Geographical discrimination of Italian carrot (Daucus carota L.) varieties: A comparison between ATR FT-IR fingerprinting and HS-SPME/GC-MS volatile profiling

The Italian carrot varieties cultivated in the Fucino upland (Abruzzo), and Ispica (Sicily), both Protected Geographical Indication specialties, together with an ordinary product grown in the Apulia region, were char-acterized by attenuated total reflectance Fourier-transform infrared spectroscopy (ATR FT-IR) and headspace solid-phase microextraction followed by gas chromatography with mass spectrometry detection (HS-SPME/GC-MS). One hundred eighty (180) infrared spectra, collected from either the internal or external side of the carrot root, and 80 volatile profiles, acquired from representative samples of the three varieties, were individually handled by partial least square discriminant analysis (PLS-DA) to trace the carrots according to their geographical origin. The PLS-DA model calibrated on 120 infrared spectra acquired on the internal side of the carrot root gave only one misclassification error in the prediction set (consisting of the left 60 spectra); whereas all validation samples were correctly classified when the model was calculated on the signals collected on the external side of the products. Further PLS-DA classification was based on 32 volatile components detected by HS-SPME/GC-MS. This model, built on 52 calibration samples and successively applied to 28 external ones, led to a global prediction accuracy of 82.5%, due to the misclassification of only one sample from Ispica and four Apulian individuals. Variance Importance in Projection (VIP) analysis was eventually conducted to identify the volatile components and vibrational modes contributing the most to the discrimination of the three carrot varieties by the respective PLS-DA models.