Two different primers in my Ion torrent sequence file


I have a fastq sequence file (have both multiplexed and demultiplexed files) from ion torrent.In total there are 192 samples in the sequence with 2 barcodes (16s and ITS). 96 samples with 16s primer and 96 samples with ITS primers and both the 16s and ITS primers are attached to the same barcodes. I find it difficult in Qiime 2 to import or analyse the data. In Qiime1 I can just change the primer sequence in mapping file and either 16s or ITS primers get excluded. Is there a simiar method to do this in Qiime2 as well?

Is it possible to separate the sequence file based on primers and do the downstream analysis. I have been trying different methods but couldn’t able to figure out.

Kindly advise how to proceed with my sequence data.

This should not be an issue to import, and should cause minor issues with processing.

Since 16S and ITS are quite dissimilar, you can use alignment-based filtering to remove these. Two options:

  1. deblur performs positive filtering as part of the pipeline (i.e., it removes sequences that do not roughly resemble the reference database). This will NOT work for you, however, as I believe deblur is not appropriate for Ion Torrent data.
  2. You can use qiime quality-control exclude-seqs to perform this same type of filtering. This would occur after denoising/OTU clustering.

So in your case, use dada2 denoise-pyro to denoise (or use OTU clustering if you prefer), then use exclude-seqs to separate out 16S and ITS sequences.

Thank you, I will try this.


I did the exclude-seqs step after denoising my sequences. If I am not wrong I should use the rep-seqs.fasta file I got as an output from the denoising step as an input here. If that is so, I understand that I will get the new file from the exclude.seqs with 16s and ITS sequences. But how do I get the OTU table for my 16s and ITS sequences? I understand that there is one OTU table output from the denoise-single.

I used following commands to separate the 16s and ITS sequences

qiime dada2 denoise-single
–i-demultiplexed-seqs Run1_trimmed16s.qza
–p-trim-left 13
–p-trunc-len 200
–p-chimera-method consensus
–p-n-threads 0
–o-representative-sequences rep_seqs_Run1.qza
–o-table table_Run1.qza
–output-dir Denoise

#Excluding/separating 16s and ITS sequences
qiime quality-control exclude-seqs
–i-query-sequences rep_seqs_Run1.qza
–i-reference-sequences RDP_16s_sequences.fasta
–p-method vsearch
–o-sequence-hits seqs_16s.qza
–o-sequence-misses seqs_ITS.qza
–output-dir Exclude_seqs

Kindly advise.


denoise-single also outputs a sequence artifact. Looks like it will be in the Denoise directory that you generated.

Is it possible to do this in qiime1 and the import the data to qiime2?

I only got one output file in the denoise folder, that is denoising_stats.qza

never mind, your table is table_Run1.qza1

No. qiime1 has no denoising methods.

Not to denoise, just to separate the sequences based on primers (as 16s and ITS). I tried running, but it didn’t work since the command doesn’t recognise the forward primers. As my data is single-end reads I am unable to use the reverse primer truncate_remove option. If I do that, I lose lot of sequences.

I think not — but you will need to ask on the qiime1 forum for more info.


if I have to do that in Qiime2 and coming back to my problem, which sequence output file I should use as an output for exclude -seqs command? I only get one output sequence file from denoise-single command. That is rep_seqs.qza file. In the denoise folder I only get some stat file.

Your exclude-seqs command is correct.

I think I finally understand what you are asking. You need to filter your feature table to include only the features that pass exclude-seqs (to get an ITS-only and 16S-only feature table). To do this, use qiime feature-table filter-features, using the correct sequence files as metadata inputs to that command.

Good luck!

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