Meta-virus resource (MetaVR): expanding the frontiers of viral diversity with 24 million uncultivated virus genomes

Mateus Bernabe Fiamenghi, Antônio Pedro Camargo, Iro N Chasapi, Fotis A. Baltoumas, Simon Roux, Artyom A. Egorov, Eleni Aplakidou, Éric Olo Ndela, Yumary M. Vasquez, I-Min A. Chen, Krishna Palaniappan, T. B. K. Reddy, Supratim Mukherjee, Natalia Ivanova, Frederik Schulz, Tanja Woyke, Emiley A. Eloe‐Fadrosh, Georgios A. Pavlopoulos, Nikos C. KyrpidesNucleic 2025 · DOI

TL;DR

MetaVR is a massive new database that dramatically expands our understanding of viral diversity by housing over 24 million uncultivated virus genomes (UViGs). It addresses limitations of previous resources by offering enhanced querying, programmatic access via an API, and integrated protein structures, making it a powerful tool for exploring the global virosphere.

What Problem Does This Paper Address?

Viruses are everywhere, playing critical roles in ecosystems, host health, and evolution. However, the vast majority of viruses remain uncultivated, making their study challenging. While genomic and metagenomic sequencing has led to an explosion of "uncultivated virus genomes" (UViGs), these discoveries are scattered across various databases that often lack comprehensive integration, large-scale querying capabilities, and programmatic access. This fragmentation hinders comparative studies and a holistic understanding of viral diversity, evolution, and host interactions. The paper introduces MetaVR as the successor to the IMG/VR database, aiming to overcome these limitations by providing a unified, expanded, and more accessible resource.

Methods

MetaVR builds upon the foundation of IMG/VR but incorporates several key methodological advancements:

  • Data Ingestion and Viral Prediction: The team mined a vast number of public metagenomic and metatranscriptomic datasets, along with isolate genomes, SAGs, and MAGs. Viral sequences were identified using geNomad, a tool that leverages score calibration for accurate false discovery rate estimation. CheckV was then used to refine predictions and assess quality.
  • UViG Quality and Confidence Assessment: A refined "viral confidence" field was introduced, distinguishing between sequences with strong orthogonal evidence of being viral and those with less definitive signatures. This provides users with a nuanced understanding of the data's reliability.
  • Viral Operational Taxonomic Units (vOTUs): UViGs were clustered into vOTUs based on average nucleotide identity (ANI) and alignment fraction (AF), creating a standardized unit for comparative analysis.
  • Taxonomic and Host Assignment: Taxonomic assignments were updated to the latest ICTV release, and a novel computational tool, iPHoP, was integrated for more robust host prediction, including eukaryotic hosts.
  • Protein Clustering and Structure Prediction: A significant innovation is the clustering of all predicted viral proteins and the subsequent prediction of 3D structures for representative sequences using AlphaFold3. This enables functional inference and comparative structural biology.
  • Database Implementation: MetaVR is built on a robust web framework with a MariaDB backend, offering a user-friendly interface and an API for programmatic access.

Datasets

The primary datasets used for MetaVR's construction include:

  • IMG/M: A large compendium of metagenomic and metatranscriptomic datasets.
  • Public Assemblies: Including bacterial and archaeal isolate genomes, single amplified genomes (SAGs), and metagenome-assembled genomes (MAGs).
  • External Studies: Data from resources like RVMT, Inoviruses, RefSeq, and ViroidDB were incorporated.
  • Giant Virus MAGs (GV-MAGs): A significant addition to expand the representation of giant viruses.
  • Human Gut Microbiome Samples (UHGV): Further enriching the human virome data.

No new benchmarks were introduced, but MetaVR significantly expands the scale of existing datasets.

Key Insights and Ideas

  • The Virosphere is Vastly Underexplored: The sheer number of UViGs (over 24 million) highlights how much viral diversity remains to be discovered, even with advanced sequencing technologies.
  • Integration is Key: MetaVR's success lies in its comprehensive integration of diverse data types (genomes, taxonomy, hosts, protein structures) and its provision of a unified access point, overcoming the fragmentation of previous resources.
  • Protein Structures Unlock Functional Insights: The inclusion of predicted protein structures is a game-changer, allowing researchers to infer function, identify evolutionary relationships, and explore conserved elements even when sequence similarity is low. This opens new avenues for understanding viral biology.
  • Host-Virus Interactions are Complex and Widespread: The enhanced host assignment capabilities, particularly with the integration of eukaryotic host data and the iPHoP tool, underscore the intricate relationships between viruses and their hosts across diverse environments.
  • Confidence Matters: The explicit distinction between "high-confidence" and "low-confidence" viral sequences empowers users to critically assess the data and tailor their analyses accordingly.

