1. Chemical characterization and classification of vegetable oils using DESI-MS coupled with a neural network.
- Author
-
Cui Y, Zhu L, Li Y, Ge K, Lu W, Ge L, Chen K, Xue J, Zheng F, Dai S, Pan H, Liang J, Ji L, and Shen Q
- Abstract
This study tackled mislabeling fraud in vegetable oils, driven by price disparities and profit motives, by developing an approach combining desorption electrospray ionization mass spectrometry (DESI-MS) with a shallow convolutional neural network (SCNN). The method was designed to characterize lipids and distinguish between nine vegetable oils: corn, soybean, peanut, sesame, rice bran, sunflower, camellia, olive, and walnut oils. The optimized DESI-MS method enhanced the ionization of non-polar glycerides and detected ion adducts like [TG + Na]
+ , [TG + NH4 ]+ . This process identified 53 lipid peaks, forming a robust lipid fingerprint for each oil type. An SCNN model was developed using fingerprints, achieving an impressive classification accuracy of 98.5 ± 2.2 %. The integration of DESI-MS with SCNN provides a fast and reliable tool for identifying and classifying vegetable oils, thereby reducing mislabeling fraud and assuring oil quality. By enabling accurate authentication, it contributes to improved transparency and integrity in food labeling and quality control practices., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)- Published
- 2024
- Full Text
- View/download PDF