MVP: a modular viromics pipeline to identify, filter, cluster, annotate, and bin viruses from metagenomes

Clément Coclet, Antônio Pedro Camargo, Simon RouxmSystems 2024 · DOI

TL;DR

This paper introduces MVP (Modular Viromics Pipeline), a user-friendly, Python-based workflow designed to streamline the complex analysis of viral genomes from metagenomic data. MVP integrates cutting-edge tools to identify, filter, cluster, annotate, and bin viruses, making advanced viromics accessible to a wider range of researchers and enabling more comprehensive studies of viral communities.

What Problem Does This Paper Address?

The study of viruses, particularly those that cannot be cultured in a lab (uncultivated viruses), has been revolutionized by high-throughput sequencing. However, analyzing the vast amounts of data generated from metagenomes, viromes, and metatranscriptomes to extract meaningful viral information is a significant challenge. Existing bioinformatics tools are numerous and constantly evolving, making it difficult for researchers, especially those with limited bioinformatics experience, to select, integrate, and effectively use them. This complexity hinders the ability to reliably identify viral genomes, assess their quality, determine their taxonomic and functional roles, and understand their abundance and distribution within ecosystems. The paper addresses this gap by providing a standardized, integrated, and user-friendly pipeline that consolidates state-of-the-art viromics analysis steps.

Methods

MVP is a modular pipeline built in Python, designed for ease of installation and use. It comprises ten modules, each performing a specific analytical task. Key modules and their functionalities include:

  • Module 01 (geNomad & CheckV): This module is central to viral genome recovery. It uses geNomad for initial identification of viral and proviral sequences from assembled metagenomes and then employs CheckV to assess the quality and completeness of these predicted viral genomes. MVP includes a smart re-analysis step for proviruses that might still contain host DNA, ensuring accurate assessment.
  • Module 02 (Filtering): This module applies conservative post-classification filters (e.g., minimum viral score, genome length, presence of hallmark viral genes) to remove low-confidence viral predictions. Users can opt for a more relaxed filtration mode if needed.
  • Module 03 (Clustering): This module performs vOTU (viral operational taxonomic unit) clustering using a high average nucleotide identity (ANI) threshold (≥95%) to dereplicate viral sequences and identify distinct viral populations.
  • Module 04 (Read Mapping): This module maps sequencing reads back to the identified viral sequences to estimate their abundance and coverage across samples.
  • Module 05 (vOTU Tables): This module generates coverage tables summarizing viral abundance and features, designed for downstream ecological analyses.
  • Module 06 (Functional Annotation): This module annotates viral proteins using various databases like PHROGS, Pfam, and optionally anti-defense system (ADS) and RdRP databases, providing insights into viral functions. It also prepares input files for tools like DRAM-v for potential auxiliary metabolic gene (AMG) identification.
  • Module 07 (Binning): An optional module that uses vRhyme to perform viral genome binning, grouping viral contigs into potential viral genomes.
  • Module 99 (GenBank Submission): An optional module to assist users in preparing metadata for submitting viral genomes to public databases like NCBI GenBank.
  • Module 100 (Summarized Outputs): This final module generates a comprehensive summary report of all commands used, running times, and key results, along with overview figures.

MVP is designed to be a dynamic wrapper, allowing for the continuous integration of new tools and databases as the field evolves.

Datasets

The paper demonstrates MVP's utility by applying it to a dataset of 20 metagenome samples collected from wetland sediments in the Loxahatchee Nature Preserve, Florida. These samples had been previously processed using the JGI Metagenome Workflow. This dataset served as a benchmark to showcase MVP's capabilities in identifying viral contigs and vOTUs, assessing their quality, and characterizing their diversity and functional potential. No new benchmark datasets were introduced, but the application to a well-characterized dataset allowed for comparison and validation of MVP's performance.

Key Insights and Ideas

  • Democratizing Viromics: MVP significantly lowers the barrier to entry for viromics research by providing a user-friendly, integrated workflow that handles the complexities of tool selection and integration.
  • Standardization and Reproducibility: By offering a standardized pipeline, MVP ensures that analyses are reproducible and comparable across different studies and research groups.
  • Comprehensive Analysis: MVP goes beyond simple identification, offering integrated steps for quality assessment, clustering, functional annotation, and even binning, providing a holistic view of viral communities.
  • Adaptability: The modular design and dynamic wrapper nature of MVP allow it to stay relevant in the rapidly advancing field of viromics, easily incorporating new tools and databases.
  • Ecological Relevance: The output of MVP is designed to be directly usable for downstream ecological analyses, facilitating a deeper understanding of viral roles in various ecosystems.

