A comparison of electronic nose and gas chromatography–mass spectrometry on discrimination and prediction of ochratoxin A content in Aspergillus carbonarius cultured grape-based medium

This study investigated discrimination and prediction of ochratoxin A (OTA) in three Aspergillus carbonarius strains cultured grape-based medium using E-nose technology and GC-MS analysis. Results showed that these strains cultured medium samples were divided into four groups regarding their log 10 OTA value using an equispaced normal distribution analysis. Partial least squares-discriminant analysis (PLS-DA) revealed that GC-MS PLS-DA model only separated the low OTA level medium samples from the rest OTA level samples, whereas all the OTA level samples were segregated from each other using E-nose PLS-DA model. Partial least squares regression (PLSR) analysis indicated that an excellent prediction performance was established on the accumulation of OTA in these medium samples using E-nose PLSR, whereas GC-MS PLSR model showed a screening performance on the OTA formation. These indicated that E-nose analysis could be a reliable method on discriminating and predicting OTA in A. carbonarius strains under grape-based medium.