Discrimination of smokers and nonsmokers based on the analysis of fingermarks for forensic purposes

For decades, the unique characteristic ridge pattern of a fingerprint (FP) has been used as a mainstay for human individualization. Most recently, vibrational spectroscopic analysis of fingermark (FM) has attracted a substantial amount of attention in providing intelligence on personal traits. Particularly, extracting phenotypical information such as age and sex can be significantly beneficial in criminal investigation. However, chemical profiling of FMs using vibrational spectroscopy has been limited to age and sex determination, while other significant traits can also aid in narrowing down the suspect pool. In the current proof-of-concept study, attenuated total reflection Fourier transform-infrared (ATR-FTIR) spectroscopy combined with partial least squares-discriminant analysis (PLS-DA) have been employed for determining smoker from non-smoker donors from their FM residues. Genetic algorithm (GA) was applied to improve the discrimination rate of the developed model by selecting regions that are significant for the distinguish between the two classes. A subject wise leave-one-out cross-validation (LOOCV) was initially used to evaluate the performance of the binary PLSDA classifier for each donor. The binary model showed 84% correct classification at spectral level and 92% correct classification at donor level in subject wise LOOCV. In addition, receiver operating characteristic (ROC) curve analysis was constructed to establish a threshold for differentiating between the smoker and non-smoker FMs, this resulted in 100% accuracy at a donor level for external validation test. This preliminary study shows a great promise for identifying smoker donors from non-smoker individuals based on chemical analysis of FM residues. After fully developed, we believe the method will offer significant potential for real crime scene investigation due to its simplicity, non-destructive nature and its prospective for in-field analysis.