Interpretation of demux.qzv

[pyee0417 (Tan Pei Yee)]

Hi good day!

I've noticed that the quality plot of my demultiplexed sequences is appearing like a series of dashes and lines without boxplot.

demux.qzv (324.2 KB)

The QIIME2 version I’m using is 2019.10 and the sequences were generated by Novaseq.
May I know does it means my sequences had been quality filtered?
Do I still need to proceed dada2/ deblur?

Thank you!

[thermokarst]

@pyee0417

generated by Novaseq

This doesn't surprise me - the quality profiles for novaseq often look quite different than a miseq run, for example.

@pyee0417

May I know does it means my sequences had been quality filtered?

Maybe, but the best way to check is to ask your sequencing center.

@pyee0417

Do I still need to proceed dada2/ deblur?

Most likely, yes - but check with your sequencing center first to see if any quality control or filtering has occurred with these reads - that will impact your decision for DADA2 vs deblur.

[pyee0417 (Tan Pei Yee)]

Thank you very much for your reply!

i have checked with my sequencer center, the sequences had not been quality controlled or filtered.

I plan to denoise it using the following parameters:

qiime dada2 denoise-paired
--i-demultiplexed-seqs paired-end-demux.qza
--p-trim-left-f 0
--p-trim-left-r 0
--p-trunc-len-f 183
--p-trunc-len-r 0
--o-table table_dada2.qza
--o-representative-sequences rep-seqs_dada2.qza
--o-denoising-stats stats_dada2.qza

or

qiime dada2 denoise-paired
--i-demultiplexed-seqs paired-end-demux.qza
--p-trim-left-f 0
--p-trim-left-r 0
--p-trunc-len-f 0
--p-trunc-len-r 0
--p-trunc-q 20
--o-table table_dada2-2.qza
--o-representative-sequences rep-seqs_dada2-2.qza
--o-denoising-stats stats_dada2-2.qza

May I know are the parameters I chose are good to go?
Besides, I've come across a forum that discussed about the similar issue as I am facing now and It was recommended to enforce monotonicity in the fitted error model by changing the error model matrix. May I know is this a necessary step?

Odd display of demux plot

I'm pushing some data through the qiime2 pipeline and am met with an odd demux plot, one that makes it difficult to choose parameters for the dada2 step. I am missing most of the box part of the boxplot that every other demux.qza that I've seen has. The plot is generated using random sampling of 1000000 out of 19442585 sequences, and looks exactly the same as when I use 10000 sequences. I wonder if it might be due to Illumina's quality score binning? Please see the demux.qza object below. demu…

Thank you in advance!

[thermokarst]

Hi @pyee0417 - unfortunately there isn’t formal support for NovaSeq data with DADA2, so you are on your own there. As suggested in that post you linked to, your best bet will be to run DADA2 in R. Fortunately you already have DADA2 installed in your QIIME 2 environment.