Sequence count by QIIME different from sequencing center


When I look at the demultiplexed sequencing counts for my files from the sequencing center and compare it with sequencing count after demultiplexing by QIIME, they are very different. QIIME shows smaller number. What would that mean?

For example, look at Negative control 2 below.

v3v5-20181012-demux-s.qzv (287.3 KB)

By looking at your provenance for the attached QZV, I learned that you used demux emp-single in QIIME 2 to demux (there are a few other ways to demux, besides this method). The method you used does not perform any sort of error correction, which is likely the cause for reduced recovery in QIIME 2.

My two cents: if you already have demuxed data from your sequencing center, don't bother redoing it in QIIME 2, just import the demux reads and get going! :qiime2: :t_rex:

1 Like

Sure. I will try that. In my post here Lots of unassigned and host DNA in my samples, we were able to recognize that what I am seeing is probably host DNA not microbiome. Do you think that trying with demultiplexed sequences from the sequencing center might change the results in terms of the host DNA problem? I have heard there might be a bioinformatic method to reduce the sequencing depth to maybe ratify the issue of having too much host DNA instead of re-sequencing. Do you know anything about that and do you think that might have anything to do with my demultiplexing problem here?

I am not sure, but my gut tells me “no” - host DNA is host DNA, no matter how you slice it.

No, these things seem unrelated to me.

Actually, doing this is much harder and more time-consuming than simply importing the multiplexed files. I am renaming the files for the Casava method. Does it matter what you use for the lane number and the set number? I am not sure I have this info. Could they be random?

Why not just use the Manifest import method? This use-case is almost exactly what it is designed for!

1 Like

Hi Mathew,

I took your advice and used the manifest method. However, the sequence counts did not change in this method compared to when I imported multiplexed sequences and demultiplexed them in QIIME. For some reason, the numbers from the sequencing center are much higher. A large portion of my reads are removed or shown lower for some reason after I import them to QIIME. This seems a big issue to me. Do you know why that might be the case?

QIIME 2 doesn’t filter reads on import — it uses exactly what it is provided. Is it possible that there was a mistake in the spreadsheet from the sequencing center? Or perhaps the counts from the sequencing center don’t mean what you think they mean?

All QIIME 2 does when you import using the manifest method is copy the entire file into the QZA, it doesn’t look into the contents at all and manipulate it.

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