Equidistant regions of interest - multivariate curve resolution for clean spectra from plant-metabolomics by gas chromatography with high resolution mass spectrometry

Abstract

Background

The application of Gas Chromatography coupled with Orbitrap High-Resolution Mass Spectrometry (GC-Orbitrap HRMS) in plant metabolomics presents challenges including substantial data storage requirements, system complexity, and particularly the prevalence of co-eluted peaks in elution profiles. These limitations necessitate the integration of chemometric smart tools to extract interpretable mass spectra from complex chromatographic fingerprints. However, the direct applicability of multivariate resolution algorithms to GC-Orbitrap HRMS datasets remains limited. The existing deconvolution platforms exhibit operational constraints such as insensitivity toward embedded peaks and prerequisite data pretreatment.

Results

Using Cyperus rotundus and Curcumae Radix as model medicinal plants, this study developed an equidistant Region of Interest (eROI) strategy to enhance metabolite annotation reliability. The eROI method systematically converts raw datasets into structured matrices with standardized dimensions, effectively preserving m/z reading for subsequent analytical phases. We implemented scenario-specific combinations of eROI with three multivariate curve resolution (MCR) techniques to isolate pure component spectra from co-eluted chromatographic features. Comprehensive metabolite annotation was achieved through systematic spectral interpretation of fragmentation patterns, supplemented by predictive mass spectral analysis when database matching proved inconclusive. Detailed validation using volatile-metabolomics fingerprints from both botanical species accompanies each methodological stage.

Significance

Our work establishes a novel data reduction framework for GC-HRMS applications, enabling robust multi-modal chromatographic deconvolution. The integrated eROI-MCR methodology provides a validated solution for obtaining reliable qualitative and quantitative results in non-targeted plant-metabolomics studies.