Qiime taxonomy classifier for 18S rRNA primers, greengenes or Silva

Dear Colleagues
We used following 18S primers to get metagenomics analysis from a mammalian species tissue samples.


When i got data, I performed QIIME analysis and performed feature-classifier using following silva classifier.

qiime feature-classifier classify-sklearn
--i-classifier silva-132-99-nb-classifier.qza
--i-reads rep-seqs.qza
--o-classification 40K-core-metrics-results/taxonomy.qza

Attached is the output. You can see that more than 99% of the sequences are Eukaryota/ Mammalia

I also used greengenes classifier by mistake but had the same results
qiime feature-classifier classify-sklearn
--i-classifier gg-13-8-99-nb-classifier.qza
--i-reads rep-seqs.qza
--o-classification 40K-core-metrics-results/taxonomy.qza

Now I have 3 questions.

  1. Did I used correct classifier? silva-132-99-nb-classifier.qza
  2. Is it normal to have such huge quantity of eukaryotes sequences?
  3. Is there a way to eliminate eukaryotes sequences and do analysis again?
  4. Since only about 1 percent of the sequences are from kingdom bacteria, is it worth to proceed with metagenomics of these data or it is useless at all?
    taxa-bar-plots-greengenes.qzv (369.1 KB) taxa-bar-plots-silva.qzv (361.8 KB) taxonomy.qzv (1.2 MB) taxonomy-silva.qzv (1.2 MB)

I do not know these primers, but unless if they are designed not to amplify host species, then you will mostly amplify host DNA, since you are using 18S eukaryote primers and amplifying host tissue samples.

Makes total sense, with the caveat above.

Yes, SILVA full-length 18S + 16S will classify 18S sequences.

18S amplifies eukaryote DNA, so yes.

Maybe you want different primers? A different target? You could filter out all mammal sequences and see what’s left.

You will have the same issue with shotgun metagenomics, since you are working with mammalian tissue. Most of the sequences will be host DNA.

Sorry to be the bearer of bad news! Sounds like you may want to find more selective primers.

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