Use of HS-SPME-GC-MS for the classification of Italian lemon, orange and citrus spp. honeys

• Published in: International journal of food science & technology Vol. 47; no. 11; pp. 2352 - 2358
• Main Authors: Papotti, Giulia, Bertelli, Davide, Plessi, Maria
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.11.2012; Wiley-Blackwell; Wiley Subscription Services, Inc
• Subjects: Biological and medical sciences; Chemometrics; Citrus fruits; Citrus limon; Citrus sinensis; Classification; Confectionery products and chocolate industries, honey; Discriminant analysis; Food industries; food quality; Food science; Foods; Fruit and vegetable industries; Fundamental and applied biological sciences. Psychology; Honey; Lemons; Mathematical models; quality assurance; solid-phase microextraction (SPME); Statistical analysis; Variance; Fruit; Citrus fruit; Use; food quality; Rutaceae; Lemon; Orange; Citrus; Solid phase microextraction; Honey; Quality assurance; Dicotyledones; Quality; Angiospermae; Classification; solid-phase microextraction (SPME); Spermatophyta; Chemometrics; Food
• ISSN: 0950-5423; 1365-2621
• DOI: 10.1111/j.1365-2621.2012.03109.x

Summary Volatile compounds from 90 Italian citrus honeys of different varieties (lemon, Citrus limon (L.) Burm., orange, Citrus sinensis (L.) Osbeck, and citrus, Citrus spp.), 30 provided by CRA‐API (Consiglio per la Ricerca e la Sperimentazione in Agricoltura‐Unità di ricerca di apicoltura e bachicoltura) (Bologna, Italy) and 60 purchased on the markets, were analysed to classify them according to the characteristic aroma profiles. A HS‐SPME‐GC method previously developed and validated was applied. Factor analysis and general discriminant analysis (GDA) were carried out on peaks and intensity data. Statistical analyses revealed a good separation between samples; the best model based on peaks data shows 66.5% and 33.5% of total variance explained by the first and the second canonical functions and a predictive capacity of 78.9%. Regarding the intensity data, the CFs of the GDA model explain 53.8% and 46.2% of the total variance and the crossvalidation shows a predictive capacity of 96.7%.