RESCRIPt plug-in error - deprelicate after extracting reads for amplicon-region specific classifier

Hi @ctekellogg,

I just copied and pasted the commands you provided on my laptop, and everything worked successfully. So, I am not sure what might have happened on your end.

Is the error repeatable? Perhaps start from the beginning again? The taxonomy can be a super set of the sequences. So, that error often arises when a given taxonomy is not present in the taxonomy file, but present in the sequence file. Make sure there isn't a file mixup somewhere. :thinking:

Perhaps remove the existing QIIME 2 environment, and reinstall anew, then try again?

FYI, if you add --p-threads 8 to the dereplicate command, and --p-n-jobs 8 to the cull-seqs command it will run much quicker. If you'd like to know what options are available add --help to any qiime command.


On another note, if making an amplicon specific classifier I'd recommend:

qiime rescript get-silva-data ...
qiime rescript reverse-transcribe  ...
qiime feature-classifier extract-reads  ...
qiime rescript cull-seqs  ...
qiime rescript dereplicate ...
qiime feature-classifier fit-classifier-naive-bayes ...

That is, there is likely a chance that the extracted region itself is of good quality, compared to the full-length sequences. That is, you might remove "good" amplicon region data when running cull-seqs, and filter-seqs-length-by-taxon prior to extracting the amplicon specific region.

That tutorial, isn't meant to be an SOP, just a bunch of simple examples about what you can do with RESCRIPt. I just used the output from the previous step to keep the tutorial easy to follow. For example filter-seqs-length-by-taxon, is an example of making a more stringent full-length classifier. Feel free to alter the order, and or skip / add commands you think might be best. :slight_smile: