BACKGROUND: Confounding factors in the analysis of human breath by thermal desorption GC-MS are reviewed, with special emphasis on the high water levels encountered in human breath samples. RESULTS: Multilinear regression optimization of breath sampling factors, along with the selection of ubiquitous sample components used as retention-time standards, enabled data registration based on retention indexing and mass spectral alignment. This was done on a component-by-component basis. The methodology developed reconciled participant safety, artefacts from accelerated hydrolysis of the stationary phase and the destructive nature of thermal desorption. Furthermore, using ubiquitous methylated cyclic-siloxanes in the thermal desorption-GC-MS chromatograms enabled secondary retention indexing for each chromatogram. This methodology enables the creation of a 'breath matrix' that is based on a combination of retention indexing and the mass spectral registration of isolated peaks. CONCLUSION: This approach facilitated more efficient data modeling and a case study from a 22-participant (10 male, 12 female) stress-intervention experiment. Principal component analysis of data registered by retention indexing did not classify successfully stressed from unstressed states. By contrast, adoption of a breath matrix approach enabled 95% separation.
BACKGROUND: Confounding factors in the analysis of human breath by thermal desorption GC-MS are reviewed, with special emphasis on the high water levels encountered in human breath samples. RESULTS: Multilinear regression optimization of breath sampling factors, along with the selection of ubiquitous sample components used as retention-time standards, enabled data registration based on retention indexing and mass spectral alignment. This was done on a component-by-component basis. The methodology developed reconciled participant safety, artefacts from accelerated hydrolysis of the stationary phase and the destructive nature of thermal desorption. Furthermore, using ubiquitous methylated cyclic-siloxanes in the thermal desorption-GC-MS chromatograms enabled secondary retention indexing for each chromatogram. This methodology enables the creation of a 'breath matrix' that is based on a combination of retention indexing and the mass spectral registration of isolated peaks. CONCLUSION: This approach facilitated more efficient data modeling and a case study from a 22-participant (10 male, 12 female) stress-intervention experiment. Principal component analysis of data registered by retention indexing did not classify successfully stressed from unstressed states. By contrast, adoption of a breath matrix approach enabled 95% separation.