DADA2, low % input passed filter

User Support
1

Hi everyone!

I'm currently processing paired reads for 16S of V3-V4, forward (341F) and reverse (805R), with DADA2 and I have some questions/concerns about my output files.

I watched the YouTube video and looked through other submissions and replies, but I think I still need a bit of help.

My denoising stats are showing 20-30% of input passed filter which seems super low since I used trunc-len-f 0 and trunc-len-r 0. The code and files are below:

qiime dada2 denoise-paired
--i-demultiplexed-seqs reads_trimmed.qza
--p-trunc-len-f 0
--p-trunc-len-r 0
--p-chimera-method consensus
--o-representative-sequences rep-seqs-denoise.qza
--o-table rep_seq_feature_table.qza
--o-denoising-stats denoising-stats.qza
--verbose

And here are the files:
denoising-stats.qzv (1.2 MB)

reads_trimmed_summary.qzv (318.6 KB)

Before trying trunc-len 0, I tried values like 240 but then only 2% passed the filter.

Does anyone have any advice as to why the % is so low or what I should do moving forward? I did fastqc on all samples and there was no adapter content and I removed the primers.

Any help is greatly appreciated! Thanks so much.

3

Hi @stephhhhanniee,
Welcome to the :qiime2: forum!
Usually if your data is not passing the filtering step it is because the quality of your data is too low towards the end of your sequences. If you are not truncating at all, it seems likely that is the cause.

Although, I would need to see your demux.qzv to confirm. Could you send that so I can take a look?

Interesting? Only 2% passed the filtering or did you get better results at the filtering step and worse results at something else

We will figure out why your data is so low! This is one of those cases where we just have to play around with parameters. It looks like the quality drops significantly for your reverse reads lets try setting --p-trunc-len-r 220 and then can you post the denoising stats for that dada2 run?
Thanks!
:turtle:

1 Like
5

Hi, thank you so much for your help!

For your first question about the demux.qzv file, I think I uploaded all the output files I had? And when I did trunc-len-f 240 and trunc-len-r 240, I did only get 2% passing filter.

I tried trunc-len-r 220 (code below) and now my % input passed filter is around 70% which is wayyy higher! Do you think 70% is ok to keep going?

denoising-stats.qzv (1.2 MB)

qiime dada2 denoise-paired
--i-demultiplexed-seqs reads_trimmed.qza
--p-trunc-len-f 0
--p-trunc-len-r 220
--p-chimera-method consensus
--o-representative-sequences rep-seqs-denoise.qza
--o-table rep_seq_feature_table.qza
--o-denoising-stats denoising-stats.qza
--verbose

2 Likes
6

Hi @stephhhhanniee,
These numbers look great! My general rule of thumb is that if you are losing more than 50% for all samples in a filtering step then there is a problem. It looks like you are getting around 70% passing all filters. Looks good to me!

1 Like
7

Thank you so much for all your help and guidance!! I really appreciate it.

1 Like
8

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