RESUMEN
The purpose of this study was to evaluate the potential of FT-IR spectroscopy as a high-throughput method for rapid differentiation among the ochratoxigenic species of Aspergillus carbonarius and the non-ochratoxigenic or low toxigenic species of Aspergillus niger aggregate, namely A. tubingensis and A. niger isolated previously from grapes of Greek vineyards. A total of 182 isolates of A. carbonarius, A. tubingensis, and A. niger were analyzed using FT-IR spectroscopy. The first derivative of specific spectral regions (3002-2801cm-1, 1773-1550cm-1, and 1286-952cm-1) were chosen and evaluated with respect to absorbance values. The average spectra of 130 fungal isolates were used for model calibration based on Discriminant analysis and the remaining 52 spectra were used for external model validation. This methodology was able to differentiate correctly 98.8% in total accuracy in both model calibration and validation. The per class accuracy for A. carbonarius was 95.3% and 100% for model calibration and validation, respectively, whereas for A. niger aggregate the per class accuracy amounted to 100% in both cases. The obtained results indicated that FT-IR could become a promising, fast, reliable and low-cost tool for the discrimination and differentiation of closely related fungal species.