qiime dada2 denoise-paired is unresponsive

Technical Support
1

Hi everyone,

I’m currently running QIIME 2 (version: qiime2-amplicon-2025.7) and encountered an issue when using the qiime dada2 denoise-paired command.

Under the same dataset and parameters that previously ran successfully (and completed within a reasonable time), the current run has been taking several times longer than usual — it has been running for hours with no output or progressreported.

Here’s the command I used:

qiime dada2 denoise-paired
--i-demultiplexed-seqs demux-trimmed.qza
--p-trunc-len-f 245
--p-trunc-len-r 200
--p-trim-left-f 0
--p-trim-left-r 0
--o-table table.qza
--o-representative-sequences rep-seqs.qza
--o-denoising-stats denoising-stats.qza

I’ve checked system resources (CPU and memory usage), and everything seems normal — no clear bottleneck or crash message.

Has anyone else experienced this issue with DADA2 in QIIME 2 2025.7? Could this be a compatibility or performance regression in this version?

Any suggestions or troubleshooting advice would be greatly appreciated!

Thank you,
Shuoyun Jiang

2

Have you tried adding “–verbose” to the parameters, this shows the progress of the step. And as long as there is progress it means DADA2 is working on it. Also if you can use the multithreaded parameter to speed it up using “--p-n-threads 0”.

However, abnormally long times could be indicative of some other deeper issue. In one of my cases, it took forever, because I used a longer read length for a smaller amplicon, which ended up sequencing a bunch of the adapters into the reads, that caused DADA2 to think they were all chimeras and remove 80% of my reads. Only way to find out the issue is to let it run. Alternatively, run just a small subset of your sequences (say 10,000 reads only) and check the output for abnormalities. It is normal for DADA2 to take upto 8hours for some large datasets.

1 Like
3

Hello @shuoyunjiang,

What OS are you using, and what kind of environment are you running QIIME 2 in (HPC, Docker, personal laptop, etc.)? Additionally, how large is the data you are using as input to denoise-paired, and have you tried running this analysis on this data in older versions of QIIME 2?

Thank you.