DADA2 to Qiime2

User Support
1

Hi I am doing meta-analysis (sample size - 80 files) downloaded from NCBI and ENA.I have done the initial sequence analysis in dada2 in R and i have saved the files in RDS formate after makeSequenceTable command.
Now my question is - can I import the individual rds files into qiime2 and merge the seq tables and perform chimera. so that I can use fragment insertion to overcome the difference of primers used.
I tried to merge the rds files in dada2 r but its showing error as ; cannot allocate vector size of 76 GB.

Kindly suggest some way

2

@kulandai1, unfortunately there is no way to perform individual DADA2 steps in the QIIME 2 implementation.

It might be possible to do something like this:

  • import individual denoised samples as SampleData[Sequences]
  • dereplicate them
  • merge the resulting individual FeatureTable[Frequency] and FeatureData[Sequence] artifacts using merge and merge-seqs
  • remove chimeras using vsearch
  • cluster and continue on with the analysis

But this is certainly a non-standard workflow, and it might be helpful to take a look at the clustering otus and the filtering chimeric feature sequences tutorials for inspiration.

Another difficulty might be importing the RDS data into QIIME 2, though this older community tutorial might be helpful here.

I can't say that I am overly optimistic about this working, but if you can't simply start by pulling all of the data into QIIME 2 initially, this might let you move forward! The easiest solution may be to merge the RDS files in R using a machine with more than 76 GB of RAM, though I am not familiar enough with that operation to know how much more you would need.

3

Hi Keegan
Thank you

Thanks, I think this would help.

Sorry for taking your time, I just want to make sure the following workflow for 16S rDNA meta-analysis (different primer regions) is good.

  1. so far I have the ASV tables in rds format (80 files) analyzed in R dada2, now I will convert and import it to qiime2 and merge the table and sequence, then I will do remove chimeras using vsearch.
  2. followed by fragment insertion with the chimera-free files and downstream analysis.
    I need some clarification regarding the merging of sequence and table and the issue of duplicate ASVs
    I read some Q&A in the qiime2 forum, but I couldn't understand. Kindly clarify me, if I use the fragment insertion it can overcome the duplication of ASVs and downstream diversity analysis.

Thank you

5

Hello @kulandai1,

What concern about duplicate ASVs do you have? When you merge tables QIIME will recognize duplicate ASVs and merge them in the expected way. What is it about the process of fragment insertion that you expect to influence/interact with the presence of duplicate ASVs?

As a note, DADA2 performs chimera removal, if that is useful or unknown information to you.

8

Hi Colin Wood
Thank you for reply.

I am planning to use fragment insertion, because my meta-analysis has different primer regions.

11

Hello @kulandai1,

If it would simplify things for you, running dada2 within qiime is a possibility. Then you wouldn't have to worry about importing into qiime afterwards. Has your issue with merging ASV tables been resolved?

12

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