1H NMR and multi-technique data fusion as metabolomic tool for the classification of golden rums by multivariate statistical analysis

José Raúl Belmonte-Sánchez; Roberto Romero-González; José Luis Martínez Vidal; Francisco Javier Arrebola; Antonia Garrido Frenich

doi:10.1016/j.foodchem.2020.126363

¹H NMR and multi-technique data fusion as metabolomic tool for the classification of golden rums by multivariate statistical analysis

Food Chem. 2020 Jul 1:317:126363. doi: 10.1016/j.foodchem.2020.126363. Epub 2020 Feb 5.

Authors

José Raúl Belmonte-Sánchez¹, Roberto Romero-González¹, José Luis Martínez Vidal¹, Francisco Javier Arrebola¹, Antonia Garrido Frenich²

Affiliations

¹ Department of Chemistry and Physics, Research Centre CIAIMBITAL, Agrifood Campus of International Excellence, ceiA3, University of Almería, Ctra. Sacramento, s/n, 04120, Almería, Spain.
² Department of Chemistry and Physics, Research Centre CIAIMBITAL, Agrifood Campus of International Excellence, ceiA3, University of Almería, Ctra. Sacramento, s/n, 04120, Almería, Spain. Electronic address: agarrido@ual.es.

PMID: 32086119
DOI: 10.1016/j.foodchem.2020.126363

Abstract

¹H NMR spectroscopy combined with chemometrics was applied for the first time for golden rum classification based on several factors as fermentation barrel, raw material, distillation method and aging. Principal component analysis (PCA) was used to assess the overall structure, and partial least square discriminant analysis (PLS-DA) was carried out for the analytical discrimination of rums. Additionally, data-fusion of ¹H NMR and chromatographic techniques (gas and liquid chromatography) coupled to mass spectrometry was applied to provide more accurate knowledge about rums. This approach provided a classification of samples with lower error rate than the one obtained by the use of a single technique (spectroscopic or chromatographic). The results showed that ¹H NMR spectroscopy is an appropriate technique for the suitable classification of >95.5% of the samples. When data fusion methodology of spectroscopic and spectrometric data was performed, the prediction efficiency can reach 100% of the samples.

Keywords: Data-fusion; Metabolomics; Multivariate analysis; NMR spectroscopy; Rum.

MeSH terms

Alcoholic Beverages / analysis*
Chromatography, High Pressure Liquid
Discriminant Analysis
Gas Chromatography-Mass Spectrometry
Least-Squares Analysis
Mass Spectrometry
Metabolomics / methods*
Multivariate Analysis
Principal Component Analysis
Proton Magnetic Resonance Spectroscopy*