From DADA2 to dereplication; error QIIME1DemuxFormat filea[Sequences]

[srosales712]

Hi,
I processed my data with DADA2 on qiime2 and I like to subsequently perform a dereplication and clustering step with vsearch. To dereplicate the data, it looks like the sequence data has to be in the 'SampleData[Sequences]' format.
I didn't find a way to go from the DADA2 output format of FeatureData[Sequence] to SampleData[Sequences].
So instead, I exported the DADA2 FeatureData[Sequence] file and then tried to import it as 'SampleData[Sequences]'. When I tried this I got the below error. Is there a way to get the qiime2 DADA2 output into the qiime vsearch dereplicate-sequence plugin? Thanks!

!qiime tools export \
  --input-path rep_merge.qza \
  --output-path exported-DADA2_seqs

!qiime tools import \
  --input-path exported-DADA2_seqs/dna-sequences.fasta \
  --output-path exported-DADA2_seqs/seqs.qza \
  --type 'SampleData[Sequences]'  

There was a problem importing exported-DADA2_240_seqs/dna-sequences.fasta:
exported-DADA2_240_seqs/dna-sequences.fasta is not a(n) QIIME1DemuxFormat filea[Sequences]

$ head dna-sequences.fasta
>00007fa1e4e48c321a7e8ac66fefb764
TACGGAGGGTGCAAGCGTTGTTCGGAATTACTGGGCGTAAAGCGCGCGTAGGCGGCTACTTAAGTCAGATGTGAAAGTCCATGGCTCAACCATGGAAGTGCATTTGAAACTGGGTAGCTTGAGTATCAGA
>00009f8c782556f942f2c23c59ed14c8
GCCATTGATACTGGCGTACTTGAGTACGGACGAGGTAGGCGGAATTTATGGTGTAGCGGTGAAATGCATAGATACCATAAAGAACACCGATAGCGAAGGCAGCTTACTAGACCGTAACTGACGCTCATGC
>0000ef72f9fb70069e941768873a1f83
CACGAACCGTCCAAACGTTATTCGGTATCACTGGGCTTAAAGCGTGCGTAGGCGGCTTGGTAGGTCGGGTGTGAAATCCCACGGCTCAACCGTGGAACTGCGCCCGAAACCCTCAAGCTCGAGGAAGATA
>00015cb0a8036853e41c83406c9c8b9a
TACGGAGGGCGCAAGCGTTACTCGGAATCACTGGGCGTAAAGAGCGTGTAGGCGGATAATTAAGTCAGGAGTGAAATCCTATAGCTCAACTATAGAGCTGCTCTTGAAACTGATTATCTAGAATATGGGAGA 

info

!qiime info
System versions
Python version: 3.6.12
QIIME 2 release: 2020.11
QIIME 2 version: 2020.11.1
q2cli version: 2020.11.1

Installed plugins
alignment: 2020.11.1
composition: 2020.11.1
cutadapt: 2020.11.1
dada2: 2020.11.1
deblur: 2020.11.1
demux: 2020.11.1
diversity: 2020.11.1
diversity-lib: 2020.11.1
emperor: 2020.11.1
feature-classifier: 2020.11.1
feature-table: 2020.11.1
fragment-insertion: 2020.11.1
gneiss: 2020.11.1
longitudinal: 2020.11.1
metadata: 2020.11.1
phylogeny: 2020.11.1
quality-control: 2020.11.1
quality-filter: 2020.11.1
sample-classifier: 2020.11.1
taxa: 2020.11.1
types: 2020.11.1
vsearch: 2020.11.1

Application config directory
/home/rosales/miniconda3/envs/qiime2-2020.11/var/q2cli

[cherman2 (Chloe Herman)]

Hello @srosales712.
Welcome to the QIIME2 community :qiime2:
The reason it is so complicated and not working is because you are adding extra steps to the pipeline!
You data is already dereplicated so you don't need to dereplicate it!

At this point you are ready to feed your data into vsearch for clustering!
Luckily vsearch takes both a SampleData[sequences] or a FeatureData[sequences](which you should have):qiime2: . I would suggest this section of our otu-clustering tutorial. This will give you some step by steps on how to go about clustering your data! It also just has a lot of good information!
Hope this helps!
Chloe

[srosales712]

Hi @cherman2,

Thanks for the response! I think I should have provided more information about my data. I'm conducting a meta-analysis with studies that used V3-V4 primers and other studies that used V4 primers. After running DADA2 on each study, I merged these datasets, and I would like to dereplicate on the merged sequences. The idea is that some of the V3 sequences would cluster with the V3-V4 sequences (assuming that it works similar to CD-Hit). I've also processed the data from DADA2 and then did OTU-clustering, but I was trying to see if adding a dereplication step would reduce the number of ASVs/batch effects.
Thanks,
Stephanie

[cherman2 (Chloe Herman)]

Hey @srosales712
Thanks for the additional information! :qiime2:
So I personally don't have a lot experience merging 2 different primers but I looked into posts on the forum and I think I found some interesting discussions that might be better then dereplicating again.

This post seems like it is doing something very similar to what you are doing

Combining data from different sequencing centers and primers

If we used two sequencing centers that used different primers (V4 only and V3+V4), would it be fine to generate two separate DADA2 feature tables, then merge the feature tables later to analyze the samples together? All samples are unique in both data sets.

This one is a little older but has some interesting discussion on it( although I don't think as relevant )

Combining Datasets with 2 sets of Primers

I have two sets of data (paired reads), with two different primer sets (341/806 and 515/806) that I need to combine. I would like to try trimming my 341/806 set to 515/806 but I’m not sure how to go about doing so before pipelining into dada2. Thank you! Julie

Both mention and link to Fragment insertion. Here is another link .
https://library.qiime2.org/plugins/q2-fragment-insertion/16/

Let me know if this is helpful or if I am way off the mark.
Chloe