Accurate viral genome reconstruction and host assignment with proximity-ligation sequencing

Abstract

Viruses play crucial roles in the ecology of microbial communities, yet they remain relatively understudied in their native environments. Despite many advancements in high-throughput whole-genome sequencing (WGS), sequence assembly, and annotation of viruses, the reconstruction of full-length viral genomes directly from metagenomic sequencing is possible only for the most abundant phages and requires long-read sequencing technologies. Additionally, the prediction of their cellular hosts remains difficult from conventional metagenomic sequencing alone. To address these gaps in the field and to accelerate the study of viruses directly in their native microbiomes, we developed an end-to-end bioinformatics platform for viral genome reconstruction and host attribution from metagenomic data using proximity-ligation sequencing (i.e., Hi-C). We demonstrate the capabilities of the platform by recovering and characterizing the metavirome of a variety of metagenomes, including a fecal microbiome that has also been sequenced with accurate long reads, allowing for the assessment and benchmarking of the new methods. The platform can accurately extract numerous near-complete viral genomes even from highly fragmented short-read assemblies and can reliably predict their cellular hosts with minimal false positives. To our knowledge, this is the first software for performing these tasks. Being significantly cheaper than long-read sequencing of comparable depth, the incorporation of proximity-ligation sequencing in microbiome research shows promise to greatly accelerate future advancements in the field.

Genomic analysis of sewage from 101 countries reveals global landscape of antimicrobial resistance

“…A web server specifically aimed at resolving ARG context based on metagenomic Nanopore reads was also recently published, and could be of help in such cases 45 . For plasmid-borne ARGs, methods involving genomic crosslinking (Hi-C) or single-cell sequencing, have also proven helpful in hosts to genes with higher certainty 46,47 . Indeed, the former has previously been used to show discrete ARG transfer networks in individual patients over merely weeks of antimicrobial therapy 48 .…”

iPHoP: an integrated machine-learning framework to maximize host prediction for metagenome-assembled virus genomes

“…Finally, iPHoP remains limited by host, virus, and host-virus databases, as illustrated by the difference in the number of phages with host prediction between the human microbiome and other ecosystems. Achieving similar performance across all biomes will require in particular expanding the catalog of potential host genomes, with a particular attention paid to CRISPR arrays which are often not fully assembled from metagenomes 28,46 , and expanding the diversity of viruses associated with a host, either from isolation or using in vitro host linkage [47][48][49][50] . In that context, to accommodate future expansions of the tool set and databases, iPHoP was intentionally designed as a modular framework, and we envision the current tool as only the first step towards a comprehensive automated in silico host prediction toolkit.…”

What Lies Beneath? Taking the Plunge into the Murky Waters of Phage Biology

“…For example, combining primary virus enrichment, either by cultivation or by phage adsorption to bacterial cell debris, with virome sequencing has identified novel phages and their hosts ( 64 , 65 ). Another extremely promising method for identifying phage-host pairs is proximity ligation sequencing, where genetic material that is in physical contact can be linked during library preparation ( 66 , 67 ). This incorporation of spatial information with de novo assembly of bacterial and viral genomes has unique potential for assembling virus genomes and establishing host links.…”

“…Using information from Hi-C linkage enables the grouping of viral contigs associated with the same host cell. Performing metagenomic binning within these groups enables reconstruction of genomes with higher overall quality and completeness than conventional workflows ( 66 , 67 ). Moreover, it captures the interaction of phages with their respective host and thus is a powerful tool for discovering phage-microbe pairs and evaluating the phage host range within a community.…”

“…Proximity ligation sequencing has the potential to transform what we know about phage host specificity. Currently, there are rarely more than hundreds of bacteria included in a typical phage host range study; however, with proximity ligation sequencing, it is possible to assess the phage host range within an entire complex and even uncultivatable sample ( 67 ). For example, Marbouty et al showed that 17 members of the crAss-like phage infect different strains of Bacteroidetes ( 68 ).…”