Rapid screening of ochratoxin A in wheat by infrared spectroscopy

The use of infrared spectroscopy for the screening of 229 unprocessed durum wheat samples naturally contaminated with OTA has been investigated. Samples were analysed by both Fourier Transform near-and midinfrared spectroscopy (FT-NIR, FT-MIR). Partial-Least Squares-Discriminant Analysis (PLS-DA) and Principal Component-Linear Discriminant Analysis (PC-LDA) classification models were used to differentiate highly contaminated durum wheat samples from low contaminated ones and the performances of the resulting models were compared. The overall discrimination rates were higher than 94% for both FT-NIR and FT-MIR range by using a cut-off limit set at 2 mu g/kg OTA, independently from the classification model used thus confirming the reliability of the two statistical approaches used. False compliant rates of 6% were obtained for both spectral ranges and both classification models. These findings indicate that FT-NIR, as well as FT-MIR analysis, might be a promising, inexpensive and easyto- use screening tool to rapidly discriminate unprocessed wheat samples for OTA content.