News

The EC Monthly Reports of Agri-Food Fraud Suspicions reports are a useful tool for estimating fraud incidents, signposted on FAN’s Reports page.  The March 2025 report was added earlier this week and can be found here.

FAN has produced this rolling 3-month graphical analysis. We have excluded cases which appear to be unauthorised sale but no intent to mislead consumers of the content/ingredients of a food pack (e.g. unapproved food additives, novel foods), excluded unauthorised health claims on supplements, and we have excluded residues and contaminants above legal limits.  We have grouped the remaining incidents into crude categories.  Our analysis is subjective but intended to give a high-level overview.  One consistent stand-out is unlicenced production, trade or import in high risk foods, often backed up by forged documentation.  Although the details behind the reports are not public, allegorical evidence is that these cases are not just "grey market" trade to small shops and market stalls.  Much of the suspicious trade enters mainstream markets.

As with all incident collation reports, interpretation must be drawn with care.  This EC collation is drawn from the iRASSF system – these are not confirmed as fraud, and the root cause of each issue is usually not public.  There are important differences in the data sources, and thus the interpretation that can be drawn, of these data compared to other incident collations.  For example:

• JRC Monthly Food Fraud Summaries (which underpin the infographics produced monthly by FAN member Bruno Sechet) - these are unverified media reports, rather than official reports, but hugely valuable in giving an idea of which way the fraud winds are blowing
• Official reports (as collated from commercial databases such as Fera Horizonscan or Merieux Safety Hud, which underpin FAN's annual Most Adulterated Foods aggregation) - these are fewer in number and give a much more conservative estimate of fraud incidence, and may miss some aspects which have not been officially reported
• Verified reports (where the root cause has been scrutinised and interpreted by a human analyst, for example the FoodChainID commercial database) - these are also few in number, less suitable for drawing overall trends, but give specific insight and information.

If looking at trends over time, you must also be wary of step-changes due to the introduction of new data sources.  For example, Turkey's public "name-and-shame" database of foods subject to local authority sanctions went online in January 2025 and has resulted in an apparant increase in incident reports from Turkey.

EFF-CoP has launched its website: 🌐 https://www.eff-cop.eu or scan the QR code to explore:

• Everything you need to know about the EFF-CoP project
• Profiles of EFF-CoP partners.
• EFF-CoP promotional materials.
• The evolving EFF-CoP landscape (including the Food Authenticity Network).
• The latest updates: news, events, newsletters, and more.

📣 Coming soon: EFF-HUB!
The EFF-HUB, your one-stop digital space to connect, collaborate, and stay informed on food fraud prevention will be vailable very soon.

🎧 And if you speak Greek, don’t miss the EFF-CoP feature on CNN Greece where EFF-CoP partner Smart Agro Hub (Dissemination- Exploitation- Communication Work Package Leader), represented by Evangelia Zavitsanou, joined the journalist Kostas Pliakos in a podcast about food fraud, food safety, and how consumers can protect themselves and make informed choices.

Listen to the podcast here:
▶️ CNN Greece
🎧Spotify
🍏 Apple Podcasts

The FAN EFF-CoP page has also been updated.

In this study (purchase required) the authors propose and develop a strategy for a field-based screening test for crude honey adulteration (adulteration with inverted sugars) using Near InfraRed Spectroscopy (NIR) hand-held scanners.  They developed a single class classification model that was sufficient to either give an “unadulterated” verdict or to refer the sample for confirmatory (IRMS) analysis.

The authors developed their SIMCA model using “genuine” adulterated honeys that had been previously seized in a Brazilian police operation that had cracked down on industrial-scale addition of invert sugar to honeys over a three year period, along with unadulterated honeys collected by police during the same operation.  The traditional SIMCA was improved by optimizing the class boundaries based on receiver operating characteristic (ROC) curves and the estimate of an uncertainty region, thus optimising the model for a screening application.

Photo by Roberta Sorge on Unsplash

Dehydrated bee pollen is a premium product with a growing market as a food supplement, functional food ingredient, and is also an ingredient in biomedical and healthcare formulations.  Pollen from certain species of stingless bees is higher priced than the equivalent from the much more common honey bee, as the former is considered to have enhanced nutritional and health benefits.  Pollens from different bee species are visually identical, giving an incentive for potential fraud.

