Discrimination of milk species using Raman spectroscopy coupled with partial least squares discriminant analysis in raw and pasteurized milk

Abstract

BACKGROUND Heat treatment is the most common practice for the microbiological safety of milk; hence determination of the heat‐treatment of milk is essential. Also, mislabeling or adulteration of expensive milk samples like ewe or goat milk with cow's milk is a growing problem in the dairy market. Thus, the determination of the authenticity of milk samples has crucial importance for both producers and consumers. The aim of this study was to discriminate milk samples firstly as heat‐treated or not, and secondly according to their species (cow, goat, ewe, mixture (adulterated)) in both raw and pasteurized milk by using Raman spectroscopy with PLS‐DA. RESULTS In this study, discrimination of milk samples as raw or pasteurized was firstly achieved using partial least square‐discriminant analysis (PLS‐DA). Both in calibration and prediction models, high sensitivity and specificity values were obtained for raw and pasteurized milk samples. The proposed method also discriminated the milk samples according to their species (cow, goat, ewe, and mixture) for both raw and pasteurized milk. In both calibration and prediction models, the sensitivity and specificity values were found above 0.857, and 0.897, respectively. Also, the accuracy values were found above 0.915. The obtained results denote satisfactory accurate classification of the samples. CONCLUSION The results suggested that Raman spectroscopy coupled with PLS‐DA was successful in discriminating milk samples according to heat treatment (raw/pasteurized) and their species within 20 s per sample. It was seen that Raman spectra provide valuable information to be used especially for discrimination of milk samples according to their origin.