How to use Silva 128 for classification?

The Silva reference provided in the mothur wiki is as alignment, while the RDP reference is just a fasta file.
If I want to use Silva for classification, should I convert the alignment to an ordinary fasta?

If yes, what is the most convenient tool to do this? Doesn’t work with simplest text file operations, as both the sequence names and the sequences contain dots.

And finally, is it still so that neither Silva or RDP is clearly better for classification, or is the most recent Silva now better?

Sorry for the dumbish questions, I’m still learning…


The alignment positions (dots and dashes in the aligned file) get ignored during classification, so don’t matter in that sense. You can remove them using degap.seqs to remove them though, because I think the pre-classification database processing would be faster without them.