In this paper (purchase required) the authors built a classification model to discriminate pollens from different bee species.  The researchers integrated digital image processing and machine learning to classify pollen loads produced by honey bees (Apis mellifera) and pot-pollen from stingless bee species based on their colour patterns. A total of 246 pollen loads and pot-pollen samples from five bee species (Apis mellifera, Melipona marginata, Melipona quadrifasciata quadrifasciata, Scaptotrigona bipunctata, and Tetragona clavipes) were collected, and high-resolution images were captured using a smartphone. Colour parameters extracted from images such as R, B, H, and V were analyzed, and classification models employing CatBoost, XGBoost, Random Forest, and k-Nearest Neighbors (kNN) algorithms were tested.

They reported that CatBoost achieved the highest performance, with accuracies of 100 % in the training phase and 98.9 % in the testing phase. Linear Discriminant Analysis (LDA) further validated the classification by grouping the pollen samples into five distinct clusters corresponding to the bee species studied.

They conclude that combining digital image processing from a smartphone with machine learning offers an effective approach to classifying pollen from different bee species, with promising applications in apiculture to ensure product quality and authenticity.

Photo by Aaron Burden on Unsplash

In this study (open access) a non-targeted method of headspace-solid phase microextraction with gas chromatography coupled to mass spectrometry (HS-SPME-GC–MS) was developed to achieve the characterization, classification, and authentication of different coffee samples according to geographical production region, and variety (arabica/robusta). Moreover, decaffeinated and non-decaffeinated instant coffee samples were analyzed. Some samples of chicory, a potential coffee adulterant, were also been included. The GC–MS fingerprints were used to classify and characterise the analyzed coffees using principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA) and partial least squares (PLS) regression.

185 reference samples were used.  42 were chicory (a typical coffee substitute or adulterant), 96 were coffee from three geographical production regions (Vietnam, Cambodia, and Costa Rica) and species (Arabica, Robusta, and Arabica-Robusta mixture), and 47 samples were soluble coffee (decaffeinated and non-decaffeinated). The chicory samples were purchased from Barcelona supermarkets and the coffee samples were from Vietnam, Cambodia and Costa Rica local supermarkets. Each paired PLS-DA model was built using 70 % of samples randomly selected from each group as the calibration set while the remaining 30 % of the samples were employed as the prediction set.  The authors compared models generated using two different GC columns and operating conditions.

The authors tested their model on mixtures prepared in-house from the same reference set:: Vietnamese Arabica Coffee adulterated with Vietnamese Robusta Coffee and Vietnamese Robusta Coffee adulterated with chicory.  They reported that the model could classify adulteration levels down to 15%.

(image from the paper)

A recent review by FranZ Ulberth and Robert Koeber describes the role of RMs in food authenticity testing, including their applications in method validation, calibration, quality control, and the definition of conventional measurement scales.

It also reviews the availability of RMs that can be used in measurement procedures to authenticate food. Furthermore, the applications of RMs in targeted adulterant detection methods, for compositional parameters used to authenticate foods and food supplements, isotopic measurements, untargeted food authenticity testing methods, and detection and quantification of genetically modified organisms (GMOs), are explored.

The review concludes by recommending the development of research grade test materials or representative test materials to harmonise untargeted testing methods and improve comparability of results across laboratories and over time.

Access the full review: DOI https://doi.org/10.1007/s00216-025-05743-0

The review has also been added to the Quality section of the FAN website.

Mechanically Separated Meat (MSM – sometimes called Mechanically Recovered Meat, MRM) must be declared on-pack if used in meat products.  Testing for undeclared MSM can be a significant analytical challenge.  Traditional official control methods rely on microscopy, which requires experienced interpretation and can be highly subjective. 

This paper (open access) builds upon previous published work from the same researchers to develop targeted LC-MSMS methods that are suitable for official control applications..

In contrast to a comparable study on MSM from poultry, the authors report that the use of cartilage/intervertebral disc material was not useful for porcine MSM. They therefore report a new marker protein from porcine MSM, protegrin-4, which allows the detection of 5/3/1 mm MSM. The validity of the developed assay was ensured by the investigation of 182 blinded samples. After unblinding, all samples containing 5/3/1 mm MSM and all negative control samples were correctly classified. Additional new results related to the investigation of the species specification of chicken, turkey, and pork also are presented.