Key Results

  • Massive Expansion of Viral Diversity: MetaVR contains 24,435,662 UViGs, a 57.6% increase over its predecessor, IMG/VR v4.
  • Over 12 Million Viral Operational Taxonomic Units (vOTUs): This provides a standardized unit for comparative analysis.
  • Extensive Protein Data: The database includes over 42 million protein clusters and predicted structures for a significant portion of these clusters, making it the largest repository of predicted virus protein structures.
  • Improved Host Assignment: Putative hosts have been assigned to a substantial number of UViGs, including previously unassigned ones.
  • Programmatic Access: The availability of an API allows for seamless integration into automated analysis pipelines.
  • Geographic Representation: While global, the data shows a bias towards samples from North America and Europe, highlighting areas for future sequencing efforts.

Limitations and Future Work

  • Unsaturated Biodiversity: Despite the massive dataset, the authors acknowledge that viral biodiversity is still far from saturated, indicating that many more viruses await discovery.
  • Computational Prediction Uncertainty: The reliance on computational predictions means there's an inherent risk of false positives. The authors emphasize the importance of using confidence and quality metrics to mitigate this.
  • Host Assignment Challenges: While improved, host assignment remains a challenge, with estimated false discovery rates of ~5-10% for in silico predictions.
  • Geographic Gaps: The current dataset shows a bias towards certain regions, suggesting a need for increased sequencing from underrepresented areas like South America and Africa.

Future work will likely involve continued expansion of the database with new sequencing data, refinement of prediction algorithms, and further development of tools for analyzing viral diversity, evolution, and host interactions. The integration of protein structures and enhanced host assignment are key steps towards a more comprehensive understanding of the global virome.

Summary updated

Abstract

Viruses are ubiquitous in all environments and impact host metabolism, evolution, and ecology, although our knowledge of their biodiversity is still extremely limited. Viral diversity from genomic and metagenomic datasets has led to an explosion of uncultivated virus genomes (UViGs) and the development of specialized databases to catalog this viral diversity, though many lack comprehensive integration. Here, we introduce meta-virus resource (MetaVR), the successor of the IMG/VR database, designed to overcome previous limitations such as large-scale querying and programmatic access. Drawing on the increase of publicly available genomes and metagenomes, MetaVR significantly expands viral diversity, now comprising 24,435,662 UViGs, a 57.6% increase from its predecessor, organized into over 12 million viral operational taxonomic units. Key enhancements include the integration of curated eukaryotic host information, the integration of protein clusters and predicted structures for comparative studies, and an API for programmatic data access. Furthermore, MetaVR features an updated taxonomic framework based on ICTV release 39, assignment to Baltimore classes, and enhanced host assignment through novel computational tools like iPHoP. These advancements position MetaVR as a unique resource for exploring viral diversity, evolution, and host interactions across diverse environments. MetaVR can be freely accessed at https://www.meta-virome.org/.
BibTeX
@inproceedings{fiamenghi2025meta,
  author = {Mateus Bernabe Fiamenghi and Antônio Pedro Camargo and Iro N Chasapi and Fotis A. Baltoumas and Simon Roux and Artyom A. Egorov and Eleni Aplakidou and Éric Olo Ndela and Yumary M. Vasquez and I-Min A. Chen and Krishna Palaniappan and T. B. K. Reddy and Supratim Mukherjee and Natalia Ivanova and Frederik Schulz and Tanja Woyke and Emiley A. Eloe‐Fadrosh and Georgios A. Pavlopoulos and Nikos C. Kyrpides},
  title = {{Meta-virus resource (MetaVR): expanding the frontiers of viral diversity with 24 million uncultivated virus genomes}},
  booktitle = {Nucleic},
  year = {2025},
  doi = {10.1093/nar/gkaf1283},
}
Figure