virus ID workflow in metagenomic NGS

[SnipSniper]

What methods does everyone employ to identify viruses in metagenomic NGS data intrigues me. Do you usually classify the remaining reads after removing the host reads, or do you compare all of the reads directly to viral databases? Or would you rather compare contigs with viral databases after de novo assembly? I'd be interested in knowing what has, in your experience, worked best.

 

[BookwormBro]

@SnipSniper Most often, after host read removal, I'll map the remaining reads directly to viral databases (like RefSeq Viral, GenBank Viral). For deeper dives or novel viruses, de novo assembly of those nonhost reads, then comparing contigs to viral databases, is usually the gold standard.

 

[TissueCrafter]

Cool, mapping to viral databases after taking out the host stuff is smart for quick results, and building from scratch is the way to find new viruses.

 

[Fantastic Mr. Fox]

What methods does everyone employ to identify viruses in metagenomic NGS data intrigues me. Do you usually classify the remaining reads after removing the host reads, or do you compare all of the reads directly to viral databases? Or would you rather compare contigs with viral databases after de novo assembly? I'd be interested in knowing what has, in your experience, worked best.

I usually go the contig route after de novo assembly if I'm hunting something novel. But for routine stuff, mapping non-host reads straight to a curated viral DB works just fine. Sometimes you learn from all the errors so if in case some things happen to you, hold on you'll be fine.