Metabolic fingerprinting of hard and semi-hard natural cheeses using gas chromatography with flame ionization detector for practical sensory prediction modeling

Abstract

Metabolic fingerprinting using gas chromatography with flame ionization detector (GC/FID) was used to generate a practical metabolomics-based tool for quality evaluation of natural cheese. Hydrophilic low molecular weight components, relating to sensory characteristics, including amino acids, fatty acids, amines, organic acids, and saccharides, were extracted and derivatized prior to the analysis. Data on 12 cheeses, six Cheddar cheeses and six Gouda cheeses, were analyzed by multivariate analysis. Prediction models for two sensory attributes relating to maturation, "Rich flavor" and "Sour flavor", were constructed with 4199 data points from GC/FID, and excellent predictability was validated. Chromatograms from GC/FID and gas chromatography/time-of-flight-mass spectrometry (GC/TOF-MS) were comparable when the same column was used. Although GC/FID alone cannot identify peaks, the mutually complementary relationship between GC/FID and GC/MS does allow peak identification. Compounds contributing significantly to the sensory predictive models included lactose, succinic acid, L-lactic acid, and aspartic acid for "Rich flavor", and lactose, L-lactic acid, and succinic acid for "Sour flavor". Since similar model precision was obtained using GC/FID and GC/TOF-MS, metabolic fingerprinting using GC/FID, which is a relatively inexpensive instrument compared with GC/MS, is easy to maintain and operate, and is a valid alternative when metabolomics (especially using GC/MS) is to be used in a practical setting as a novel quality evaluation tool for manufacturing processes or final products.