Abstract:
Adulteration detection and origin identification of yak milk powder were essential to ensure food safety and safeguard consumer rights interests, thereby promoting the healthy development of the dairy product market. Traditional DNA detection methods and isotope analysis showed long detection time, which were inapplicable to rapid, low-cost on-site analysis. To address these issues, a rapid adulteration detection and identification of the origin of yak milk powder based on near-infrared Spectroscopy (NIRS) technology was established in this study. Yak milk powder samples from nine brands from Sichuan, Gansu, Yunnan, and Qinghai were collected. Before preparing adulterated samples, polymerase chain reaction (PCR) technology and DNA gel electrophoresis were used to verify whether the collected yak milk powder were adulterated with cow milk powder. Then adulterated samples were prepared and NIRS data were collected. The K-nearest neighbors (KNN) method was employed to establish a classification model. Partial least squares regression (PLSR) was used to establish a quantitative prediction model. The predictive ability of quantitative prediction model was improved by optimizing spectral preprocessing methods and variable selection methods. Results showed that KNN achieved 100% correct classification for adulteration detection (pure cow milk powder, pure yak milk powder, yak milk powder adulterated with cow milk powder) and origin identification (Sichuan, Gansu, Yunnan, Qinghai). The calibration set correlation coefficient (
Rc), the prediction set correlation coefficient (
Rp), the root mean square error of prediction (RMSEP), and the ratio of performance to deviation (RPD) of the adulteration quantitative prediction model were 0.9975, 0.9913, 1.9823%, and 7.2522, respectively. This method enables rapid and accurate prediction of cow milk powder adulteration in yak milk powder and the identification of the origin of yak milk powder, providing technical support for the quality control of yak milk powder.