Inquiry about Demultiplexed sequence counts summary

Hello everyone.
I am a university student who has started to do research on intestinal microbiome.
I am analyzing 16 fecal samples, and I have created a manifest file and then got a qzv file using the following command, but when I look at the resulting qzv file, forward and reverse appear respectively.

manifest file
sample-id absolute-filepath direction
S1 /Users/inukaiyousuke/Desktop/qiime2/ngs1/1730_S22_L001_R1_001.fastq.gz forward
S1 /Users/inukaiyousuke/Desktop/qiime2/ngs1/1730_S22_L001_R2_001.fastq.gz reverse
S2 /Users/inukaiyousuke/Desktop/qiime2/ngs1/1828_S23_L001_R1_001.fastq.gz forward
S2/ Users/inukaiyousuke/Desktop/qiime2/ngs1/1828_S23_L001_R2_001.fastq.gz reverse

①qiime tools import --type 'SampleData[PairedEndSequencesWithQuality]' --input-path atemanifest1.csv --output-path paired-end-demux-ate1.qza --input-format PairedEndFastqManifestPhred33

②qiime demux summarize --i-data paired-end-demux-ate1.qza --o-visualization ate1.qzv

When I analyzed another sample with the same command before, forward and reverse were not separated.
Are forwad and reverse not merged?
I was hoping you could give me a solution.

Welcome to the forum, @yokmok !

Q2-demux doesn't join paired-end reads. You can do that during denoising with DADA2 (example), or you can do your own qc and join them with q2-vsearch.

You can use the 'Provenance' tab of your prior visualization to see what your commands were, and deduce why you don't have forward and reverse reads there.

Chris :bird:


Thanks for the reply.

As I mentioned above, when I used another samples in same command before, it was not necessary to join paired-end reads. Do you know why this is?
Sample files and manifest files look same, like
S1 /Users/inukaiyousuke/Desktop/qiime2/ngs1/1730_S22_L001_R1_001.fastq.gz forward
S1 /Users/inukaiyousuke/Desktop/qiime2/ngs1/1730_S22_L001_R2_001.fastq.gz reverse

I used Provenance tab in both qzv files, action details looks same for me.

If you'd like to share the one with only one set of reads, I can take a look and let you know what I think. Here is great, or by private message if you're concerned about sharing too broadly.

Thank you for your help.
Here's one of my samples.

1730_S22_L001_R1_001.fastq.gz (6.7 MB)
1730_S22_L001_R2_001.fastq.gz (5.7 MB)

I'm sorry.
Now that I checked, the problem seemed to be that the sample ended up at 150 sequence base.
The correct sample is up to 300 sequence base.
From here, I can proceed to denoise in the normal way, right?

Not exactly sure what you mean here. Are these 2x150 paired-end reads? Or 2x300?

Side note - I can probably be more helpful if you share your demux-summary.qzvs than I can with raw sequences. Sorry I wasn't clearer about that!

Sorry for the late reply.
Here is my demux-summary.qzv.
These are 2x300 paired-end reads.
qzv shows forward read and reverse reads, but both same sequence counts.
Is this correct?
In next step, I'm going to try denoise using p-trunc-len-f 295 --p-trunc-len-r 250.

startate.qzv (317.1 KB)

and another qzv is here.
I Used the same method to analyze, this dosen't separate forward and reverse reads.
LEN.qzv (290.6 KB)

Thanks for sharing those visualizations, @yokmok. LEN.qzv was created with a rather old version of q2-demux (2019.7), while startate.qzv was created with something more recent.

In 2020.6, this pull request improved the visualization, allowing it to display separate summaries for forward and reverse reads.

If you click on the "Interactive Quality Plots" tab in either viz, you'll see that F/R reads are there in both - you'll have to join your reads downstream of demuxing in either case.

Oh I understood now.
Thanks for your great kindness!

1 Like

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