Analysis of volatile metabolites are of increasing interest to clinical and laboratory phenotyping for a host of respiratory diseases. For respiratory diseases caused or exacerbated by microbes, headspace analysis of in vitro cultures is particularly useful for elucidation of extracellular metabolism. In this study, a headspace sampling method was developed and tested by cultivating A. fumigatus in vitro. The headspace profile revealed the presence of known fungal-specific volatiles, for example pyrazines and terpenoids. Furthermore, volatiles (e.g. monoterpenes) were found to be associated with fungal germination and early sporulation. However, measurement of volatile metabolites is prone to high temporal and inter-sample variation, especially within complex matrices. Therefore, a strong emphasis on analytical quality assurance is required. To this end, a 24-component in-house mixture of volatile chemical standards was developed, specifically for analysis with breath samples of patients with lung infection. The mixture was found to be suitable for analytical batch correction (majority of components had an RSD < 10 %), where nonane was used to normalise samples due to its stability (RSD 2 %). In addition, the mixture was used to confirm and quantify VOC identification, and for comparing intra- (r = 0.87) and inter-instrument (r = 0.77) similarity using Procrustes analysis. Large clinical studies are particularly susceptible to high variation due to the influence of additional patient cohorts and sampling sites. To investigate this problem, exhaled breath VOC data from the U-BIOPRED study were evaluated, where results showed significant differences in pentane (p = 0.027) and dodecane (p < 0.001) intensities between sites. Additionally, batch correction and instrument similarity were assessed using the Bhattacharya distance and Procrustes superimposition, respectively. As a result, several recommendations were made for future research, specifically to assess variation within large-scale and multi-site studies. Finally, a VOC gas sensor based on micro-cantilever technology was developed and assessed using acetone, ethanol, octane, and water vapour. The acquired resonant frequency response correlated with external temperature and pressure fluctuations, and therefore systematic drift correction was necessary. Measurement of VOC vapour mixtures indicated discrimination of VOCs with micro-cantilevers coated with three different polymers. Furthermore, the sensor chips showed inter-measurement reproducibility and over a period of six months (p = 0.553), however not between duplicate micro-cantilevers on the same chip (p = 0.0039). The work presented in this thesis addresses several aspects in the sampling and analysis of volatile metabolites. The research performed also demonstrates the potential for volatile metabolites to serve as markers of several respiratory diseases, as an alternative route to clinical and laboratory disease phenotyping, towards improved patient outcomes.
|Date of Award
|1 Aug 2019
- The University of Manchester
|Royston Goodacre (Supervisor) & Stephen Fowler (Supervisor)