How to decide Forward and reverse primers for training a classifier


I tried to use the pre trained classifier’s but had a following error:

The scikit-learn version (0.20.2) used to generate this artifact does not match the current version of scikit-learn installed (0.21.2). Please retrain your classifier for your current deployment to prevent data-corruption errors.

so i tired to use this command for that
conda install --override-channels -c defaults scikit-learn=0.21.2

After that i tried to run the same command:
qiime feature-classifier classify-sklearn --i-classifier gg-13-8-99-nb-classifier.qza --i-reads rep_seqs.qza --o-classification taxonomy.qza

But i met same error. So i decided to train my own classifier. for that i need to give --p-f-primer and --p-r primer can you please guide me how can i decide that.

I will be used following command to train own classifier:

qiime feature-classifier extract-reads
–i-sequences gg-13-8-99-nb-classifier.qza
–p-f-primer ???
–p-r-primer ???
–o-reads ref-seqs.qza

Let me know if i have missed on anything here.


1 Like

They should be the primers that you use to create the amplicon libraries.

You can cheat a little if you know the V region of the 16S that you’re amplicons are from (e.g., V3 or V3V4, etc.).

edit: So if you are using the Earth Microbiome Protocol these would be the primers that you use:

edit/edit: maybe ask your sequencing person or your core regarding the flanking primers that were used to generate the sequences - for V4 I tend to use the EMP protocol primers


1 Like

Thanks Ben. It worked.

1 Like




@ben, Completely off topic, but you have an amazing reaction gif collection.

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