Raman spectroscopy and Advanced Multivariate Analysis for the Identification and Age Estimation of Body Fluid Traces on Interfering Substrates for Forensic Purposes

Lecture Center 3C

In forensics, identifying body fluid evidence is a large priority. Body fluids, such as blood, semen, and sweat, can provide DNA for profiling and insight onto the nature of the crime committed. Our laboratory has developed a methodology that can universally detect and identify body fluid stains using Raman spectroscopy and statistics. This methodology was created under laboratory conditions, which are a rarity out in the field. Body fluid evidence can be found on any number of substrates which yield large Raman signals that impede detection and identification. This project proposes a new program called HAMAND (Hypothetical Addition Multivariate Analysis with Numerical Differentiation) as a solution to this complication. The software uses a reference spectrum of the expected body fluid to search for any reference components in the sample spectrum and then uses these to create a new spectrum belonging to just the analyte. This proof of concept experiment focused on testing blood on denim, which is a common substrate in crime scenes and produces a large Raman signal. In combination with HAMAND, we were able to produce high quality spectra that were highly resembling of average blood spectra collected on lab-grade substrates. When run through identification models, these spectra were all classified correctly as blood, whereas their original spectra were not. This shows that HAMAND has great potential in rectifying the current obstacles in body fluid identification for forensic purposes.