Related Experiment Video
Updated: Jun 28, 2025
11:23
Purifying the Impure: Sequencing Metagenomes and Metatranscriptomes from Complex Animal-associated Samples
Published on: December 22, 2014
37.3K
Benchmarking bioinformatic virus identification tools using real-world metagenomic data across biomes
Ling-Yi Wu1, Yasas Wijesekara2, Gonçalo J Piedade3,4
1Theoretical Biology and Bioinformatics, Science4Life, Utrecht University, Padualaan 8, Utrecht, 3584 CH, The Netherlands.
Genome Biology
|April 15, 2024
Summary
Choosing the right virus identification tool for metagenomic data is complex. This study benchmarks nine tools, finding PPR-Meta most effective, and suggests parameter adjustments for better virus discovery.
Area of Science:
- Virology
- Bioinformatics
- Computational Biology
Background:
- Metagenomics is the primary method for discovering uncultivated viruses.
- Identifying viruses in metagenomic data is challenging due to numerous available tools.
- Lack of independent benchmarking hinders tool, parameter, and cutoff selection.
Purpose of the Study:
- To independently benchmark state-of-the-art virus identification tools.
- To provide objective guidance for selecting bioinformatics tools for virus discovery.
- To offer suggestions for parameter adjustments in viromics research.
Main Methods:
- Performance comparison of nine virus identification tools across thirteen modes.
- Utilized eight paired viral and microbial datasets from three biomes, including Antarctic coastal waters.
- Evaluated tools based on true positive and false positive rates.
Main Results:
- Tools exhibited highly variable performance, with true positive rates from 0-97% and false positive rates from 0-30%.
- PPR-Meta demonstrated superior performance in distinguishing viral from microbial contigs, followed by DeepVirFinder, VirSorter2, and VIBRANT.
- Most tools identified unique viral contigs, and performance improved with adjusted parameter cutoffs.
Conclusions:
- Independent benchmarking aids researchers in selecting appropriate bioinformatics tools for virus identification.
- Adjusting parameter cutoffs is recommended to enhance tool performance.
- This study facilitates informed choices for viromics researchers.
