Silva classifier scikit learn version issue

I am having an issue using Silva to classify my reads. I used the following command
qiime feature-classifier classify-sklearn --i-classifier silva-132-99-515-806-nb-classifier.qza --i-reads allreps_rep-seq.qza --o-classification allrepstaxonomy.qza

And I received this error
The scikit-learn version (0.21.2) used to generate this artifact does not match the current version of scikit-learn installed (0.22.1).

I’m guessing I need a newer version of Silva but I can’t find it on their archive, and I’ve tried re-downloading the one from but it isn’t recognizing it as a qza file.

Any help would be much appreciated!!

Hi @smurray4,
You’ll need to either update your latest QIIME 2 version to 2020.2 in order to use the linked classifier or if you can’t update for some reason, just download an older version of the classifier that matches your current Q2 version, for example :

Hi @Mehrbod_Estaki thank you for your reply!! I am actually already running Qiime2 2020.2, the problem seems to be the classifier is the older version (0.21.2) when Qiime2 2020.2 uses (0.22.1) and therein lies the incompatibility

Hi @smurray4 - any chance you’re just mixing up files you have downloaded already? The classifiers that you linked to in your first post are trained with 0.22.1, you can see it in the provenance:

I think Bod might’ve just misread the error message, but the same point holds true: the version of the classifier needs to match with the version of QIIME 2, basically.

My guess is this is just a minor file bookkeeping issue on your end - I triple-checked the linked data resources and all are fine - every one is the latest, trained with 0.22.1. Maybe try downloading again and using all new file names to help you keep your files clearly separated. Keep us posted!


1 Like