Research Information System
Analytica Chimica Acta, vol.1363, 2025 (SCI-Expanded)
Background: The extraction of relevant information from proton nuclear magnetic resonance (1H NMR) spectra through preprocessing and multivariate analysis requires integrating multiple software tools and extensive manual intervention, compromising efficiency and reproducibility when the technique is used. Consequently, the development of automated, versatile, and reliable methodologies has become imperative to streamline workflows, improve analytical performance, and broaden the applicability of multivariate methods for the analysis of diverse sample types and experimental conditions. Results: This work presents the development and application of Interval Resonance Analysis (InRA), an alternative software tool focused on 1H NMR multivariate analysis. InRA includes a novel algorithm for resonance signal detection (intervals), specifically designed to operate with flexibility across diverse 1H NMR spectra. All intervals are integrated using multivariate curve resolution with alternating least squares (MCR–ALS) and analyzed by exploratory analysis. The performance of InRA was tested by evaluating the 1H NMR spectra of hydrophilic sugar beet root extracts cultivated in three different fields and their discrimination by partial least squares – discriminant analysis (PLS–DA). The workflow provided by InRA yielded consistent results regarding the distribution of samples according to their field, enabling the identification of subtle sources of variation and achieving classification accuracies ≥ 88.9 %. Significance: The proposed methodology represents an advancement in the multivariate analysis of 1H NMR spectra for untargeted studies and enhances analytical efficiency by reducing manual intervention and reliance on analyst experience. InRA is versatile and can be applied to various sample types and analytical objectives, as it is not restricted by specific experimental conditions.