Dataset Information

Comprehensive Classification and Regression Modeling of Wine Samples Using ¹H NMR Spectra.

ABSTRACT: Recently, ¹H NMR (nuclear magnetic resonance) spectroscopy was presented as a viable option for the quality assurance of foods and beverages, such as wine products. Here, a complex chemometric analysis of red and white wine samples was carried out based on their ¹H NMR spectra. Extreme gradient boosting (XGBoost) machine learning algorithm was applied for the wine variety classification with an iterative double cross-validation loop, developed during the present work. In the case of red wines, Cabernet Franc, Merlot and Blue Frankish samples were successfully classified. Three very common white wine varieties were selected and classified: Chardonnay, Sauvignon Blanc and Riesling. The models were robust and were validated against overfitting with iterative randomization tests. Moreover, four novel partial least-squares (PLS) regression models were constructed to predict the major quantitative parameters of the wines: density, total alcohol, total sugar and total SO₂ concentrations. All the models performed successfully, with R² values above 0.80 in almost every case, providing additional information about the wine samples for the quality control of the products. ¹H NMR spectra combined with chemometric modeling can be a good and reliable candidate for the replacement of the time-consuming traditional standards, not just in wine analysis, but also in other aspects of food science.

SUBMITTER: Baratossy G

PROVIDER: S-EPMC7824661 | biostudies-literature | 2020 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Comprehensive Classification and Regression Modeling of Wine Samples Using 1H NMR Spectra.

Barátossy Gábor G Berinkeiné Donkó Mária M Csikorné Vásárhelyi Helga H Héberger Károly K Rácz Anita A

Foods (Basel, Switzerland) 20201230 1

Recently, 1H NMR (nuclear magnetic resonance) spectroscopy was presented as a viable option for the quality assurance of foods and beverages, such as wine products. Here, a complex chemometric analysis of red and white wine samples was carried out based on their 1H NMR spectra. Extreme gradient boosting (XGBoost) machine learning algorithm was applied for the wine variety classification with an iterative double cross-validation loop, developed during the present work. In th ...[more]

PMID: 33396655

Similar Datasets

Project description:INTRODUCTION:Proton nuclear magnetic resonance spectroscopy (1H-NMR)-based metabolomic profiling has a range of applications in plant sciences. OBJECTIVES:The aim of the present work is to provide advice for minimizing uncontrolled variability in plant sample preparation before and during NMR metabolomic profiling, taking into account sample composition, including its specificity in terms of pH and paramagnetic ion concentrations, and NMR spectrometer performances. METHODS:An automation of spectrometer preparation routine standardization before NMR acquisition campaign was implemented and tested on three plant sample sets (extracts of durum wheat spikelet, Arabidopsis leaf and root, and flax leaf, root and stem). We performed 1H-NMR spectroscopy in three different sites on the wheat sample set utilizing instruments from two manufacturers with different probes and magnetic field strengths. The three collections of spectra were processed separately with the NMRProcFlow web tool using intelligent bucketing, and the resulting buckets were subjected to multivariate analysis. RESULTS:Comparability of large- (Arabidopsis) and medium-size (flax) datasets measured at 600 MHz and from the wheat sample set recorded at the three sites (400, 500 and 600 MHz) was exceptionally good in terms of spectral quality. The coefficient of variation of the full width at half maximum (FWHM) and the signal-to-noise ratio (S/N) of two selected peaks was comprised between 5 and 10% depending on the size of sample set and the spectrometer field. EDTA addition improved citrate and malate resonance patterns for wheat sample sets. A collection of 22 samples of wheat spikelet extracts was used as a proof of concept and showed that the data collected at the three sites on instruments of different field strengths and manufacturers yielded the same discrimination pattern of the biological groups. CONCLUSION:Standardization or automation of several steps from extract preparation to data reduction improves data quality for small to large collections of plant samples of different origins.

Dataset Information

Comprehensive Classification and Regression Modeling of Wine Samples Using ¹H NMR Spectra.

Publications

Comprehensive Classification and Regression Modeling of Wine Samples Using <sup>1</sup>H NMR Spectra.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets