News

Analytical tests that are underpinned by statistical classification models are often based on reference sets of relatively few (in statistical terms) “authentic samples”.  There is a risk that these may not reflect the entire scope of variability within an authentic food population.  Laboratories building these models therefore need to take great care in how they process the reference data (e.g. dimension reduction, feature selection) to avoid the problem of over-fitting.  Over-fitting results in a statistical model too tightly tailored to the reference set which then fails when applied to a sample that differs in some way.  There is best-practice guidance available for laboratories – see signposts on FAN.

This latest research (open access) develops specific recommendations and a workflow for laboratories to deal with dimension reduction.  It comes from a statistical, rather than analytical, scientific journal. The authors evaluated different statistical approaches, using a model dataset of ICP-MS data from 28 apples of 4 origin classes.  They compared Linear Discriminant Analysis (LDA) and Partial Least Squares Discriminant Analysis (PLS-DA) algorithms. Their workflow integrated Principal Component Analysis (PCA) for feature extraction, followed by supervised classification using LDA and PLS-DA. Model performance and stability were systematically assessed. The dataset was processed with normalization, scaling, and transformation prior to modeling. Each model was validated via leave-one-out cross-validation and evaluated using accuracy, sensitivity, specificity, balanced accuracy, detection prevalence, p-value, and Cohen’s Kappa.

They report that, as a linear projection-based classifier, LDA provided higher robustness and interpretability in small and unbalanced datasets. In contrast, PLS-DA, which is optimized for covariance maximization, exhibits higher apparent sensitivity but lower reproducibility under similar conditions. They also emphasise the importance of dimensionality reduction strategies, such as PCA-based variable selection versus latent space extraction in PLS-DA, in controlling overfitting and improving model generalisability.

They conclude that their proposed algorithmic workflow provides a reproducible and statistically sound approach for evaluating discriminant methods in chemometric classification.

The EU Joint Research Centre (JRC) have now published their monthly collation of fraud media reports for July 2025 and September 2025 (these collations are published retrospectively, and August’s report was published in advance of July’s).  The full index of reports can be found here

These new reports have also been added to the JRC database that underpins a searchable front-end for media reports of food fraud incidents.  It allows filtering by commodity, country, fraud type and other key criteria.

The JRC collation is just one of the incident databases available.  Different databases collect different information, in different ways, and therefore show a different angle on the true picture.  All of these sources are signposted on FAN.  Best practice is to use a combination of all sources, but the final critical question is “how vulnerable is my own supplier”.

• JRC – These are solely media reports.  They exclude cases not in the public domain, and can be biased by shocking but highly localised incidents in local food supply within poorly regulated countries.  For the past few years, FAN member Bruno Sechet has produced a useful infographic based on each month's data
• EU Agri-Food Suspicions – These are solely EU Official Reports, and only suspicions.  The root cause of each incident is unknown.  The data include cases less likely to be deliberate fraud such as pesticide residues above their MRLs or unpermitted (but labelled) additives.  FAN produce our own infographic on a rolling 3-month basis.
• Food Industry Intelligence Network Fiin SME Hub – These are aggregated anonymised results from the testing programmes of large (mainly UK) food companies.  The testing programmes are targeted and risk-based, not randomised, and the fraud risks within the suppliers of large BRC-certified retailers and manufacturers may be different than the companies supplying small manufacturing businesses or hospitality firms.

Many testing laboratories also supply their own customers with incident collations, and there are many commercial software systems that scrape reports from the internet.  All collect and treat the data slightly differently.  FAN produce a free annual aggregate of "most adulterated foods" from three of the largest commercial providers, which gives very high level smoothed data.

It is notoriously difficult to collate fraud incidents in order to track trends and prioritise generic risks by either food commodity or country.  One of the more useful free tools for the past 10 years has been the monthly EU Joint Research Centre (JRC) collation of fraud media reports.

The JRC have just launched a searchable front-end for their database of reports.  It allows filtering by commodity, country, fraud type and other key criteria.

The JRC collation is just one of the incident databases available.  It must be remembered that different databases collect different information, in different ways, and therefore show a different angle on the true picture.  All of these sources are signposted on FAN.  Best practice is to use a combination of all sources, but the final critical question is “how vulnerable is my own supplier”.

• JRC – These are solely media reports.  They exclude cases not in the public domain, and can be biased by shocking but highly localised incidents in local food supply within poorly regulated countries.  For the past few years, FAN member Bruno Sechet has produced a useful infographic based on each month's data
• EU Agri-Food Suspicions – These are solely EU Official Reports, and only suspicions.  The root cause of each incident is unknown.  The data include pesticide residues above their MRLs.  FAN produce our own infographic on a rolling 3-month basis.
• Food Industry Intelligence Network Fiin SME Hub – These are aggregated anonymised results from the testing programmes of large (mainly UK) food companies.  The testing programmes are targeted and risk-based, not randomised, and the fraud risks within the suppliers of large BRC-certified retailers and manufacturers may be different than the companies supplying small manufacturing businesses or hospitality firms

Many testing laboratories also supply their own customers with incident collations, and there are many commercial software systems that scrape reports from the internet.  All collect and treat the data slightly differently.  FAN produce a free annual aggregate of "most adulterated foods" from three of the commercial providers, which gives very high level smoothed data.

One of the limiting factors in developing any untargeted analysis is sourcing samples for the reference database.  The samples labelled as “authentic” in the model developed must be of absolute trustworthiness (fully traceable back to authentic production) and also fully representative of every variable within the “authentic” scope (e.g. different permitted agricultural inputs, harvest seasons, species, variety and geographic origin).

This pilot study (open access) uses a statistical approach to circumvent this need.  The authors still attempted to source reference samples that were representative of an “authentic” scope but they made no attempt to verify the samples; all reference samples were purchased from online direct-to-consumer vendors.  They then built a results dataset of 38 different fatty acid methyl esters, tocopherols and phytosterols measured by GC-MS and LC-MS.  They selected from within this dataset using Monte Carlo sampling to choose different “reference databases” and build a large number of One Class Classification models.  They then used statistical analysis to see if each of these models appeared internally consistent (i.e. suggestive that all of the reference samples had been “authentic”) or not (i.e. suggestive that some of the reference samples had been “inauthentic”).  They chose the model with best internal consistency.

They piloted this approach on 40 avodado oils, identifying 6 of them a potentially adulterated.  Subsequent targeted chemical analysis showed that these 6 adulterated samples had been correctly identified.

This publication (open access) describes the launch of FISH-FIT.  FISH-FIT is a biobank of seafood species samples which are linked to an authentic database of morphology, genetic information, and other physical characteristics. It also contains a library of PCR analytical methods.   It was developed under an EU-funded project and free access is currently only available to EU regulatory bodies, although wider access is planned.  The databank is hosted by the Max Ruber Institute.

FISH-FIT has been added to FAN’s index of authenticity reference databases, a useful search tool for existing databanks or commercial testing services..

(image from the paper)

Food is at the center of the world’s most urgent challenges and largest opportunities.

According to the World Health Organization, malnutrition is the leading cause of death and disease globally. In fact, there is a “triple burden” of malnutrition at all levels of the population:

• Undernutrition: The lack of food and/or access to it. 
• Overnutrition: The consumption of too many calories.  
• Poor nutrition: Not the right nutritional content (vitamin and mineral deficiencies).

Given advances in the quality and cost of mass spectrometry, bioinformatics, machine learning and big data, along with the growing recognition of the important health impacts of food, the time is ripe for the PTFI.  

The PTFI will strengthen and support ongoing work by developing low–cost mass spectrometry kits, standards, methods, cloud-based analytical tools, and a public database that will include a quantitative and qualitative analysis of 1,000 foods that are representative of geographic and cultural diversity worldwide.

The PTFI will establish a Working Group, composed of experts around the globe, who will inform the selection of the first 2,000 foods based on specific criteria. The overarching goal of this selection process is to ensure inclusivity. The following dimensions we are considering arise out of provocations that help define the plenum of global food options: 

• Biology: Where in the phylogenetic tree did the organisms that become food originate? 
• Tissue: What part of organisms are used for food? Entire organisms or portions of plants, animals, or microbes? 
• Geography: Where do foods originate and where do they thrive? 
• Consumers: Who are specific foods targeted to? 
• Processing: Broadly speaking, how are foods treated after “harvest”? 
• Domestication: How has human intervention modified organisms from their native (wild) state? 
• Derivation and Formulation: Is the organism (plant, animal, microbe) consumed as a food as is, or is it a derived ingredient in a formulated product or recipe? 
• Proportional Abundance: From rice to spice – which foods are the center of a meal and the core of a cuisine, and which are tiny fractions of the diet, but can be just as frequently consumed? 
• Affordability: Which foods are luxury and which are staples? 
• Frequency: Which foods are consumed on a regular basis and which are associated with rare festive events, life transitions, spiritual celebrations? 
• Complementarity: Which foods are historically consumed as ensembles? 

Once the database is in place, the scientific community and private sector can build on this public resource by adding analysis of additional foods, varieties, and cooking methods. The PTFI technical platform will enable conditions for a rapid acceleration in research and innovation in both the public and private sectors.

Visit the website for further information.

Food adulteration is a growing concern worldwide. The collation and analysis of food adulteration cases is of immense significance for food safety regulation and research.

Research led by the Chinese Academy of Sciences collected 961 cases of food adulteration between 1998 and 2019 from the literature reports and announcements released by the Chinese government. Critical molecules were manually annotated in food adulteration substances as determined by food chemists, to build the first food adulteration database in China (http://www.rxnfinder.org/FADB-China/). This database is also the first molecular-level food adulteration database worldwide.

Additionally, the researchers propose an in silico method for predicting potentially illegal food additives on the basis of molecular fingerprints and similarity algorithms. Using this algorithm, we predict 1,919 chemicals that may be illegally added to food; these predictions can effectively assist in the discovery and prevention of emerging food adulteration.

The publication of this research has been published in Food Chemistry, doi: https://doi.org/10.1016/j.foodchem.2020.127010   

The FADB-CHINA database has been added to the 'Services' page of the Food Fraud Mitigation section.