Rapid discrimination of beef, pork and rabbit meat using HS-SPME coupled with GC-MS and chemometric analysis

Abstract

Species adulteration in meat products remains a persistent challenge for food safety, regulatory compliance, and consumer confidence, especially in products labeled as rabbit, pork, or beef. This study developed and validated a rapid, non-targeted volatilomic strategy based on headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS-SPME–GC–MS) for the authentication of minced samples from these three species. A total of 28 volatile compounds (VOCs), predominantly aldehydes and fatty acids associated with lipid oxidation pathways, exhibited statistically significant interspecies differences and distinct species-specific distribution patterns. Multivariate data exploration using t-Distributed Stochastic Neighbor Embedding (tSNE) demonstrated complete cluster separation among rabbit, pork, and beef samples. Supervised machinelearning algorithms, including Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Artificial Neural Network (ANN), achieved high classification performance, with overall accuracies ranging from 90.3 to 94.4%; the ANN model yielded the best predictive performance. Rabbit meat was distinguished with near-perfect precision, primarily attributable to distinctive fatty acid-derived volatile markers. The proposed method requires less than 60 min per sample, involves minimal sample preparation, and relies on widely available analytical instrumentation, supporting its applicability as a practical, cost-effective tool for routine meat authenticity assessment in both raw and processed products.