Classification of olive oils according to geographical origin by using (1)H NMR fingerprinting combined with multivariate analysis (Journal article)
Longobardi, F./ Ventrella, A./ Napoli, C./ Humpfer, E./ Schutz, B./ Schafer, H./ Kontominas, M. G./ Sacco, A.
Authentic extravirgin olive oils from 7 different regions (Italy - 3 regions, Greece - 4 regions) have been investigated by (1)H Nuclear Magnetic Resonance (NMR) fingerprinting in combination with multivariate statistical analysis. In order to cover the dominating lipid signals as well as signals from compounds of low abundance in the oil, both a simple one pulse experiment and an experiment with multiple saturation of the lipid signals was applied to each sample. Thus, the dynamic range of concentrations covered by the two experiments was of the order of 100,000 allowing for a more comprehensive NMR assessment of the samples. Monte-Carlo embedded cross-validation was used to demonstrate that a combination of principal component analysis, canonical analysis, and classification via nearest class mean can be used to predict the origin of olive oil samples from (1)H NMR data. Given the rather limited number of samples tested, correct prediction probabilities of 78% were achieved with region specific correct predictions between 53% and 100%. (C) 2011 Elsevier Ltd. All rights reserved.
|Institution and School/Department of submitter:||Πανεπιστήμιο Ιωαννίνων. Σχολή Θετικών Επιστημών. Τμήμα Χημείας|
|Keywords:||(1)h nmr,fingerprinting,multi-signal suppression sequence,multivariate statistical analysis,olive oil,geographic origin,principal component analysis,fatty-acid-composition,pattern-recognition,authentication,cultivars,chemometrics,triglyceride,spectroscopy,phenotypes,networks|
|Link:||<Go to ISI>://000295666100025|
|Appears in Collections:||Άρθρα σε επιστημονικά περιοδικά ( Ανοικτά)|
Files in This Item:
|Longobardi-2012-Classification of ol.pdf||692.45 kB||Adobe PDF||View/Open Request a copy|
Please use this identifier to cite or link to this item:This item is a favorite for 0 people.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.