Prediction of chemical composition, volatile fatty acids and volatile compounds in Montasio cheese by Fourier transformed near-infra-red reflectance spectroscopy (FT-NIR).

Efficacy of Fourier transform (FT)-NIR reflectance spectroscopy in the determination of the chemical composition of Montasio cheese, at different periods of ripening, and in predicting the contents of volatile fatty acids (VFA) and other compounds was investigated. A total of 128 cheeses was collected and analysed for moisture, protein, fats and VFA. Additionally, volatile compounds were determined using a reverse carrier gas headspace sampling technique coupled to GC-MS. The NIR spectra of the cheese samples were acquired with an FT-NIR spectrometer. By relating spectrum properties to chemical data, calibration equations were developed and computed. A full cross validation was performed during the calibration process, which permitted a selection of calibration equations with the highest coeff. of determination (R 2) and the smallest standard errors of calibration (SEE) and cross validation (SEP). Results obtained for the chemical composition were very satisfactory, with R 2 values of 0.977, 0.953 and 0.966 for moisture, fat and protein, respectively. Encouraging calibrations were also computed for acetic acid and the sum of the VFA contents, with R 2 values >0.93. SEE and SEP values were consistently low for both the chemical composition and the volatile FA contents. However, calibration equations for volatile compounds were less satisfactory, with R 2 values