They concluded that LC-MS/MS-based detection of of undeclared MSM has been successfully extended from poultry to porcine MSM. The assay was successfully transferred to a triple quadrupole LC-MS system to facilitate routine use.

Photo by Branimir Petakov on Unsplash

This paper (open access) reviews the history and exportation of turmeric in Africa and the safety issues of some toxic adulterants.

Priority adulterants were determined from global food safety alerts. A systematic bibliographic search was performed to identify appropriate methods and techniques for authentication and safety testing. The quality of each study was assessed according to PRISMA guidelines/protocol.

The authors report that African turmeric exportation is on the rise due to recent insights into the suitability of local cultivars, soil and climate for growing high-quality turmeric. There are limited data on turmeric adulteration for domestic consumption and export markets..

Global alert databases revealed lead chromate as the top hazard identified of all adulterants. Current techniques to detect adulterants are laboratory-based, and while efficient, there is a need for more rapid, field-friendly, non-destructive analytical tools.  The authors consider that – if lead chromate is considered to be the main tisk - then pXRF would be ideally suited as a field-based test in Africa. In the hope that it could be further developed and calibrated to detect below the regulatory level of 1.5 mg/kg lead in turmeric powder. There would be a need to cross-check pXRF screening results against a validated and accredited ICP-MS method as a reliable confirmatory tool.

Photo by Md Shakil Photography on Unsplash

Authenticity testing of honey is the best-known example of the need for a weight-of-evidence approach.  One analytical test is unlikely to give a definitive answer.  Using a panel of different tests, using different techniques and principles, can give an incremental list of suspicions.

The routine use of machine learning for constructing reference databases has enabled the rapid expansion of techniques that – in principle – can discriminate differences between “authentic” and “inauthentic” reference samples, and thus could be added to this weight-of-evidence armoury.  Two recent publications are a case in point.

In the first study (purchase required) the authors produced their own reference set by adulterating honey with syrups, then showed that they could be discriminated from authentic honey using Differential Scanning Calorimetry (DSC)  (they used graph-based semi-supervised learning to construct the classification model).  DSC is based on melting curves, and is an indirect measure of water content.  It is a cheap, routine, test widely used in many industries and – as such – would be ideal as the first step in an analytical workflow in order to screen out the most crudely adulterated samples.

This second study (purchase also required) is an example of building much more focussed and granular reference databases to address a highly specific authenticity question.  The authors measured carbon-13 ratios in honeys (and in their constituent protein) from 196 authentic honeys sourced from 56 cities in Turkey.  This analytical technique is usually used as a marker for exogenous sugar addition, but in this study the authors used the more subtle variations (driven by changes in flora, temperatures and humidities) to build a classification model for regional origin.  They were able to cluster the honeys into one of 7 distinct geographical regions of Turkey based upon their carbon-13 ratios.

[thanks to FAN-member Peter Farnell for spotting that the original version of this blog referenced "Differential Scanning Colourimetry" - which would, indeed, have been a novel technique worthy of comment]

A recent New Zealand case is a reminder that any food – or related  product - where the price is based on an indirect marker of quality can be a target of adulteration.  This is an increasing watch-out with the industry focus on waste valorisation and the circular economy.

In this case, the perpetrators (commercial meat producers) manipulated the Free Fatty Acid profile of tallow by adding extraneous fats and oils.  A lower FFA profile increases tallow’s value in the biofuels market.  The adulterated tallow was destined for export into biofuels feedstock.

The fraud was brought to light by a whistleblower.  The perpetrators have been fined 1.6 million NZ dollars.

A full media report is here

Photo by Wilman Aro on Unsplash

.

The UK Serious Fraud Office (SFO) have issued guidance that clarifies that if a corporate self-reports suspected wrongdoing and co-operates fully with investigators, it can expect to be invited to negotiate a Deferred Prosecution Agreement (DPA) rather than face prosecution.

Whether a DPA applies is a judicial decision, but a key consideration is how promptly, after the fraud was uncovered, the company then self-reports.  The guidance makes clear that “waiting for an internal investigation to conclude” is not an excuse to delay reporting.

SFO Director Nick Ephgrave put it in these terms at the Guidance launch event on 24 April. “As soon as you have reasonable suspicion – it doesn’t have to be ‘beyond reasonable doubt’, doesn’t have to be ‘on the balance of probabilities’, just reasonable suspicion you’ve got some offending going on – that’s the point at which you stop, that’s the point at which you talk to us.”

