The reduce of sample num after dada2

[Ash1One (Hao Huang)]

Hi,I am new user of qiime2 and I met a problem when I first used it.
Virtualbox-QIIME 2-2018.11
Commands:
qiime dada2 denoise-paired
--i-demultiplexed-seqs chicken-paired-end-demux.qza
--p-trim-left-f 5
--p-trim-left-r 5
--p-trunc-len-f 300
--p-trunc-len-r 300
--o-table table.qza
--o-representative-sequences rep-seqs.qza
--o-denoising-stats denoising-stats.qza

The chicken-paired-end-demux.qzv:

The table.qzv:

I wonder why there are only four samples left.Then I checked the denoising-stats.qzv:

I want to know if the tablefile is normal or any mistakes I did.

[jwdebelius (Justine Debelius)]

Hi @Ash1One,

Welcome!

It looks like your reads drop in quality at around 290 bp on the forward read and around 210 on the reverse. (At least, this is my impression from the quality control plot.) My suspecion is that DADA2 is having trouble with read joining, and therefore discarding your samples.

Let me make three suggestions.

1. First, Id discuss the quality of your reads with your sequencing center, because the last bases seem problematic.
2. I might trip the last 10 bp-20bp at least from both the forward and reverse reads before joining, because it looks like you have absolutely no quality there.
3. You could try a single ended technique on your forward reads, but again, Id drop the last 20bp because of your poor quality.

Best,
Justine

[Ash1One (Hao Huang)]

Thanks for your suggestions that enlightened me.I will try your suggestion.
Could you tell me what kind of trouble with read joining would cause that DADA2 discards samples?
Like sequencing error that the match position’s quality of both forward reads and reverse reads are too low resulting in discarding? or reads is too short?

[jwdebelius (Justine Debelius)]

Hi @Ash1One,

Dada2 does filtering based on quality. You can see your reads are dropping out on quality filtering to begin with (input -> filtered). But, the surviving sequences may then not be good enough for joining. Your issue in joining will be too low of quality between the two reads.

Best,
Justine

[Ash1One (Hao Huang)]

Thank you:grimacing:

[Ash1One (Hao Huang)]

Thank you so much again! @ jwdebelius
I have tried you second sugguestion that spending 6 hours in virtualbox on the command:
qiime dada2 denoise-paired
--i-demultiplexed-seqs demux.qza
--p-trim-left-f 5
--p-trim-left-r 5
--p-trunc-len-f 295
--p-trunc-len-r 226
--o-table table.qza
--o-representative-sequences rep-seqs.qza
--o-denoising-stats denoising-stats.qza
Then denoising-stats.qzv:

[jwdebelius (Justine Debelius)]

@Ash1One,

Your stats look a lot better to me, at least isn that they’re surviving filtering.

Best,
Justine