Location: Quality & Safety Assessment ResearchTitle: Feasibility of using near-infrared (NIR) spectroscopy for quantitative detection of Kojic Acid in wheat flour Author
|Zhao, Xin - China Agricultural University|
|Wang, Wei - Tarim University|
|Li, Chunyang - Jiangsu Academy Agricultural Sciences|
|Wang, Wei - China Agricultural University|
|Ni, Xinzhi - US Department Of Agriculture (USDA)|
Submitted to: Journal of Food Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/25/2017
Publication Date: N/A
Interpretive Summary: The browning of wheat flour and its products, such as fresh noodles and dough, is a common quality defect during processing and storage. Browning not only affects the appearance of the products but also affects customers’ acceptability adversely. Meanwhile, browning leads to the losses of protein nutrition and lowers the value of the products. Kojic acid is an antioxidant and a fungal metabolite produced mainly by species of Aspergillus, Penicillium and Acetobactor. It is widely used as a chemical additive to suppress browning during food storage and processing, including wheat flour. However, its safety could be a health concern in application. Research has shown that kojic acid causes mutations in bacteria and also is toxic to animals. Therefore, its content in food is not desirable. The objective of the present study was to investigate the feasibility of detecting added kojic acid in wheat flour using NIR full spectra. Our results show that a partial least square model based on NIR spectra can be used to predict kojic acid content in wheat flour rapidly and non-destructively.
Technical Abstract: The possibility of using NIR spectroscopy technology to detect kojic acid (KA) added in wheat flour was studied. Three common types of white flour samples, i.e. high-gluten flour, plain flour and low-gluten flour were added with different contents of KA (0.0%, 0.5%, 1.0%, 3.0%, 5.0%, and 10.0%) respectively as the tested samples. Spectra of all samples were collected with a range of 1000-2400 nm. For high-gluten flour samples, three common preprocessing algorithms were compared with each other, as well as non-preprocessing. SGD was found to be the best performance and used as the preprocessing method. Then interval partial least squares (iPLS) was adopted to optimize spectral interval. The optimal spectral interval was determined from 1088.8 to 1153.5 nm. PLS model based on the optimal spectral interval showed better performance than model based on full wavelength.