Best classifier for viral reads--PhyloSift? BLAST? Others?

[clintp]

I'm trying to compare methods of classifying RNA and DNA virus reads from metagenomic sequeincing data. I've got MetAMOS set up, and I'm mainly interested in the annotation steps right now. So far, out of Kraken, PhyloSift, FCP, PhymmBL, and BLAST, only PhyloSift and BLAST give somewhat correct results (I'm running a sample known to contain lots of orthomyxovirus cDNA). BLAST annotation is fine but slow. The trouble with PhyloSift is that for the majority of its virus calls, when I go back to find the contig that resulted in a virus call, none of it will BLAST to what it should. Not even with very loose parameters.

I'm probably looking at this the wrong way--PhyloSift works differently than I'm used to I know. If it's as simple as that, how can I go about verifying hits (or at least become confident) that result from PhyloSift?

Alternatively, should I be trying a different classifier? The others I mention seem to default in MetAMOS to Bacteria/Archea classification, or DNA viruses only. Any advice here?

Thanks

[guliseedol]

Hi,

I know it's been a year since you posted this, so i am wondering if you found some good classification tools for viruses, and how did phyloshift work out for you.
Also i am wondering if you are looking for novel viruses how would you go about in finding them properly, basic assmebly and then mapping?

Cheers