-P-trim and -p-trunc values selection for qiime dada2 denoise-paired

Dear support team
I know there is not a definite rule for selection on trim and trunc values and it purely depends upon the researchers choice and data quality.
But here I am facing a problem that the overall quality of the reads is very poor. It takes me 2 days to run dada2 command.
Seeing this data, can you please advise me how much trim and trunc values I should use?
I have attached demux graph here.


Hi @drmusk,
Can you give us a bit more info first please. What is the expeted amplicon size, region target and the primers you used? We need to calculate the overlap region before making those decisions. Also, what was your exact command when you ran dada2? Two days completion for DADA2 might be normal if you have a big run, low # of cores and RAM and not a very good processor.
What did the stats results of DADA2 look like after it was completed?
The link you provided is also broken, so we only have the image to go by. Unless this is a complete overlap region I’m going to take a guess that you should just ditch the forward reads and use your reverse only since they seem to be in much better shape for some reason.

1 Like

Dear Mehrbod
Sorry for late reply. Let me answer all questions one by one

What is the expeted amplicon size, region target and the primers you used?
We followed illumina metagenomics protocol which employee v3,v4 region primers. Expected amplicon size is 550bp. With adopters and paired indexes the size of amplicon would be 630bp.
I used -p-trim 13 and -p-trunc 220 for both forward and reverse but got I got empty output file.

Hi @drmusk,
If you are following this protocol using V3-V4, the expected amplicon size is actually 460bp. Meaning on a 2x300 run you have 140bp overlap. DADA2 requires at least 20nt overlap for proper merging, so the maximum combined truncation to have effective merging should be ~ 120. You are truncating 80bp on both the forward and reverse reads, so 160bp. That means you do not have sufficient overlap and dada2 will fail to merge most if not all of your reads. Given the very poor quality of your forward reads, I would actually recommend just ditching those and using your reverse reads alone, try truncating those at 220 as you were. You will lose some resolution but you will retain most of your reads, besides I’m not sure if you have another option at the moment.
Let us know how that goes.

Dear Mehrbod
Thanks. I worked with reverse reads only and I got something to present in then end. Thanks for your advice.
Beside this. In the V3-V4 protocol you mentioned above, the amplicon size is 460bp. But after doing PCR, in this protocol, they say that the product size would be about 550bp. Why is so? Primers?

1 Like

Hi @drmusk,
Glad you got something working! :partying_face:
the 460 bp is the region which your primers flank and amplify in the first PCR. The additional bp come from your barcodes (index) and flow-cell adapters that are attached during your 2nd PCR and are non-biological sequences, so those do not get included in your run/calculation of amplicon length. Hope that clarifies it a bit.