Quick question about which import method to use for preprocessed fastq.gz files

Hey there!
Long time user of QIIME, and excited to try QIIME 2! I have a quick question though.
I recently received sequencing files that consist of 2 sequencing runs on the same samples, forward and reverse reads. So 4 fastq.gz files per sample. I have done quite a bit of pre-processing, including the following (in order):

  1. Merged paired-ends using PEAR
  2. Concatenated the 2 runs (2 merged paired-end files) together for each sample
  3. Adapter and quality filtering using Trimmomatic
  4. Separated files into folders according to research project

So now I have directories with merged, concatenated, and trimmed fastq.gz files, one per sample.

My question is which method of importing should I use? I assume that since I already merged the paired-ends, that I should use one of the single-end options. At first I was thinking “SingleEndFastqManifestPhred33” might be appropriate, since Phred33 was used in the Trimmomatic filtering step. But then I read about “Casava 1.8 single-end demultiplexed fastq” which also pretty accurately described my files.

Also, once imported, would you still recommend running Deblur or DADA2 prior to chimera checking, OTU picking, and other downstream analysis?

I am using QIIME 2018.4 on macOS High Sierra version 10.13.4.

Thanks for any help!

See this tutorial for importing and denoising. There is an example of importing as JoinedSequencesWithQuality near the bottom of that page.

Use deblur. Dada2 does its own paired-end read joining and can have issues with pre-joined reads, since some joiners arbitrarily alter the quality scores. The tutorial covers that step.

Deblur has its own chimera filtering step, so don’t worry about that.

Note that QIIME2 can also perform similar steps to your pre-processing steps. See the tutorial I linked to for read joining, and see q2-cutadapt for trimming adapters/primers.

I hope that helps!

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