Geographical screening of EVO oil, developed a non-targeted model to identify if the area of origin is Italian

 
Researchers from the IZSVe Experimental Chemistry Laboratory

 

 

Researchers from the Experimental Chemistry Laboratory  of the Istituto Zooprofilattico Sperimentale delle Venezie (IZSVe) have developed a simple and effective method of geographic screening to discriminate extra virgin olive oil (EVO) of Italian origin from those of other sources. The non-targeted method, based on the integration of Raman and near infrared (NIR) spectroscopic techniques, was published in the Journal of Near Infrared Spectroscopy .

 

 

Due to climate change and the Xylella fastidiosa epidemic, the annual production of Italian EVO oil has recently decreased. To meet the needs of consumers and the food industry, Italy is forced to import oils from other countries, running the risk of seeing an increase in the number of frauds, i.e. blends of Italian oil and oils of European origin and extra-European sold as oils of Italian origin. Researchers from the IZSVe Experimental Chemistry Laboratory (SCS8) have developed a quick method to distinguish Italian oil from foreign (in this case Greek) oil based on the chemical composition of the sample.

 

 

 

The research

 

 

The IZSVe approach combines spectroscopic and machine learning techniques: starting from the exploratory analysis of the data collected in the laboratory, through the identification of the most significant variables, the creation of a statistical model capable of identifying the specific chemical characteristics of oils under investigation to unmask any commercial fraud. Subsequently, the model was subjected to validation with unknown samples, thus determining its predictive capabilities.

 

 

33 oil samples were analyzed by Raman and NIR spectroscopy, that is non-destructive techniques, which provide answers in a short time and do not require sample preparation, minimizing time and costs. Furthermore, these techniques have the advantage of being sustainable (for example no solvents are used) and of responding to the principles of green chemistry .

 

 

Raman spectra were acquired in the 3200-300 cm-1 wave number range, while NIR spectra were acquired in the 12000-4000 cm-1 range. All the acquired spectra were subjected to normalization and then statistically merged by means of mid-level data fusion, which was followed by supervised statistical partial least squares discriminant analysis (PLS-DA), which has allowed to clearly distinguish the two study groups, Greek and Italian oils.

The PLS-DA model thus obtained was then validated reaching an accuracy of 83% and sensitivity and specificity of 80% and 100%.

 

 

 

A tool for fraud prevention

 

 

The integrated analysis system proposed by the IZSVe study allows to obtain a rapid and complete identikit of the composition of the food product, also characterizing it at a geographical level. Thanks to this information, which is extremely valuable for both producers and consumers, it is possible to guarantee the traceability and quality of extra virgin olive oil, preventing commercial fraud that may also have health implications.

 

 

Data fusion obtained with NIR and Raman spectroscopy was very promising: the classification model can be considered to all intents and purposes an effective tool for the discrimination of EVO oils from Italian olives from those of foreign origin. As this is a preliminary study, new samples will be added to the initial data set to improve the accuracy of the model. The tests on the oil matrix are added to previous studies conducted by the Experimental Chemistry Laboratory on other food products (for example oreganofish and milk)), which have equally confirmed the effectiveness and strategic nature of the non-targeted approach in defining reliable the authenticity of the food.

 

 

 

Read the article on Journal of Near Infrared Spectroscopy 

 

 

 

 

 

Source: IZSVe

 
 
 

NATIONAL REFERENCE CENTRE FOR VETERINARY EPIDEMIOLOGY, PROGRAMMING AND INFORMATION AND RISK ANALYSIS (COVEPI)
Daniela Morelli

National Reference Centre for Risk Analysis
Armando Giovannini

Epidemiology
Paolo Calistri

Statistics and GIS
Annamaria Conte

EDITORIAL STAFF

e-mail benv@izs.it

fax +39 0861 332251

Cookie Policy

 

Coordination
Simona Iannetti
Francesca Dall'Acqua

Editorial board
Barbara Alessandrini, Annamaria Conte, Fabrizio De Massis, Armando Giovannini, Paolo Calistri, Federica Monaco, Giovanni Savini

Istructional designer
Alessandro De Luca

Web master
Sandro Santarelli

logo IZSAM