Detection of adulteration of goat and sheep meat through NIRS and chemometric analysis

This research aimed to assess the capability of near-infrared (NIR) reflectance spectroscopy coupled with chemometric analysis in detecting adulteration in Goat and Sheep meat. A total of 16 samples were prepared, consisting of 2 pure samples and 14 adulterated samples. Spectra were collected using DLP® NIRscan™ Nano Software to detect adulteration. Partial least square and principal component regression models were developed for calibration and validation using The Unscrambler X software. The accuracy of the calibration models was assessed using root mean square error of calibration (RMSEc), root mean square error of cross-validation (RMSEcv), coefficient of calibration (R2c), and coefficient of cross-validation (R2cv). Typically, a regression model is considered excellent when R2 ≥ 0.90. For the PCR model, the predicted R2cv was 0.62, and for PLSR, it was 0.99 after leverage correction. Conversely, through cross-validation, the R2cv for the PCR model was 0.29, and for PLSR, it was 0.18. The results suggest that NIR spectroscopy coupled with chemometric analysis was reasonably efficient in detecting adulteration in goat meat with sheep meat.