The guidance also includes details on how to report, including a secure online self-reporting form.

This paper (open access) introduces the workflow MEATiCode, a comprehensive proteomic liquid chromatography tandem mass spectrometry (LC-MS/MS) method for the simultaneous identification of species in meat authentication.

This novel database search approach enabled the differentiation of meat species (as demonstrated for beef, pork, chicken and lamb) in raw and cooked food products following a simple sample preparation procedure and LC-MS/MS analysis of extracted meat peptides.  Peptides and proteins were characterised from reference samples using an untargeted protocol.  The MEATiCode database was then constructed in the Mascot Server search engine, with the objective of creating artificial proteins comprising the concatenated amino acid sequences of the peptides identified as specific for each species.

The authors report that the efficacy of the MEATiCode method was demonstrated through its application to a range of meat products, achieving high sensitivity (0.5 % Limit of Detection (LoD)) and reliability in the detection of adulteration, even in highly processed or cooked meats.

📥🧾🪧🔖 EFF-CoP promotional materials are ready for distribution. To help spread the word and engage new members, EFF-CoP has developed a range of promotional materials designed to inform, inspire, and connect!

Here’s what you can find:

• Leaflet - A quick and clear introduction to the EFF-CoP mission, objectives, and activities. Perfect for events, classrooms, and partners.
• Stickers - Fun, visual reminders of the EFF-CoP identity - great for students, notebooks, or toolkits.
• Bookmarks - A small but meaningful way to stay connected with EFF-CoP while promoting learning and awareness.

These materials reflect the heart of the project: to build a robust, collaborative European network focused on eradicating food fraud through collaboration, innovation, and shared knowledge.

🌐 EFF-CoP’s website and HUB are currently under construction and will be available soon.

By becoming a member, you will be able to read and write articles focused on food fraud, and engage in conversations with people, experts, and stakeholders who are passionate about this topic. You will also be able to download all promotional materials.

In the meantime, stay tuned to EFF-CoP’s LinkedIn profile, where we will keep you informed about all upcoming updates.

The Instutute of Food Science and Technology (IFST) has published a Technical Brief on the difference between Food Risks vs. Hazards.

John Points (FAN's very own Technical Director) and Peter Wareing dissect this critical distinction with clarity and real-world relevance. From unpacking ISO terminology to examining practical case scenarios, such as allergen mislabelling or aflatoxins in confectionery, this is essential reading for anyone involved in food production, safety, or regulation.

The brief explores:
– How to assess risk magnitude using likelihood and consequence
– When to withdraw, recall, or trade through
– Why pre-emptive planning beats reactive chaos
– What ‘tolerable risk’ really means under UK law.

Access this free Technical Brief here.

This brief has also been added to the 'Report' part of the 'Tools /Guides /Reports' tab of FAN's Food Fraud Prevention section.

There is a premium market, particularly in the US, for beef labelled as “grass-fed”.  From 2007 to 2016, the United States Department of Agriculture (USDA) carried voluntary marketing claim standards to help regulate grass-fed beef (GFB). These standards were discontinued, but producers can still seek approval from the USDA to market GFB.  This can only come from meat derived from cattle fed 100% forage, but the USDA also allows for partial claims (e.g., 50% grass-fed). Participating producers can define their own claim and need to comply with written protocols and sign an affidavit, but no audits are conducted.

In this study, (open access) three reference populations were used; 100% grass-fed, grain-fed, and grape-supplemented.  Red Angus steers (n = 54) were randomly allocated to one of the three feed regimes. Beef samples were collected in September 2019 and October 2020 in a USDA-regulated slaughter facility. All animals were slaughtered on the same day at 16–18 months old for GRAIN and GRAPE and 24–26 months old for GRASS. Ribeye samples were collected from the left side of the carcass between the 11th and 13th rib.

A multi-omics approach (gene expression quantification, metabolomics, and fatty acid [FA] profiling) was used to classify the three groups.  FAs were measured by gas chromatography-mass spectrometry (GC–MS), secondary metabolites were identified using ultra-high-performance liquid chromatography tandem mass spectrometry (UPLC–MS/MS), and gene expression analysis was performed using quantitative reverse transcription polymerase chain reaction (RT–qPCR).

