Establishment of a spectral database for classification of edible oils using matrix-assisted laser desorption/ionization mass spectrometry

Abstract

In this study, we aim to establish a comprehensive spectral database for analysis of edible oils using matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). More than 900 edible oil samples, including 30 types of edible oils, were analyzed and compared, and the characteristic peaks and spectral features of each edible oil were obtained. Edible oils were divided into eight groups based on their characteristic spectral patterns and principal component analysis results. An overall correct rate of 97.2% (98.1% for testing set) was obtained for classification of 435 edible oil products using partial least square-discriminant analysis, with nearly 100% correct rate for commonly used edible oils. Differentiation of counterfeit edible oils, repeatedly cooked edible oils and gutter oils from normal edible oils could also be achieved based on the MALDI-MS spectra. The establishment of this spectral database provides reference spectra for spectral comparison and allows rapid classification of edible oils by MALDI-MS.