News

In this study (open access) the researchers proposes using a MALDI-ToF and LC-Q-ToF dual approach, following trypsin digestion, as a method to verify fish species.  Trypsin digestion breaks the proteins down into peptides, and they used peptide fingerprints to identify peptides that were unique markers for specific species. The advantage of their approach over DNA methods, and in comparison to MALDI-ToF-MS analysis of undigested proteins, is that it can be applied to complex and heat-processed samples.

The study aimed to differentiate six fish species—carp, mackerel, pike, pollock, salmon and trout. Matrix-assisted laser desorption/ionization–time ff flight mass spectrometry (MALDI-TOF MS) was employed to identify characteristic species-specific m/z values to differentiate raw and cooked fish meat. Additionally, liquid chromatography–electrospray ionization–quadrupole–time tf flight (LC-ESI-Q-TOF) was used to determine specific amino acid sequences in carp and salmon, selected as model species.

Two or more distinct species-specific m/z markers were identified for all six fish species, enabling their differentiation in both raw and processed form. A slightly larger list of distinct markers were found for cooked, compared to raw, fish.  In carp and salmon, hundreds of peptide sequences were detected, leading to the identification of a panel of peptide markers that determine both the fish species and the type of meat processing. The results confirm that mass spectrometry-based proteomic approaches can serve as effective tools for the authentication of fish meat.

The authors conclude that it is possible to use two complementary mass spectrometry techniques for reliable and rapid authentication of fish species. By focusing on peptide-level markers and leveraging accessible tools, they believe that the approach offers a cost-effective and innovative alternative for fish meat authentication.

Nontargeted analysis for food authenticity by liquid chromatography–mass spectrometry (LC-MS) can provide data on thousands of chemical features. However, most studies that train machine learning models for food authentication use sample sizes in the tens or hundreds.  Such training sets are typically considered too small to be optimal, as it introduces the problem of overfitting when working with such a large feature-to-sample ratio.

This study (open access) aimed to mitigate this issue with a machine learning protocol designed for sub-optimal training sets, using honey as an example.   A recursive feature elimination (RFE) pipeline was developed specifically to address the challenges of optimizing the honey chemical fingerprint for multiclass machine learning classifiers on a limited number of samples with imperfect labels. A support vector machine was used for both RFE and classification to reduce the 2028 nontargeted features down to just 54 features (a 97.3% reduction) without any loss of classification performance.

The authors report that the resulting model was a 6-class classifier, capable of identifying monofloral blueberry, buckwheat, clover, goldenrod, linden, or other honey with a nested cross-validation Matthews correlation coefficient (MCC) of 0.803 ± 0.046. The development of a k-nearest neighbours filter and the decision to continue the RFE process beyond the iteration with the highest classification score were instrumental in achieving this outcome.

They conclude that this work shows a complete pipeline that automates feature selection from nontargeted LC-MS spectra when working with a limited number of samples and imperfect labels. This process can also be expanded to other food groups and spectral data.

Photo by Andrea De Santis on Unsplash

Further to the review of methods used in the EU co-ordinated survey of honey collected 2021-22 (“From the Hives” survey) – see previous blog here:

The Joint Research Centre of the European Commission have now published (open access) more details of some of the test methods used and results interpretation.  This publication relates particularly to the two qualitative Liquid Chromatography–High-Resolution Mass Spectrometry (LC-HRMS) methods that were developed to detect mannose (Man), difructose anhydride III (DFA III), 2-acetylfuran-3-glucopyranoside (AFGP), and oligo-/polysaccharides with degrees of polymerization (DPs) of 6 to 11.

The presence of mannose and unusual oligo-/polysaccharides was the main reason that many of the samples were flagged as “suspicious” in the previously published report.

There is no single definitive test for dilution of honey with foreign sugar syrups.  An untargeted test, often used to contribute to an analytical weight of evidence, is proton NMR followed by chemometric pattern recognition based on variations in the sugars profile.  One disadvantage of this technique is a lack of sensitivity. 

LCMS is a more sensitive technique and could – in principle – be used in a similar untargeted manner to drive pattern recognition statistics based on the sugar profiles of a database of reference honeys.  The limiting factor has been the computing power that would be needed to “re-set” the database each time a new chromatographic peak is measured or data from different chromatographic systems are combined. (this is why untargeted LCMS is often used in authenticity testing as a 1-off development tool to identify marker compounds, which are then used as the basis for a more routine targeted test, rather than being used as a routine untargeted test).

In this paper (open access), the authors resolved the computing power limitation by using their Bucketing of Untargeted LC-MS Spectra (BOULS) data processing approach which they have previously published.  They demonstrated that untargeted LCMS testing (combining data from different systems, HILIC column with MS in both positive and negative ionisation mode) could discriminate a range of adulterated honeys (rice, beet and high-fructose corn syrups added at 5% to a reference set of 34 North German honeys) from their unadulterated counterparts.

As is the case with all untargeted analytical techniques, the key to using this method routinely would be building a robust reference database of verified authentic honeys that is fully representative of all types and origins on the market.

Photo by Roberta Sorge on Unsplash