Vibrational spectroscopy as a green technology for predicting nutraceutical properties and antiradical potential of early-to-late apricot genotypes.

• Antioxidant compounds of early-to-late apricots were predicted using ATR-FTIR and PLS. • Very good multi-cultivar models were developed for total phenols and flavans. • MIR predicted the parameters better than NIR. Vibrational spectroscopy is considered a green alternative to the traditional analytical methods for quantitative determination of antioxidants in agricultural products. In this study, the potential of NIR and MIR spectroscopy for the prediction of antioxidant compounds in seven early-to-late apricot genotypes from different locations was reported. NIR and MIR calibration models were developed using Partial Least Squares Regression (PLS) with cross-validation, and the performance of final models was evaluated using Rp2, root mean square error of prediction (RMSEP) and relative predicted determinant (RPD). Five independent PLS models were developed, using both NIR and MIR spectra from 300 samples per year and their corresponding total phenols (TPs), total flavans (FLs), total carotenoids (TCs), antiradical capacity (AC) reference values, yielding coefficient on average always > 0.95. Very good models were obtained for TPs (Rp2 = 0.99, RMSEP = 0.73 g kg−1 GAE, RPD = 12.0; Rp2 = 0.99, RMSEP = 0.95 g kg−1 GAE, RPD = 9.9 for MIR and NIR models, respectively), and FLs (Rp2 = 0.99, RMSEP = 0.99 g kg−1 CAT, RPD = 8.5; Rp2 = 0.98, RMSEP = 1.10 g kg−1 CAT, RPD = 6.4, for MIR and NIR models, respectively). Besides, both MIR and NIR models for total carotenoids (TCs) quantification showed a satisfactorily predictive capability (Rp2 = 0.98, RMSEP = 0.01 g kg-1 β-carotene, RPD = 6.5; Rp2 = 0.95, RMSEP = 0.01 g kg-1 β-carotene, RPD = 4.5, for MIR and NIR models, respectively). Good model performances were also obtained for the prediction of antiradical potential of samples, developed herein for the first time on freeze-dried apricot samples. In general, ATR-MIR demonstrated advantages over NIR in predicting all nutraceutical parameters, whereas similar model performances were obtained for the antiradical potential tested versus ABTS+• using MIR and NIR spectral regions. Results reported in the present study suggest that vibrational spectroscopy, both in MIR and NIR range, is adequate to predict the nutraceutical properties of apricots for screening purposes quickly and sustainably. [ABSTRACT FROM AUTHOR]