The animal species of raw meat and processed meat products was determined by 1 H NMR spectroscopy with subsequent multivariate data analysis. Sample preparation was based on comprehensive lipid extraction to capture nonpolar and polar (amphiphilic) fat components of meat. A nontargeted approach was used to analyze the 1 H NMR data, followed by a principal component analysis, linear discrimination analysis, and cross-validation embedded in a Monte Carlo re-sampling approach. A total of 437 raw meat samples (pork, beef, poultry, and lamb) and 81 processed meat samples (pork, beef, and poultry) were collected to build and/or test the classification model. On average, 98% of the analyzed raw meat samples and 97% of the processed meat products were correctly classified with respect to meat species. Furthermore, relevant spectral regions to identify potential chemical markers such as linoleic acids, trans-fatty acids, and cholesterol for the meat species classification were described.
Meat species of raw meat and processed meat products were investigated by 1H NMR spectroscopy with subsequent multivariate data analysis. Sample preparation was based on aqueous extraction combined with ultrafiltration in order to reduce macromolecular components in the extracts. 1H NMR data was analyzed by using a non—targeted approach followed by principal component analysis (PCA), linear discrimination analysis (LDA), and cross-validation (CV) embedded in a Monte Carlo (MC) resampling approach. A total of 379 raw meat samples (pork, beef, poultry, and lamb) and 81 processed meat samples (pork, beef, poultry) were collected between the years 2018 and 2021. A 99% correct prediction rate was achieved if the raw meat samples were classified according to meat species. Predicting processed meat products was slightly less successful (93 %) with this approach. Furthermore, identification of spectral regions that are relevant for the classification via polar chemical markers was performed. Finally, data on polar metabolites were fused with previously published 1H NMR data on non-polar metabolites in order to build a broader classification model and to improve prediction accuracy.
Abstract1H NMR spectroscopy was applied to analyse samples of “Swabian–Hall Quality Pork” with protected geographical indication (PGI). To obtain maximum chemical information sample preparation was based on both polar extraction and non-polar extraction. A non-targeted approach was used to analyse the 1H NMR data followed by principal component analysis (PCA), linear discriminant analysis (LDA), and cross-validation (CV) embedded in a Monte Carlo (MC) resampling approach. A total of 275 raw pork samples were collected in the years 2018 to 2021. The correct prediction rate of “Swabian–Hall Quality Pork” was about 92% on average for both models based on either the polar or non-polar metabolites. In addition, 1H NMR data describing the polar and non-polar metabolites were combined in a classification model to improve the prediction accuracy. By performing a mid-level data fusion, a correct prediction rate of 98% was achieved. Furthermore, spectral regions in the NMR spectra of the polar and non-polar metabolites that are relevant for the classification of the pork samples were identified to describe potential chemical marker compounds.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.