taxonomic q2 feature-id doesn't match dada2 feature id table

Hello,

So my problem is linked to the taxonomic assignation.... In fact, I 've downloaded the silva 138 release version provided by this website (the one that has been re-arranged with Rescript software) and when I do it with the NB classifier it gave me this results:

https://www.casimages.com/i/210625045658187371.png.html

But when I'me trying to create a barplot under qiime2 it did'nt worked because the q2 feature id column is not the same (under dada2, it's something like a list of string character i.e. 855cdsr4e6fd1s1fes)

Can you help ?
Thanks in advance :slight_smile:

The image you are showing looks like the reference sequences used to make the SILVA classifier. The labels are basically the accession and the start-stop positions of the reference sequence.

Can you provide all of the exact commands you ran? It appears you may have classified the SILVA reference sequences against itself, and not your own sequence data?

-Mike

Hi @SoilRotifer,
Thank you for your answer :slight_smile:

Here are the commands used:

#Train the classifier
qiime feature-classifier fit-classifier-naive bayes \
--i-reference-reads ref-seqs.qza \ *#the reference sequences I extracted from silva 138*
--i-reference-taxonomy silva-138-99-tax.qza \
--o-classifier $HOME/NB_cl.qza

#test the classifier
qiime feature-classifier classify-sklearn \
--i-classifier $HOME/NB_cl.qza \
--i-reads $HOME/ref-seqs.qza \
--o-classification $HOME/NB_taxo.qza

I think I misunderstand the naive bayes classifier training....

EDIT: I've just realized where the mistake came from .... instead of putting the representative sequences obtained on dada2 in --i-reads ,I've put the extracted sequences in silva 138 .... and as you said I have classified the SILVA reference sequences against itself... :sweat_smile: Sorry again for the mistake...

2 Likes

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