Microbial and seminal traces of sexual intercourse and forensic implications

Predicting recent intercourse is crucial in forensic casework on sexual assaults. Justice is no longer only interested in matching a trace to its cellular origin and donor (source level) but also in which actions have led to the deposition of a trace (activity level). In this work, we assessed whether sexual intercourse can be predicted based on the vaginal microbiome and compared these predictions to the gold standard method of semen detection. First, we trained a predictive model on our large Isala dataset with microbial community profiles of 3,043 women. On this dataset, intercourse was predicted with 71% accuracy in a balanced cross-validation machine learning setting. Next, we evaluated our prediction model in a newly conducted longitudinal intervention study of 10 women who were requested to have sexual intercourse on fixed days (GeneDoe project). Vaginal community profiles were generated, on which our predictor could accurately establish intercourse in 82%. We then compared these microbiome-based predictions to prostate-specific antigen-based semen detection, which showed an accuracy of 84%. In addition, we assessed whether semen detection in underwear compared to vaginal samples could be used as a less invasive alternative. The prediction accuracy of intercourse by detecting semen on underwear was 95%. This greater evidential value of underwear than vaginal swabs for intercourse prediction was confirmed through a retrospective analysis of 207 forensic sexual assault cases. Taken together, this study revealed the potential of both microbiome profiling on vaginal swabs and semen detection on underwear for forensic casework. In addition, our study provides ecological insights into microbiome dispersal across the vagina, underwear and semen, opening new perspectives on alternative approaches in the microbiome field to address women's health and safety.