Skip to content

Virus and Phage Identification Tools

Last Updated: November 29, 2025

This section details tools for identifying viral sequences in metagenomic data, including tools for general metagenome analysis, integrated viruses (prophages), and RNA virus identification.

Important: False Positives and Best Practices

Viral identification tools can produce false positives! Cellular genes, mobile genetic elements, and other sequences may be incorrectly classified as viral. To minimize false positives:

  • Use multiple tools: Run at least 2-3 different tools (e.g., VirSorter2 + VIBRANT + geNomad) and take consensus predictions
  • Check confidence scores: Use appropriate score thresholds (see tool-specific guidance below)
  • Validate with CheckV: Always run CheckV to assess genome quality and remove contamination
  • Set minimum contig length: For metagenomes, use ≥5kb; for isolates, ≥1kb minimum
  • Review results critically: Unexpected taxonomic assignments or unusual gene content may indicate false positives

Remember: High sensitivity often comes with lower specificity. Adjust thresholds based on your research goals.

Metagenome Analysis

Tools for identifying viral sequences in mixed metagenomic data:

VirSorter2

VirSorter2 is a multi-classifier, scalable pipeline for identifying viral sequences from diverse environments.

  • Version: v2.2.4, 2023
  • Installation: conda install -c bioconda virsorter=2
  • Language: Python
  • Key Features:
  • Random forest classifiers trained on diverse viral genomes
  • Can detect novel viral sequences with limited homology to reference databases
  • Scalable to large metagenomic datasets
  • Improved sensitivity compared to the original VirSorter

Usage Example: 

virsorter run -w output_dir -i input_contigs.fa --min-length 5000 --min-score 0.5 --include-groups dsDNAphage,ssDNA -j 4 all

Recommended Parameters: - --min-length 5000: Use ≥5kb for metagenomes to reduce false positives - --min-score 0.5: Default threshold (0.5-0.9 for high confidence) - --include-groups: Specify expected viral types to reduce search space - Consider --keep-original-seq to retain boundary information

Score Interpretation: - ≥0.9: High confidence viral sequences - 0.5-0.9: Medium confidence (validate with other tools) - <0.5: Low confidence (likely false positives)

VIBRANT

VIBRANT (Virus Identification By iteRative ANnoTation) identifies viral contigs from metagenomic data.

  • GitHub Stars: ⭐ 159
  • Key Features:
  • Combines boundary detection with annotation
  • Identifies viruses without reliance on sequence similarity
  • Includes metabolic and auxiliary gene annotation
  • Can extract sequences directly from metagenomic assemblies

How It Works: VIBRANT uses multiple neural networks to scan for viral signatures, identifies viral protein boundaries, and performs annotation for identified viral sequences.

geNomad

geNomad (v1.6.0, 2023) is a tool for identifying viral sequences, including integrated proviruses.

  • GitHub Stars: ⭐ 219
  • Key Features:
  • Joint classification of viral and plasmid sequences
  • Annotation of viral genes with protein families
  • Extraction of genomic features
  • Integrates both marker and gene content-based approaches

DeepVirFinder

DeepVirFinder uses deep learning to identify viral sequences in metagenomic data.

  • Key Features:
  • Neural network approach
  • Trained on both prokaryotic and viral genomes
  • Performs well on short contigs
  • Works on novel viruses with limited similarity to references

Integrated Viruses

Tools specifically designed to identify prophages (viruses integrated into bacterial genomes):

PhiSpy

PhiSpy (v4.2.23, 2023) identifies prophages in bacterial genomes using a combination of similarity and composition-based approaches.

  • Installation: pip install phispy or conda install -c bioconda phispy
  • Key Features:
  • Identifies prophages in both complete and draft genomes
  • Uses multiple measurements: GC skew, coding density, strand switching
  • Combines similarity and composition-based features
  • Outputs prophage coordinates and sequences

Usage Example: 

PhiSpy.py -o output_dir -t data/trainers/Generic_taxonomy.pt input_genome.fasta

Phigaro

Phigaro is a tool for prophage prediction that performs well on both complete and draft genomes.

  • Key Features:
  • High-throughput prophage prediction
  • Works with both complete genomes and metagenomic contigs
  • Identifies partial prophages
  • Outputs prophage coordinates and gene annotations

viralintegration

viralintegration is a Nextflow pipeline for detecting viral integration sites in host genomes.

  • Key Features:
  • Identifies viral integration in both germline and somatic contexts
  • Works with both DNA and RNA sequencing data
  • Supports multiple alignment strategies
  • Part of the nf-core collection of pipelines

RNA Virus Identification

Tools specifically for identifying RNA viruses:

palmID

palmID is an RNA virus RdRp (RNA-dependent RNA polymerase) search tool with an R interface.

  • Key Features:
  • Identifies RNA viruses through their RdRp sequences
  • Interactive visualization of phylogenetic placement
  • Integrates with Serratus for large-scale analysis
  • Web interface for easy use

RdRp-scan

RdRp-scan is a tool for searching sequences against the RdRp database to identify RNA viruses.

  • Key Features:
  • Sensitive detection of viral RdRp sequences
  • Works with divergent viral sequences
  • Can be integrated into RNA virus discovery pipelines
  • Command-line tool for high-throughput analysis

metaviralSPAdes-RNA

metaviralSPAdes-RNA is a module of the SPAdes assembler specifically designed for RNA virus detection and assembly.

  • Key Features:
  • Specialized assembly of RNA virus genomes
  • Works with both short and long reads
  • Optimized for metagenomic RNA sequencing data
  • Part of the widely-used SPAdes assembler

Comparison Table

Tool Type Method Best For Installation
VirSorter2 Metagenome Random Forest Large metagenomic datasets conda install -c bioconda virsorter=2
VIBRANT Metagenome Neural Networks Detailed annotation conda install -c bioconda vibrant
geNomad Metagenome Hybrid Both viruses and plasmids conda install -c bioconda genomad
DeepVirFinder Metagenome Deep Learning Short contigs pip install deepvirfinder
PhiSpy Integrated Hybrid Complete bacterial genomes pip install phispy
Phigaro Integrated HMM Draft genomes Custom installation
viralintegration Integrated Alignment Integration sites Nextflow pipeline
palmID RNA RdRp Search RNA virus discovery Web interface
RdRp-scan RNA HMM Divergent RNA viruses git clone
metaviralSPAdes-RNA RNA Assembly RNA virus assembly Part of SPAdes

For comprehensive viral discovery in metagenomic samples, consider using multiple tools with different approaches:

  1. Basic Workflow:
  2. Run VirSorter2 for general viral detection
  3. Check for prophages with PhiSpy

  4. If working with RNA, add RdRp-scan

  5. Advanced Workflow:

  6. Run multiple tools: VirSorter2, VIBRANT, geNomad
  7. Compare results and take consensus predictions
  8. Perform quality assessment with CheckV

  9. Annotate with specialized tools like Pharokka

  10. RNA Virus Workflow:

  11. Assemble with metaviralSPAdes-RNA
  12. Screen for RdRp with palmID or RdRp-scan
  13. Validate with sequence-based methods

Further Reading