viralFlye: assembling viruses and identifying their hosts from long-read metagenomics data

Area of Science:

• Metagenomics
• Virology
• Bioinformatics

Background:

• Assembling viral genomes from complex metagenomic data is challenging.
• Short-read sequencing methods often yield fragmented viral genomes.
• Long-read sequencing offers improved contiguity but requires specialized tools.

Purpose of the Study:

• To introduce viralFlye, a tool for analyzing metagenome-assembled viruses from long-read assemblies.
• To evaluate the performance of viralFlye in improving viral genome assembly.
• To compare virus-host association predictions from long-read versus short-read data.

Main Methods:

• Development and application of the viralFlye tool.
• Analysis of long-read and short-read metagenomic data.
• Identification of CRISPR arrays for virus-host association prediction.

Main Results:

• viralFlye significantly improves the assembly of viral genomes from long-read data.
• Long-read assemblies reveal a substantially larger number of potential virus-host associations compared to short-read assemblies.
• Novel CRISPR arrays were identified, providing insights into new virus-host relationships.

Conclusions:

• Long-read sequencing, coupled with tools like viralFlye, is superior for assembling complex viral communities.
• This approach expands the discovery of viral diversity and their ecological interactions.
• CRISPR array analysis is a valuable method for predicting novel virus-host associations.