the commonality of primary metabolites across diverse plant species. The analysis of real-world samples yielded encouraging results, underscoring LAMAIS's versatility as an auxiliary tool suitable for a variety of botanical sources. For analyst convenience, interactive graphics are utilized as the output format. Significance: LAMAIS excels, demonstrating competitiveness and reliability. The approach minimizes repetitive tasks and sample wastage, improving the efficiency of 1D 1H NMR qualitative analysis. Constructing a reference library effectively preserves knowledge, mitigates reliance on human experience, and addresses gaps in the analysis of plant source samples.
What problem does this paper attempt to address?