Key Results

When applied to the 20 wetland sediment metagenomes, MVP successfully identified:

  • 11,656 viral contigs after relaxed filtration.
  • 8,145 viral operational taxonomic units (vOTUs) after clustering (ANI ≥ 95%; AF ≥ 85%).
  • A clear beta-diversity pattern was observed across different locations within the Loxahatchee Nature Preserve, indicating distinct viral community compositions.
  • The majority of identified vOTUs belonged to the Caudoviricetes class (94.7%), a common group of tailed bacteriophages.
  • Functional annotation revealed that a significant portion of viral genes were related to DNA, RNA, and nucleotide metabolism, as well as integration and excision mechanisms.
  • The pipeline successfully processed the data, generating comprehensive summary reports and figures that facilitated interpretation.
  • A comparison with the ViWrap pipeline indicated that MVP was faster to run for the same dataset and provided filtered outputs that were directly usable, whereas ViWrap's unfiltered outputs could be biased.

Limitations and Future Work

  • Pre-processing Not Included: MVP does not include modules for raw read pre-processing (quality control) or assembly. Users are expected to provide these inputs.
  • Host Prediction: While MVP can prepare input files for host prediction tools like iPHoP, it does not natively perform host prediction itself.
  • Future Development: The authors plan to continuously update MVP by integrating new tools and databases, such as vConTACT3 for taxonomic classification and iPHoP for host prediction. They also aim to collect user feedback to guide future development and ensure MVP remains a state-of-the-art tool. The modular design is specifically intended to facilitate these future integrations.

Summary updated

Abstract

While numerous computational frameworks and workflows are available for recovering prokaryote and eukaryote genomes from metagenome data, only a limited number of pipelines are designed specifically for viromics analysis. With many viromics tools developed in the last few years alone, it can be challenging for scientists with limited bioinformatics experience to easily recover, evaluate quality, annotate genes, dereplicate, assign taxonomy, and calculate relative abundance and coverage of viral genomes using state-of-the-art methods and standards. Here, we describe Modular Viromics Pipeline (MVP) v.1.0, a user-friendly pipeline written in Python and providing a simple framework to perform standard viromics analyses. MVP combines multiple tools to enable viral genome identification, characterization of genome quality, filtering, clustering, taxonomic and functional annotation, genome binning, and comprehensive summaries of results that can be used for downstream ecological analyses. Overall, MVP provides a standardized and reproducible pipeline for both extensive and robust characterization of viruses from large-scale sequencing data including metagenomes, metatranscriptomes, viromes, and isolate genomes. As a typical use case, we show how the entire MVP pipeline can be applied to a set of 20 metagenomes from wetland sediments using only 10 modules executed via command lines, leading to the identification of 11,656 viral contigs and 8,145 viral operational taxonomic units (vOTUs) displaying a clear beta-diversity pattern. Further, acting as a dynamic wrapper, MVP is designed to continuously incorporate updates and integrate new tools, ensuring its ongoing relevance in the rapidly evolving field of viromics. MVP is available at https://gitlab.com/ccoclet/mvp and as versioned packages in PyPi and Conda.IMPORTANCEThe significance of our work lies in the development of Modular Viromics Pipeline (MVP), an integrated and user-friendly pipeline tailored exclusively for viromics analyses. MVP stands out due to its modular design, which ensures easy installation, execution, and integration of new tools and databases. By combining state-of-the-art tools such as geNomad and CheckV, MVP provides high-quality viral genome recovery and taxonomy and host assignment, and functional annotation, addressing the limitations of existing pipelines. MVP's ability to handle diverse sample types, including environmental, human microbiome, and plant-associated samples, makes it a versatile tool for the broader microbiome research community. By standardizing the analysis process and providing easily interpretable results, MVP enables researchers to perform comprehensive studies of viral communities, significantly advancing our understanding of viral ecology and its impact on various ecosystems.
BibTeX
@inproceedings{coclet2024mvp,
  author = {Clément Coclet and Antônio Pedro Camargo and Simon Roux},
  title = {{MVP: a modular viromics pipeline to identify, filter, cluster, annotate, and bin viruses from metagenomes}},
  booktitle = {mSystems},
  year = {2024},
  doi = {10.1128/msystems.00888-24},
}
Figure