Creating A SILVA Classifier

Background Information:
I am currently re-running some data through the QIIME 2 pipeline and have gotten to the classifier step when I hit a snag. The current version of the Silva Classifier posted to the documents page is not compatible with my current version of QIIME 2 (2020.2) in a conda environment.
As such I decided to create my own classifier using the 138 Silva release and it is not going well. I have unzipped the file (SILVA_138_SSURef_tax_silva_full_align_trunc.fasta) and have vsearch and biopython installed.

Current Predicament:
I am trying to follow the Mike Robeson code on github ( but I do not know enough about python or coding or the general classifier making process itself to understand what I am doing and where I maybe going wrong. I have tried to look through the forum and other resources, but in all honesty they just made me more confused and made me feel more hopeless.

So I am trying to run the very first step of the pipeline ( and I have no idea what I’m doing because I have no idea what this means and Google was not helpful in the least.

from cogent.parse.fasta import MinimalFastaParser
from string import maketrans

HI @gpgarcia,

This is because you are running an old version of QIIME 2! I recommend just installing the latest version if you can.

If you notice, the README on that repo notes that that has been superceded by another repository, which has been superceded by RESCRIPt.

And the great @SoilRotifer himself recently wrote a tutorial for using this new plugin to create a SILVA classifier:

Though the QIIME 2 data-resources page also hosts some pre-formatted SILVA sequences and taxonomies that you can use to train a classifier, no need to follow that whole tutorial.

But the easiest thing of all would be to install the latest version of QIIME 2 so that you can use the latest pre-trained classifiers.

Good luck!

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.