I would like to use SILVA 138 database for taxonomy classification. However, I am not sure which is the correct file to use.
For example the code for taxonomy classification:
If you are using blast, etc.. you can simply use one of these pre-made sequence files from the Data resources page. These are the files used to make the naïve bayes classifiers.
Also, in order to take full advantage of a "consensus taxonomy" I'd strongly recommend against using --p-maxaccepts 1 otherwise you'll simply be accepting the first, and not necessarily the best, hit to a reference sequence.