Detection of deoxynivalenol in wheat flour using fluorescence fingerprint

AbstractThis study focused on using fluorescence fingerprint (FF) to detect DON non-destructively. The fluorescence fingerprint is also called excitation-emission matrix, which consists of an excitation wavelength axis, an emission wavelength axis, and a fluorescence intensity axis. It is acquired by measuring the fluorescence intensity of a sample at consecutive excitation and emission wavelengths. The samples were wheat grains artificially contaminated with 4 levels of DON. They were crushed into flour. FF measurement was made by a fluorescence spectrometer using a powder cell. Both excitation wavelength and emission wavelength ranges were from 200 to 900nm at each 10nm interval. Three analyses for FF data were carried out to confirm the possibility of DON detection. Principal component analysis (PCA) was applied to the FF data. The result of PCA indicated that PC2 could become an index of DON contamination degree. Discriminant analysis was also carried out to the same data. Canonical 1 can discriminate the samples based on the DON contamination. Partial Least Squares Regression was applied to the quantitative sample data. Half of data set was used for the calibration and the rest was used for validation. R2 for calibration and validation set were 0.990 and 0.983, respectively. As a result, fluorescence fingerprint has possibility to become practical technique for screening of DON contamination