Recovery and host linking of MGEs in microbiomes

Mobile genetic elements (MGEs) are major drivers of horizontal gene transfer, including the spread of antimicrobial resistance (AMR) genes. However, determining the host of an MGE in complex natural microbiomes remains challenging. Here, we spike a Bacillus velezensis strain carrying an extrachromosomal phage and plasmid into a batch bioreactor simulating the human gut, and use it as a spike-in control to assess the performance of Hi-C sequencing and Oxford Nanopore Technologies (ONT)-enabled DNA methylation detection to identify MGE-host pairs. To improve recovery of low-abundance genomes, we used a novel ONT adaptive sampling (AS) strategy that depletes de-novo assembled high-abundance contigs, rather than relying on reference genomes. This approach led to ~2-fold enrichment of low-abundance replicons, including the spike-in strain. Methylation-based host assignment failed for the B. velezensis MGEs, likely due to absence of DNA-methylation. In contrast, Hi-C successfully linked the phage to its host, but not the plasmid, likely due to non-intact cells, and only after careful bioinformatic processing to remove artefactual signals. For a native E. coli strain, the Hi-C and methylation data indicated a link to two plasmids, which could be confirmed by selective isolation and genome sequencing. Our results highlight that Hi-C and methylation data can provide powerful insights into MGE–host associations, but their interpretation requires careful computational analysis and biological validation. Moreover, our AS strategy offers a cost-efficient method to boost coverage of low-abundance genomes, improving metagenomic investigation of MGEs in complex natural microbiomes.