Nature-inspired machine learning algorithm to evaluate authenticity in various food matrices via MALDI-TOFMS technology

Background

During the last two decades, food authenticity studies have been found on the frontline of scientific research. The development in the field of food science aimed to achieve high nutritious, superior quality and safe food. Global recognition of products, known for their unique nutritional 15 properties, such as olive oil and dairy products, has raised awareness for quality assurance. The financial importance of these foods leads producers to various fraudulent practices, such as mislabelling or adulteration of products with illegal substances for profit purposes.

Technology Overview

The present invention relates to a methodology and a system for analysing food samples to determine product authenticity and ensure quality of the product and its compliance with particular specifications. The present invention further relates to a food traceability system that is configured to trace and authenticate food products at different steps of the production chain.

Stage of Development

Today, the invention is in TRL* 4, with the next steps being:

Optimization of the specific model, through the feedback of information (input data) and the parallel enrichment of the database for the application in different food products.

Benefits

• Fast and reliable holistic workflow to verify food authenticity.
• CNN-based workflow: quick learning ability, utilizing the whole profile.
• Applicable in various food matrices.
• Reliable chemical analysis, user-friendly GUI application and QR coding accessible to the end user.

Applications

Analytical Chemistry applied in Food Authenticity– Development of HRMS-based methodologies for food authenticity characterization in various food matrices/Exploitation of chemometrics and advanced statistical tools for the evaluation.

Opportunity

National & Kapodistrian University of Athens is seeking partners for the development of digital authenticity and traceability platforms.

Patents

• EP22201755.0
• PCT/EP2023/077945