The application and development of high throughput, high-resolution mass spectrometry techniques
: towards faster and more robust profiling of small molecules in complex matrices.

  • Sufyan Pandor

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

Incorporating mass spectrometry-based metabolite profiling into food integrity studies holds the potential to detect and identify markers to tackle many issues around food safety, complex food fraud, and other important attributes such as food quality. The last decade has seen significant technological advances as well as increased accessibility of high-resolution mass spectrometry in routine laboratory settings rather than considering the technique as purely a research tool. The range of potential contaminants and residues requiring routine monitoring is growing rapidly and includes pesticides, veterinary drugs, synthetic chemicals, natural toxins, and adulterants. Full mass spectral acquisition in high resolution has clear benefits of unit mass instruments allowing an unlimited number of compounds to be measured as well as having the ability to multiplex methods. Important challenges do exist such as the successful transfer of analytical methods to industry, and robust quality control of the entire workflow from sample preparation to the resulting data analysis is required.

Typically, the workflow used to profile small molecules in food is similar to that used in the field of metabolomics applied to studying small molecules in cells, biofluids, tissues, or organisms. However successful profiling in food has very different challenges due to the complex chemical composition of food, with many different classes of compounds combining to form challenging and increasingly complex matrices. Seasonal variation and changes in farming practices, alongside the journey through the transport chain can impact raw ingredients, whilst availability, cost, and changing consumer preferences can see processed food products change in composition over relatively short time frames. In addition, the act of fraud itself cannot be easily predicted and manifests itself in many forms, from adding alum to bread hundreds of years ago to the current common practice of adding syrups to honey. It is generally defined today as “Any actions taken by businesses or individuals that deceive other businesses and/or individuals in terms of misrepresenting food, food ingredients or food packaging that brings about a financial gain1.” The issues are largely the same today as they were hundreds of years ago, but what has changed today is the impact on the consumer; food no longer has borders and passes along complex supply chains, it can last much longer, and the world today consumes more packaged food than ever. A two-tier approach to detecting and confirming food adulteration has been successfully applied in many food commodities. The first tier usually involves a simple and low-cost technique such as molecular spectroscopy, which is used to rapidly screen samples. If a sample is found to be suspect a more complex and time consuming second-tier approach is applied usually involving mass spectrometry2, which can identify the type of fraud and in many cases, measure the contaminants or adulterants present. This thesis aims to focus primarily on the more complex second-tier testing based on mass spectrometry. This will be undertaken by firstly assessing the robustness of existing techniques and workflows, then investigating how methods can be made faster and more robust, showing improvements to the complex data analysis workflows, and concluding by presenting recent and upcoming technologies that have the potential to advance the detection of food fraud with mass spectrometry.

Firstly, a study was undertaken to assess a typical metabolomics workflow for robustness and reproducibility, using a simple pharmaceutical formulation. paracetamol. Paracetamol was chosen as it is subject to rigorous manufacturing controls which ensure consistency of production. The analyte is also highly relevant for food fraud as it is widely used by criminals as an adulterant in various food products, particularly health supplements which have associated claims of treating migraines, allergies, or joint pains.

In this experiment, mass spectrometry was used to distinguish the origin of the manufacture of paracetamol, simulating a common requirement to protect the geographical origin of food. The study aimed to look at the feasibility of detecting small differences between samples assuming minimal variability would be present. Unlike complex food matrices, simple small molecule drug formulations are relatively easy to analyse - as most are produced synthetically, they are tightly regulated in both manufacture and design, so should have minimal sample variance and even when variance is observed it can often be explained from the perspective of detection analysis. The results of the study showed differences in manufacturing origin could be observed as well as differences between over-the-counter paracetamol brands. However, when tested with unknown samples there was very low confidence in class prediction due to lack of statistical power. In addition, the experiment outlined the need for better experimental design and data analysis for forthcoming projects.

The second project made use of the newly released Agilent 6560 Ion Mobility Mass Spectrometer (IM-MS). This instrument was the first commercial instrument to feature a drift tube, allowing measurement of collision cross section (CCS). With any new technology, this was seen as a research tool with very specific application areas and not initially for the analysis of small molecules in complex mixtures. In addition to reviewing the literature and discussing the potential of this technology, many optimisation experiments were performed to improve the performance in the measurement of such small molecules. This work was proven to apply to many other application areas such as the biological fluids often measured in the field of metabolomics and the measurement of lipids in mycobacteria. From learning the potential of this technology further enhancements were made to utilise the ion mobility dimension to enhance, for example, the speed of separation in the chromatographic dimension. By making some changes to an Agilent 1290 liquid chromatography (LC) system, and effectively turning it into two parallel separations, with one running hydrophilic interaction liquid chromatography (HILIC) and the other Reverse Phase, it was possible to double the throughput. This created a system configuration which was ideal for quickly running small pilot scale batches. Hundreds of samples of honey were tested using this methodology and a novel software data processing workflow proposed.

In a third experiment, various issues in the honey industry were examined. Partnering with honey distributors allowed genuine and traceable samples to be collected and issues in claims around various geographies to be understood. For this, more than 1400 samples were examined using a multi-tier testing approach. The first tier utilised Fourier Transform Infra-Red (FTIR), the second tier used gas chromatography mass spectrometry (GC-MS), and the final tier made use of liquid chromatography mass spectrometry (LC-MS). Each technique presented its own strengths, with FTIR being able to detect high levels of syrup adulteration. The detection of syrup adulteration, which is the most common and problematic issue for the industry, was important. The same data was analysed further to review the issues around the geographical origin of honey, such as Manuka where LC-MS was able to distinguish samples harvested in different locations in New Zealand. The work carried out provided a workflow that can easily be adapted for other food commodities as several major obstacles preventing such large studies to be carried out have been addressed such as acquiring reproducible data using short liquid chromatography gradients, maintaining reproducibility with large sample sizes, treatment of data pre and post-statistical analysis and fusing data sets from different instruments and techniques together.

The final chapter looks at how mass spectrometry will fit into the digital lab of the future with particular emphasis on how automation plays a major role in this ecosysten and the current state of software to be able to process data in the cloud. It was important here to highlight the hyphenation of separation techniques such as 2D chromatographic separation and supercritical fluid chromatography (SFC). These techniques, whilst not strictly in the scope of this PhD project, are very important considerations when considering the design of a multi-tier testing approach. The outlook also reviews the contribution analytics and AI have in the processing of data.

Thesis is embargoed until 31 July 2026.
Date of AwardJul 2024
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SponsorsAgilent Technologies, USA
SupervisorNick Birse (Supervisor), Christopher Elliott (Supervisor) & Olivier Chevallier (Supervisor)

Keywords

  • food fraud
  • mass spectrometry
  • metabolomics

Cite this

'