The authors report that all target genes were upregulated in beef from GRASS compared to the other two groups. Multivariate analyses showed that long-chain n-3 polyunsaturated FAs, the n-6:n-3 ratio, vitamin E, organic acids, amino acid derivatives, and the nephronectin isoform X1 (NPNT-1) gene were the most important compounds for group separation. These compounds showed higher concentrations in beef from GRASS.

The success of beef separation by dietary treatment was highlighted by the 90.4% prediction accuracy of the random forest model, with beef from GRASS being 100% accurately predicted and beef from GRAPE being 94.4% accurately predicted. Beef from GRAIN was 76.5% accurately predicted.

The authors conclude that coupling gene expression analysis to metabolomics and FA profiling allowed for the separation of beef samples from varying dietary backgrounds with a high degree of confidence.

Thank you to FAN member Lucas Krusinski for flagging this article

This is a first in a series of invited blogs from FAN's laboratory Centres of Expertise.  In this article, Christophe Noel from SGS's food analysis molecular biology team, discusses the likely challenges in authenticating lab-grown meat.

The Infinite Animal: Mutation, Identity, and Authentication in Lab-Grown Meat

Lab-grown meat, or cultured meat, presents a new frontier in food production.  It introduces complex authentication challenges involving traceability, product integrity, and regulatory compliance.

With novelty and uncharted territory comes the need for new and emerging technological and regulatory strategies, as well as a combination of different methodological approaches to address these issues.

For example, the use of molecular and isotopic fingerprinting to distinguish lab-grown meat from conventional meat products—by analysing specific metabolic markers or isotopic ratios unique to cell culture processes—appears to be a promising option. DNA-based methods, including barcoding and genetic tracing, are also being proposed to verify cell line origins and production authenticity. Additionally, blockchain technology could offer transparent supply chain management, providing immutable records of production steps from cell culture to final product labelling, which would be extremely valuable.

Regulatory bodies are contributing by drafting frameworks that require rigorous documentation and verification at each stage of production, helping to establish standards for what legally constitutes "cultured meat."

Despite these advances, the field remains in its infancy, and ongoing research is crucial for validating these methods across different production platforms and global regulatory systems.

When considering the potential role of DNA/RNA analysis, its scope can include confirming the animal species present, detecting potential adulteration, identifying the nature of heterogeneous scaffolding, or even verifying the brand of the product if a unique genetic signature—either naturally occurring or engineered as a molecular label—is used.

An important aspect to consider in terms of authentication is the genetic stability of the product. A fundamental characteristic of cultured meat is that once cells are collected from the animal, this step is not repeated. The cell line must become immortal, ideally multiplying indefinitely and remaining genetically identical. However, cell culturing and propagation are, by nature, driving forces for potential mutations, raising the question of how we monitor these effects and their impact on the integrity of the final product.

Ultimately, an important question for authentication arises: are we destined to eat a definitive, unchanging version of one animal forever?

SGS Analytics United Kingdom Ltd is a UKAS (ISO17025) accredited analytical laboratory specialising in molecular and immunological detection of contaminants and adulterants in foodstuffs. With specific reference to authenticity testing our core area of expertise is the detection of meat, plant, fish and crustaceans using endpoint PCR; real time-PCR, PCR-sequencing and Next Generation Sequencing based technologies.  FAN has more information here.

In this study (purchase required) the researchers used bioinformatics methods to identify specific sequences of cattle, pig, chicken, and duck, and designed primers and probes accordingly.

They developed a method based on recombinase polymerase amplification (RPA) combined with lateral flow dipstick (LFD) for rapid visual authentication of beef and beef products. The RPA reaction was conducted at 37℃ for 20 min. The amplification products were then diluted and applied to the sample pad of the LFD. Results were visible to the naked eye within 5 minutes.

They report that the results demonstrated the method could specifically differentiate components of bovine, porcine, chicken, and duck origin, with a limit of detection (LOD) of approximately 20 copies for each species.

They applied the method to 10 commercially available beef products. Of which, five samples were detected with porcine-derived components. The results of the RPA–LFD method were verified using PCR and observed to be consistent between the methods.

The researchers conclude that this method is easy to use, requires no specialized equipment, and delivers results in about 30 min from amplification to detection, making it suitable for rapid visual detection on-site.

