Organic food fraud is a growing concern, particularly in the beef industry due to high price premiums and challenges in detection. This study addresses the need for rapid, cost-effective methods to verify organic beef authenticity and accurately identify meat cuts, which are often mislabeled. Previous methods such as stable isotope ratio analysis (SIRA) have shown some promise in distinguishing organic from conventional beef based on isotopic ratios related to feed sources. However, SIRA suffers from limitations such as high cost, time-consuming sample preparation, and susceptibility to inaccurate results due to variations in feed composition. This study explores the potential of REIMS, a rapid and cost-effective ambient mass spectrometry technique, as a more robust alternative. REIMS involves direct analysis of meat samples with minimal preparation, producing lipid profiles that reflect differences in production systems (organic vs. conventional) and meat cuts. The study's objective is to assess the accuracy of REIMS in differentiating organic and conventional beef and identifying different cuts, paving the way for a more reliable and efficient method for detecting beef fraud.

The literature highlights the vulnerability of organic beef to fraud due to high price premiums and difficulties in establishing reliable detection methods. While stable isotope ratio analysis (SIRA) has been investigated, its effectiveness is limited by its high cost, time-consuming sample preparation, and dependence on feed composition which can vary greatly between organic and conventional producers. Other methods such as fatty acid profiling have shown potential, but none have proven consistently robust in detecting organic fraud in beef. The use of veterinary medicines and antibiotics represents another potential avenue of detection, as their presence can often indicate non-organic rearing practices. However, even this approach has limitations as regulated antimicrobial use is decreasing in conventional farming as well. REIMS, a type of ambient mass spectrometry, has shown promise in food fraud detection in various products. This study builds upon prior research utilizing REIMS in detecting meat and fish fraud, adulteration and differentiating between different types of poultry, to address the specific challenge of organic beef authentication.

The study involved collecting 561 beef samples (285 conventional, 276 organic) from various carcass regions (neck, rump, shin) across Ireland and the UK. Samples were collected over a 13-month period from multiple abattoirs and included various cattle breeds (Aberdeen Angus, Limousin, Hereford, Simmental) for better representation of commercial beef production. Samples were stored at -80°C before analysis. REIMS analysis was performed using a Waters REIMS ion-source connected to a Waters Xevo G2-XS QTOF mass spectrometer. A bipolar probe burned the meat sample, creating an aerosol that was analyzed by the mass spectrometer. A 0.1 ng/µL lockmass solution of leucine enkephalin was infused for accurate mass correction. Data analysis involved using Waters MassLynx and Waters Abstract Model Builder (AMX) software. Preprocessing steps included background subtraction and normalization. PCA and PCA-LDA models were generated using the m/z 600–1000 mass range to differentiate between meat cuts (three-class model) and production systems (two-class model). A combined model assessed both aspects simultaneously. Cross-validation was used to assess model performance. The study employed a 20% leave-out cross-validation approach to evaluate the predictive power of models.

REIMS successfully identified meat cuts with high accuracy. An initial model showed 88.5% accuracy, and after correcting for mislabeled samples, the accuracy improved to 98.2%, demonstrating the system's ability to distinguish between neck, rump, and shin cuts based on lipid profile differences. In contrast, distinguishing organic from conventional beef proved more challenging. Initial models showed lower accuracy (below 60%), but with increased sample numbers, the cross-validation percentage reached 84%. Models created for individual meat cuts (neck, rump, shin) showed consistent accuracy (70-75%) in identifying the production system, suggesting that larger sample sizes are needed to improve accuracy. No individual biomarkers were identified to differentiate between organic and conventional beef; rather, the separation appeared to be driven by overall differences in lipid profile intensities. The study highlights the considerable difference in performance between identification of meat cuts and production system; meat cut being reliably identified at greater than 98%, and production system at 84%.

The high accuracy of REIMS in identifying meat cuts demonstrates its potential for detecting substitution fraud where cheaper cuts are misrepresented as higher-value cuts. The ability to simultaneously identify both meat cut and production system with reasonable accuracy makes REIMS a valuable tool for beef authentication. Compared to SIRA, REIMS offers significant advantages in terms of speed, cost, and ease of use, requiring no sample preparation and producing results within seconds. While the 84% accuracy in differentiating organic and conventional beef is lower than the accuracy in identifying meat cuts, it still represents a substantial improvement over SIRA and other existing methods. The limited separation observed in the models indicates that it is unlikely a single or small group of biomarkers will emerge to differentiate organic and conventional beef. Instead, many or all of the lipid species present likely contribute to the differences observed. The study suggests that minor changes in animal diet and husbandry, particularly those related to medications, can have a small but detectable effect on the resulting lipid profile. The results support the use of REIMS as a rapid and cost-effective method for beef fraud detection.

This study demonstrates the significant potential of REIMS as a rapid and cost-effective method for detecting various forms of beef fraud. The high accuracy in identifying meat cuts and the considerable accuracy in differentiating organic and conventional beef, coupled with the ease of use and minimal sample preparation required, makes REIMS a highly valuable tool for the beef industry. Future research could focus on expanding the sample size to further improve accuracy in differentiating organic and conventional beef and exploring the potential of incorporating organically produced biomarkers detectable by REIMS.

The study's accuracy in differentiating organic and conventional beef (84%) could be improved with larger sample sizes, representing a limitation of the current findings. Further investigation may be required to investigate the possibility of incorporating organically produced biomarkers detectable using REIMS. The current study focused on identifying meat cuts and production systems from samples obtained from Ireland and the UK. The generalizability of the findings to other geographical regions or different cattle breeds may need further validation. The study also relied on the accuracy of sample labelling, with errors in labelling having a significant impact on model performance.