Zaqlul Iqbal, Annelies Postelmans, Joni Kusnadi, Wouter Saeys
Minced beef is widely used as a primary ingredient in various processed meat products and is vulnerable to adulteration with lower-priced meats such as pork. While near-infrared (NIR) spectroscopy has been investigated for detecting such adulteration, the role of fat content as key feature for adulteration detection remains underexplored. To address this gap, this study investigated the potential of near infrared (NIR) spectroscopy to detect pork adulteration in minced beef across varying levels of fat content and adulteration, and identified the most informative NIR wavebands for detection. As water is a strong absorber in the NIR range, samples were dried and powdered prior to spectral scanning to minimize its masking effect. Synergy interval partial least squares discriminant analysis (SiPLS-DA) was applied to identify influential spectral regions for distinguishing pork and beef-pork mixes. Overall, detection models performed better in classifying samples of fat-rich mixtures than lean mixtures which indicate that fat plays an important role in NIR detection of pork adulteration in minced beef. This finding suggests that species-specific differences in fat signature contribute to distinctive spectral characteristics, which support the models to classify samples to their respective classes. Across 120 combinations of double cross-validation, models including the wavelength range from 1521 to 1719 nm achieved test set accuracies of 86.3 ± 6.3% with a high sensitivity for detecting pork-adulterated samples (up to 93.0 ± 8.0%), but lower specificity (72.1 ± 18.0%). These findings demonstrate that NIR spectroscopy has potential for rapid detection of pork adulterated minced beef after appropriate sample preparation. However, the models showed limited performance in classifying lean beef and samples containing 20% pork adulteration, which exhibited the highest misclassification rate of 46.2% and 20.97%, respectively.