This study (open access) proposes a strategy to verify the authenticity of Mozzarella di Bufala Campana (MdBC).  MdBC is, a Protected Designation of Origin (PDO) cheese, Buffalo breeds are highly similar genetically, so detecting foreign buffalo milk in commercial cheese is more complicated than identifying cow, goat, or sheep milk. Fraud involving cow milk is particularly concerning because it is cheaper and more widely available, especially during peak MdBC production seasons

The researchers used a reference set of sixty-four anonymized PDO MdBC and foreign mozzarella samples provided by the Italian Central Inspectorate for Fraud Repression and Quality Protection of the Agrifood Products and Food, Ministry of Agricultural and Forestry Policies (Rome, Italy).  They used a sequential approach to verifying foreign milk species in buffalo mozzarella.  As a first screen, the casein was separated on a polyacrylamide gel.  This was generally sufficient to identify extraneous cows’ milk proteins.  In a second stage, the isolate casein was then digested with trypsin and the peptides analysed by MALDI-ToF-MS.

In cases requiring confirmation, nano-liquid chromatography coupled to electrospray tandem mass spectrometry (nano-LC-ESI-MS/MS) is used in central state laboratories for the highly sensitive detection of extraneous milk proteins in PDO buffalo MdBC cheese. The researchers report that analysis of the pH 4.6 soluble fraction from buffalo blue cheese identified 2828 buffalo-derived peptides and several bovine specific peptides, confirming milk adulteration.

They conclude that, despite a lower detection extent in the pH 4.6 insoluble fraction following tryptic hydrolysis, the presence of bovine peptides was still sufficient to verify fraud. This integrated proteomic approach, which combines electrophoresis and mass spectrometry technologies, significantly improves milk adulteration detection.

Photo by Audric Wonkam on Unsplash

In common with most jurisdictions, India has regulatory analytical criteria for authentic honey.  This includes various stable isotope ratios.

In this study (open access) the researchers set out to construct an analytical database of fully traceable authentic honeys in order to verify the criteria set by the Food Safety and Standards Authority of India.

They collected 98 authentic samples (covering 19 botanical sources, 42% multifloral and 58% monofloral).  They covered 17 states and provinces.  Sample were from collection centres of the All-India Coordinated Research Project on Honey Bees and Pollinators (AICRP, HB&P), under the auspices of the Indian Council of Agricultural Research (ICAR).   In addition, beekeepers registered with the National Bee Board (NBB) were also identified for sample collection. All samples were fully traceable.

The researchers generated a database of stable carbon isotope ratios (¹³C/¹²C) by Elemental Analyzer/Liquid Chromatography–Isotopic Ratio Mass Spectrometry (EA/LC-IRMS). The samples were analyzed for the parameters δ¹³C_Honey(δ¹³C_H), δ¹³C_Protein(δ¹³C_P), δ¹³C individual sugars, ∆δ¹³C_{Protein-Honey}(δ¹³C_P-H), C4 sugar, ∆δ¹³C_{Fructose-Glucose}(δ¹³C_Fru-Glu), ∆δ¹³C_max, and foreign oligosaccharides as per the official methods of analysis of the Association of Official Analytical Chemists (AOAC 998.12) and the FSSAI.

The results were evaluated against the published literature and Indian regulatory criteria for authentic honey. The δ¹³C value for honey (δ¹³C_H) ranged from −22.07 to −29.02‰. It was found that 94% of samples met the criteria for Δδ¹³C_P-H (≥−1.0‰), Δδ¹³C_Fru-Glu (±1.0‰), and C4 sugar content (7% maximum), with negative C4 sugar values treated as 0% as prescribed by the AOAC method.  86% of samples met the accepted foreign oligosaccharide criteria (maximum 0.7% peak area).

They conclude that the data of this study provide scientific backing for these four parameters as per the FSSAI regulation. However, the non-compliance of a high number (47%) of authentic honey samples for Δδ¹³C_max (±2.1‰) compels further systematic investigation with a special focus on bee feeding practices. Further, they found that honey samples with a Δδ¹³C_P-H greater than +1‰ and a C4 sugar content more negative than −7% also did not comply with the Δδ¹³C_max criteria. They suggest that Δδ¹³C_P-H values (>+1‰ equivalent to C4 sugar < −7%) could be an indicator of C3 adulteration to some extent.

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.