Hi,
i am trying to *Assign taxonomy to my data using this classifier
i ran this command: qiime feature-classifier classify-sklearn
--i-classifier gg-13-8-99-515-806-nb-classifier.qza
--i-reads 5_rep_seq_deblur.qza
--o-classification 6_deblur_taxonomy.qza
I got this message
Plugin error from feature-classifier:
The scikit-learn version (0.22.1) used to generate this artifact does not match the current version of scikit-learn installed (1.2.2). Please retrain your classifier for your current deployment to prevent data-corruption errors.
This error happens when you update / change software inside your Qiime2 conda environment. The scikit-learn classifier can break when using a different version than it was trained on, so it's important to keep your conda environment consistent.
Hi,
I reinstalled Qiime2 and tried to run the command, but I got the same message with a different version:
Plugin error from feature-classifier:
The scikit-learn version (0.22.1) used to generate this artifact does not match the current version of scikit-learn installed (0.24.1). Please retrain your classifier for your current deployment to prevent data-corruption errors.
I also tried to create a new conda and install a new software and then switch between qiime2 env, but it did not work too
Debug info has been saved to /tmp/qiime2-q2cli-err-r__fuwy6.log
Ah! It appears we how have a version of Qiime2 that is newer than that database version. I suppose you could install an older version of Qiime, but it's probably best to get a newer (matching!) version of the database from the data resources page.
Thanks for sticking with me while we figure this out.
Well that's no good! It should be possible to have several copies of Qiime2 installed in separate conda environments...
i downloaded several files, but it did not work too
I downloaded the files in qiime2 env, then i tried to run the command:
qiime feature-classifier classify-sklearn
--i-classifier gg-13-8-99-515-806-nb-classifier.qza
--i-reads 5_rep_seq_deblur.qza
--o-classification 6_deblur_taxonomy.qza
i still see the message:
Plugin error from feature-classifier:
The scikit-learn version (0.22.1) used to generate this artifact does not match the current version of scikit-learn installed (0.24.1). Please retrain your classifier for your current deployment to prevent data-corruption errors.
Debug info has been saved to /tmp/qiime2-q2cli-err-dbnclk28.log
Hi @hadeerismail22,
The classifier you are using isn't supported with your current qiime2 env. You will need to do to the Data-resource page that @colinbrislawn linked and install an classifier that is compatible with qiime2-2023.2.
hi
thanks for your recommendation
I am using this classifier gg-13-8-99-515-806-nb-classifier.qza
i tried to follow the page of data resource and i applied all the steps on my qiime2 env, but it did not work too
now I am trying to install scikit learn version 0.22.1 inside qiime2 env as i tried to download freely in another environment, and got the same result
Could you recommend specific steps, please?
Hi @hadeerismail22,
Can you link to where you downloaded that classifier from?
I would recommend against trying to lower the version of scikit learn in your qiime 2 env unless it is absolutely necessary. You could download an older release of qiime 2 with an older version of scikit learn as well.
i would not lower the scikit version in qiime env,
my question is; can i add another version of scikit learn; 0.22.1, or qiime2 could work with 2 versions of scikit learn or not?
In conda environments, only one version of a library like scikit learn can be installed at a time. Adding one removes the other.
(It may be possible to install both at once in the same env, but that would cause conflicts and this is what conda is trying very hard to avoid.)
The goal is to get a set of software and databases that all match. You can start at either side: choose a classifier.qza file and install the right version of scikit learn and qiime2, or find a matching classifier for the skl you are running.
Thanks so much for your help
I changed the classifier and it worked, i used this instead:
qiime feature-classifier classify-sklearn
--i-classifier classifier.qza \