Evaluation of targeted Massively Parallel Sequencing methods for forensic metagenomics.

Massively Parallel Sequencing (MPS) is effective for monitoring the microbial composition of environmental samples. Soil microbial signatures are critical for pinpointing the geographic location of forensic evidence, but standard 16S rRNA methods lack species-level resolution. Targeted sequencing panels, consisting of informative DNA fragments, can overcome this shortcoming and are highly desirable for forensic investigations. To address this, we evaluated three target enrichment methods for metagenomic analysis. First, we used Whole Metagenome Sequencing (WMS) data from 134 soil samples across 46 locations in Poland to extract a set of 200 markers. Using these markers, we created prototype targeted sequencing panels to compare two amplicon capture-based methods (Thermo Fisher AmpliSeq™ and Integrated DNA Technologies xGen™) and one hybridization capture-based method (Roche KAPA HyperPlus). The comparison of the technologies was guided by the results of classification of sample origin by machine learning classifier trained on feature profiles from WMS. The methods were assessed on technical parameters including data quality, reproducibility, sensitivity, and practical implementation for forensic laboratories. The performance and precision varied depending on technology and DNA concentration. The Roche KAPA HyperPlus hybridization capture-based method consistently demonstrated superior performance. Across various DNA input quantities, it showed the highest correlation with WMS data and achieved an exceptional F1 score of 0.94 at 5 ng, significantly outperforming the amplicon-based methods. This indicates that hybridization capture is a more robust and accurate approach for forensic soil microbiome profiling, particularly for low-template evidence, providing a highly reliable tool for predicting geographic origin. KEY POINTS: • Targeted Massively Parallel Sequencing methods for forensic soil microbial analysis • Targeted sequencing allowed the determination of the place of origin of soil samples • Roche KAPA HyperPlus: the most accurate classification of the soil samples origin.