BLAST comparison of 16S rRNA against the SILVA database

[Nilaksha]

I'm currently running metagenomics pipeline through galaxy, and at the taxonomic mapping step I use ncbi blast+ wrappers to blast against a locally installed nt database. But ncbi nt database contains many data which are not totally relevant to metagenomics , so I want to locally install a more curated database like Silva or greengenes. But is it possible to do a megablast using blast+ wrappers against the silva or greengenes database? I have no idea of how to proceed. Any feedback or alternative suggestions are much appreciated.


Thanks a lot.

[rhinoceros]

IMO blast is a very poor solution for assigning taxonomy to 16S reads. The proper way would be to do this in QIIME or mothur. I think QIIME now defaults to uclust for taxonomy assignment (it was previously RDP classifier). Not sure about mothur..

[Nilaksha]

IMO blast is a very poor solution for assigning taxonomy to 16S reads. The proper way would be to do this in QIIME or mothur. I think QIIME now defaults to uclust for taxonomy assignment (it was previously RDP classifier). Not sure about mothur..

Thanks I will look into that.

[overflow]

http://qiime.org/scripts/assign_taxonomy.html

or even faster

http://qiime.org/scripts/parallel_assign_taxonomy_blast.html

use that + silva nr ssu.
however, you still have to separate sequences and taxonomy first.