ABSTRACT
The present investigation uses proton transfer reaction mass spectrometry (PTR-MS) combined with multivariate and univariate statistical analyses to study potential biomarkers for altered metabolism in urine due to strenuous walking. Urine samples, in concurrence with breath and blood samples, were taken from 51 participants (23 controls, 11 type-1 diabetes, 17 type-2 diabetes) during the Dutch endurance walking event, the International Four Days Marches. Multivariate analysis allowed for discrimination of before and after exercise for all three groups (control, type-1 and type-2 diabetes) and on three out of 4 days. The analysis highlighted 12 molecular ions contributing to this discrimination. Of these, acetic acid in urine is identified as a significant marker for exercise effects induced by walking; an increase is observed as an effect of walking. Analysis of acetone concentration with univariate tools resulted in different information when compared to breath as a function of exercise, revealing an interesting effect of time over the 4 days. In breath, acetone provides an immediate snapshot of metabolism, whereas urinary acetone will result from longer term diffusion processes, providing a time averaged view of metabolism. The potential to use PTR-MS measurements of urine to monitor exercise effects is exhibited, and may be utilized to monitor subjects in mass participation exercise events.
ABSTRACT
Real-time measurements of many low-abundance volatile organic compounds (VOCs) in breath and air samples are already feasible due to progress in analytical technologies, such as proton transfer reaction mass spectrometry (PTR-MS). Nevertheless, the information content of real-time measurements is not fully exploited, due to the lack of suitable data handling methods. This study develops a data scientific procedure to enhance data analysis and interpretation of longitudinal, multivariate data sets from real-time, in vivo, aroma-release studies. The developed procedure includes an automated data preprocessing and a multivariate assessment of the test panel performance. A large multifactorial PTR-MS data set is investigated that includes four experimental protocols, two tested food products, four aroma compounds, and eight panelists. Real-time measurements are converted into standardized breath profiles by preprocessing, and 10 kinetic parameters are derived. Next to this, panel performance is evaluated per experimental protocol and food product. Comprehensive information about panel performance, individual panelists, studied products, aroma compounds, and kinetic parameters is extracted, demonstrating the great value of the developed approach.
