Molecular structural discrimination of chemical compounds in body odor using their GC–MS chromatogram and clustering methods

In the present study, body odor samples have been collected according to different sampling protocols and characterized using the gas chromatography (GC)–mass spectrometry (MS) technique with the objective to investigate the existence of different volatile organic compounds (VOCs) belonging to several chemical classes in body odor composition. Moreover, the characterization outcomes have been validated by analyzing spectral information using substantial clustering methods. Specifically, the data matrix based on peak height of chemical compounds in each experiment has been analyzed by using six clustering methods, including principal component analysis (PCA), k-means clustering, fuzzy c-means clustering, hierarchical cluster analysis, fuzzy clustering, and k-medoids clustering. Chemical compounds were well clustered into several groups with each of the implemented clustering methods which endorse the experimental characterization outcomes and establishes the existence of VOCs from multiple chemical classes in body odor composition.