A bioinformatics framework using public 16S rRNA gene amplicon data to assess the presence of target bacteria in bat and rodent samples.
TL;DR
A dual-strategy bioinformatics pipeline that leverages publicly available 16S rRNA gene amplicon sequencing data to reliably and inexpensively confirm target bacterial presence and distinguished target-positive from negative samples, with phylogenetic support for specificity is described.
Abstract
Validating the ecological distribution of a newly isolated bacterial species in natural hosts remains challenging due to the lack of specific detection assays and the cost of large-scale screening. Here, we describe a dual-strategy bioinformatics pipeline that leverages publicly available 16S rRNA gene amplicon sequencing data to reliably and inexpensively confirm target bacterial presence. The method first extracts hypervariable regions from the target bacterium's full-length 16S rRNA gene and evaluates their specificity by calculating an A-value-defined as the highest sequence similarity to any non-target strain in reference databases. Regions with an A-value below the 98.7% species threshold are selected. These are then aligned against Amplicon Sequence Variants (ASVs) from public datasets to compute a B-value (highest similarity to ASVs within a sample). A novel classification logic (B > A) is applied to designate samples as positive or negative, reducing false positives. The pipeline incorporates multi-level controls, including process/biological negatives and positives. Testing with novel species (Clostridium sp. nov.) and a formally described species (Streptococcus lishijunsis), along with common commensal species demonstrated that region-specific performance varies, highlighting the need for pre-validation. The framework successfully distinguished target-positive from negative samples, with phylogenetic support for specificity. This approach provides a rigorous, cost-effective, and accessible workflow that links in vitro isolation to in vivo ecological validation using existing public data.