Sex Differentiation from Human Scent Chemical Analysis

Abstract

Human scent traces are often the only pieces of evidence providing information about individuals that were present at a crime scene. In this study, the possibility of sex differentiation using detailed chemical analyses of human scent samples for forensic purposes is discussed. The human scent samples were analyzed through the use of headspace-gas chromatography/mass spectrometry (HS-GC/MS). The results of these chemical analyses were evaluated using several data processing approaches (Linear Support Vector Machine, Quadratic Discriminant Analysis, and Ridge Regression), which were applied to distinguish between sexes from the human scent samples obtained from the palms of six volunteers for twelve weeks. This study indicates that sex differentiation based on the chemical analysis of human scent samples using HS-GC/MS is possible. The best results were obtained using the Ridge Regression with thresholding providing accuracy and a critical sensitivity of the sex differentiation of better than 91%.