Error in Denoising Step - Loosing 4 samples of 6 that I need to Analyze

[fellora (Felipe Huanachin Carahuanco)]

Hello all,

I'm using Qiime2: qiime2-2021.11

I already used the dada2 script with no problems but now I have one with a new dataset.
I have a problem with the qiime dada2 denoise-paired step.

When I do this step:

qiime dada2 denoise-paired
--i-demultiplexed-seqs aguas.qza
--p-trim-left-f 0
--p-trim-left-r 0
--p-trunc-len-f 235
--p-trunc-len-r 235
--p-n-threads 31 
--o-table aguas_table.qza \
--o-representative-sequences aguas_rep-seqs.qza \
--o-denoising-stats aguas_stats.qza \
--verbose 

I have the following error message appearing.

Command: run_dada_paired.R /tmp/tmp1ubwn1ot/forward /tmp/tmp1ubwn1ot/reverse /tmp/tmp1ubwn1ot/output.tsv.biom /tmp/tmp1ubwn1ot/track.tsv /tmp/tmp1ubwn1ot/filt_f /tmp/tmp1ubwn1ot/filt_r 235 235 0 0 2.0 2.0 2 12 independent consensus 1.0 31 1000000

R version 4.0.5 (2021-03-31)
Loading required package: Rcpp
DADA2: 1.18.0 / Rcpp: 1.0.7 / RcppParallel: 5.1.4

1) Filtering The filter removed all reads: /tmp/tmp1ubwn1ot/filt_f/Up01_10_L001_R1_001.fastq.gz and /tmp/tmp1ubwn1ot/filt_r/Up01_10_L001_R2_001.fastq.gz not written.
The filter removed all reads: /tmp/tmp1ubwn1ot/filt_f/Up45_10_L001_R1_001.fastq.gz and /tmp/tmp1ubwn1ot/filt_r/Up45_10_L001_R2_001.fastq.gz not written.
The filter removed all reads: /tmp/tmp1ubwn1ot/filt_f/Down45_10_L001_R1_001.fastq.gz and /tmp/tmp1ubwn1ot/filt_r/Down45_10_L001_R2_001.fastq.gz not written.
The filter removed all reads: /tmp/tmp1ubwn1ot/filt_f/Down01_10_L001_R1_001.fastq.gz and /tmp/tmp1ubwn1ot/filt_r/Down01_10_L001_R2_001.fastq.gz not written.
Some input samples had no reads pass the filter.
x.xx.x

2) Learning Error Rates
456320650 total bases in 1941790 reads from 2 samples will be used for learning the error rates.
456320650 total bases in 1941790 reads from 2 samples will be used for learning the error rates.

3) Denoise samples ..
..

4) Remove chimeras (method = consensus)

6) Write output

Saved FeatureTable[Frequency] to: aguas_table.qza
Saved FeatureData[Sequence] to: aguas_rep-seqs.qza
Saved SampleData[DADA2Stats] to: aguas_stats.qza

Notice that my 6 samples appear in the filter step, but in the result I only have 2 samples.

I tried do do the following: changing --p-trunc-len-f and --p-trunc-len-r for possible merging problems, but nothing.

Why is it not working ? How can I know where is the problem in my data?
When I looked at my demux.qzv file, all the counts were good in "Per-sample sequence counts".

aguas_stats.qzv (1.2 MB)
aguas.qzv (317.7 KB)
aguas_table.qzv (454.2 KB)

Thank you in advance for your help.

[lizgehret]

Hi @fellora,

Welcome back to the forum, and thanks for your patience here!

@fellora

I tried do do the following: changing --p-trunc-len-f and --p-trunc-len-r for possible merging problems, but nothing.

It doesn't look like your trim/trunc values are the source of the issue here. Is it possible that you haven't removed all adapter, linkers, and barcodes? Non-biological signals in your reads could certainly cause the issues you are seeing. If that's the case, you'll need to start there prior to running DADA2.

Maybe as a sanity check you could try processing just your forward reads, and see if you see this same aggressive filtering. No need to re-import, just pass your paired end reads to qiime dada2 denoise-single and it will use the forward reads.

Hope this helps!

Cheers