Abstract
We used fluorescence fingerprint (FF), also known as excitation-emission matrix spectroscopy, to develop a novel method to determine the mycotoxins nivalenol (NIV) and zearalenone (ZON) in wheat grain infected with Fusarium head blight. Grain samples were divided into four categories according to their level of blight damage. The amounts of NIV and ZON in milled samples were determined by liquid chromatography. NIV was not detected in all samples. ZON ranged from not detected to 0.71 ppm; its concentration was not correlated with damage, but it was predicted by FF values at various wavelengths by a partial least-squares regression model. The model showed a good fit with the calibration dataset (R2 = 0.993, SEP = 0.02 ppm) and the validation dataset (R2 = 0.974, SEP = 0.04 ppm). Plots of regression coefficients indicate that the prediction was based on peaks unique to each mycotoxin and some other weak fluorescence signals. These results confirm that FF can take rapid, non-destructive measurements and predict mycotoxins in wheat flour.
