Recently, a research project entitled ‘Food fingerprinting as a tool to control food authenticity’, has been successfully ended. This project focused on the detection of various types of food fraud in oregano and rice. Partners in the project were Ciboris (part of Primoris Holding), Ghent University, ML2Grow (company for advanced machine learning) and CRA-W.
Several analytical techniques, such as near infrared and mid-infrared spectroscopy, hyperspectral imaging, gas chromatography coupled to mass spectrometry and proton-transfer reaction time-of-flight mass spectrometry, combined with chemometrics, were examined to evaluate their potential to solve different food fraud and quality control issues.
Concerning oregano, successful models were made to determine the country of origin, to identify adulteration and for batch-to-batch control. With the available database of genuine oregano samples and chemometric models, it was possible to differentiate oregano samples from Italy, Turkey, Israel and South-America. It was also shown that batch-to-batch control from incoming raw materials can be achieved. Adulteration with sumac, myrtle, olive tree and cistus leaves was detected starting up from 10% adulteration. The model built was successfully tested on a set of blind samples (in red in the figure)
Concerning rice, it was possible to distinguish genuine rice samples coming from different countries i.e. Thailand, Vietnam, Spain, Italy and Pakistan. Additionally, different varieties as Basmati, White, Glutinous, Loto, Jsendra and Puntal rice could be also differentiated.
Data fusion was also performed to obtain more robust and more accurate classification models. For origin assessment of rice, a combination of NIR and GC-MS data permitted to develop a more performant classification model.
Funding:
VLAIO (Vlaamse Agentschap Innoveren en Ondernemen)
