Food fraud has become a topic of major concern over the past decade, primarily due to major incidents, such as the 2008 Chinese melamine scandal and the 2013 European horse meat scandal, and subsequent increased customer awareness and media coverage.
The cost of food crime is considerable. This expense has helped refocus attention on developing measures to ensure the integrity of the food supply chain, with an increase in demand for food fraud detection to be proactive, rapid, and reliable to maintain the security of the food chain while also acting as a deterrent.
There are many types of food fraud, including:
- Substitution of part or all the food with a lower value commodity;
- Addition of a component to increase the value of the overall product; and
- False claims on product labels that increase their value, such as “organic,” “welfare friendly,” “fair trade,” or “country of origin.”
Direct Analysis Using Mass Spectrometry (Direct MS)
The primary objective of authenticity testing is rapid verification, from raw ingredients through to finished (processed) products, to support traceability systems. Development of technologies that can be used to rapidly differentiate authentic products from fraudulent ones represents a significant challenge. One approach is the direct analysis of samples using mass spectrometry (MS) without any pre-treatment (e.g., extraction or chromatography). MS generated using various types of ambient ionization are used to create multivariate statistical models. Most applications have used linear discriminant analysis on principal component analysis (PCA-LDA) reduced data for the generation of predictive models, but others have explored machine learning approaches.
The result of the subsequent sample classification is presented and refreshed in real time. In all cases, validation is essential to evaluate the accuracy of the models. Analysis of a sample and the generation of results takes only a few seconds, enabling faster decisions and support for next steps. Let’s look at a few examples.
Rapid Evaporative Ionization Mass Spectrometry (REIMS)
REIMS allows for the collection of mass spectrometric data directly from the surface of biological samples, without any sample preparation. The technique was originally demonstrated to show promise for detection of cancerous tissues during surgery but has subsequently been used for investigation into food and beverage fraud, especially in the seafood and meat sectors. This work is conducted on a high-resolution instrument, the quadrupole time-of-flight (Q-TOF) mass spectrometer, to ensure enough selectivity to differentiate components and increase the specificity of the statistical model. REIMS typically uses a surgical diathermy sampling device, the iKnife, but there is growing interest in alternative means to generate the aerosol from the sample, such as other designs of monopolar probes, bipolar forceps, and use of lasers.
When it comes to fish and shellfish, we often don’t get what we ask for. Fraud is common. For example, one can get high-quality salmon substituted with lower quality salmon species, wild swapped for farmed, and, in some countries, rainbow trout is often mislabelled and sold as salmon. Typically, polymerase chain reaction (PCR) methods are used, which exploit minor differences in DNA sequence between different fish species. A small piece of fish DNA is copied many times using PCR and compared with a large, authenticated database of fish species using matcher software to ensure accurate fish species identification.
However, such techniques comprise multiple steps and can take hours. REIMS offers an accurate, high-throughput, cost-effective alternative to screen large numbers of samples for discrimination among fish species. After construction and validation of well-established models, the identity of blind fish fillets can be given in real time without any sample preparation. It has been demonstrated that REIMS can be applied as a rapid screening technique to detect various species of white fish, salmon, tuna, and other sea creatures, to complement existing DNA methods. In addition, there is some evidence that the same approach can be used to monitor the quality of products, such as shelf life and degree of lipid oxidation of fish oils during storage and in real time during cooking.