Full-length 16S rRNA-based exploration of body site-specific bacterial signatures for origin determination and individual identification

Abstract

When the available human-derived information at a crime scene is limited, it poses challenges in determining the origin of the biological materials and identifying their donors. In this context, microorganisms have gradually emerged as a valuable complementary tool. Nowadays, the application of third-generation sequencing technology for full-length 16S rRNA sequencing to explore the specific bacterial biomarkers in various biological materials holds significant research and practical value. In this study, we performed full-length 16S rRNA gene sequencing on sterile swabs from palmar skin, oral mucosa, and nasal cavity using the PacBio single-molecule real-time sequencing (SMRT) platform. Alongside identifying specific bacterial biomarkers for these biological materials from different body sites, the study also preliminarily explored the specific bacterial taxa in 19 individuals at the phylum, genus, and species levels. The results showed that the palmar skin bacteria primarily consist of Cutibacterium, Staphylococcus, and Streptococcus, the oral mucosal bacteria are dominated by Streptococcus, Neisseria, and Haemophilus, while the dominant bacteria in nasal cavity are Staphylococcus and Cutibacterium. Beta diversity analysis revealed significant differences in the bacterial community composition across the three origins of biological materials. Furthermore, classification models based on the bacterial species were constructed using the Random Forest, XGBoost, and KNN algorithms. The results showed that both Random Forest and XGBoost models achieved an accuracy of 97 %, significantly outperforming the KNN model (79 %). The prediction accuracy at the OTU level was comparable to that at the species level. In addition, bacterial community differences between individuals were observed at both the genus and species levels. Overall, this study further explores the potential of classification prediction methods based on bacterial features for distinguishing the body site origins of different biological materials and enabling individual traceability, thereby providing valuable data to support the application of microbiological techniques in forensic practice.