A volatilomics approach for off-line discrimination of minced beef and pork meat and their admixture using HS-SPME GC/MS in tandem with multivariate data analysis.

Beef, pork and mixed (70% beef and 30% pork) minced meat samples were obtained from a meat processing plant in Athens during a two-year survey and analyzed both microbiologically and by headspace solid-phase microextraction in combination with gas chromatography-mass spectrometry (HS-SPME/GC-MS). A validated method for the discrimination of minced meat was developed based on the volatile fingerprints. Unsupervised (PCA) and supervised (PLS-DA) multivariate statistical methods were applied to visualize, group and classify the samples. The data-set was divided 70% for model calibration and 30% for model prediction. During model calibration 99, 100 and 100% of the samples were correctly classified as beef, pork and mixed meat samples, respectively, while for model prediction the respective percentages were 100, 100 and 95% respectively. In both datasets, the overall correct classification rate amounted to 99% on average. Among the volatile compounds identified, heptanal, octanal, butanol, pentanol, hexanol, octanol, 1-penten-3-ol, 2-octen-1-ol, 3-hydroxy-2-butanone, 2-butanone and 2-heptanone were positively correlated with beef samples. Furthermore, pentanal, hexanal, decanal, nonanal, benzaldehyde, trans-2-hexenal, trans-2-heptenal, trans-2-octenal and 1-octen-3-one were positively correlated with pork. Lastly, the alcohols, 2-butanol and 1-octen-3-ol showed positive correlation with mixed samples. The results indicated that the volatilomics approach employed in this study could be used as an alternative method for robust and reliable discrimination and classification of meat samples in an off-line mode.