Quant Biol. 2026 Sep 14;14(4):e70055. doi: 10.1002/qub2.70055. eCollection 2026 Dec.
ABSTRACT
16S ribosomal RNA (rRNA) gene sequencing is a standard tool for microbial community analysis. Challenges can occur, particularly in low-biomass samples, when host DNA triggers off-target amplification. Low-biomass microbiome studies are particularly vulnerable to contamination from host DNA, which can obscure microbial signals and bias interpretation. This contamination presents a significant barrier to accurately characterizing microbial communities, especially in clinical or environmental samples with limited bacterial DNA. To systematically quantify and mitigate host DNA interference, we constructed a bacterial mock community dilution series spiked with controlled proportions of human DNA. Using 16S rRNA gene sequencing, we assessed how increasing host DNA affects microbial community profiles and evaluated several computational approaches for removing host-derived sequences, including pre-clustering filtering, post-clustering operational taxonomic unit (OTU) filtering, and the R package Decontam. We found that off-target amplification was more prevalent when the bacterial content was less than 10% relative to host DNA. Total DNA concentration induced minimal bias. Post-clustering OTU filtering and reference genome mapping effectively reduced host contamination. Among the tested correction strategies, post-clustering OTU filtering proved most effective and computationally sustainable, achieving nearly complete removal of host-derived reads with minimal effect on microbial diversity estimates. Although 16S rRNA gene sequencing remains a cost-effective and high-throughput technology, it requires rigorous methodological controls in low-biomass contexts. Our study offers a systematic evaluation of off-target amplification effects and practical mitigation strategies to improve the accuracy of microbial community analysis. The presented framework provides a robust and scalable approach for identifying and removing host contamination from low-biomass 16S rRNA sequencing data.
PMID:42845678 | PMC:PMC13642109 | DOI:10.1002/qub2.70055