Forward, Reverse or both?

Hi. I want to denoise my paired-end-demultiplexed data with dada2.
I have trimmed the primers and adapters using cutadapt with your help :
qiime cutadapt trim-paired \

i- *demultiplexed-sequences paired-end-demux.qza *

–p-front-f CCTACGGGNGGCWGCAG **

–p-front-r GACTACHVGGGTATCTAATCC **

*–p-match-adapter-wildcards *

*–p-match-read-wildcards *

*–p-discard-untrimmed *

–o-trimmed-sequences paired-end-demux-trimmed.qza

I have read many of the questions in the forum and find out I have to keep about 12 bp and some normal variation.
according to PCR catalog: Order amplicon primers–The protocol includes the primer pair sequences for the V3 and V4 regions that create a single amplicon of approximately ~460 bp.
I think if I want to use both forward&reverse sequences I can use
–p-trim-left-f 0 **

–p-trim-left-r 0 **

–p-trunc-len-f 283 **

–p-trunc-len-r 247 **
(please let me know if I made a mistake in choosing them)
so I will have (283+247) - 460 = 70 overlapping region and its enough
I don’t know if there is a threshold for the percentage of input non chimeric sequences but I don’t think my results were enough as you can see in the picture:


also I tried these steps with dada2-single denoising in this way :
–p-trim-left 0
–p-trunc-len 283
and I think it was better :

my question is that if I did all the steps correctly ( in using of cutadapt, identify trim & trunc parameters , etc.) which result is more reliable to continue? are the paired-denoise result enough or its better to use just the forward reads?
I’m really sorry for my long question but after I read many questions in the forum I couldn’t decide which one is better.
Thank you very much in advance for your time.

@mohsen_ej, are these low-biomass samples? Do you expect low read counts? It looks like you’ve only got between 700 and 1600 features per sample.

Chris :whale2:

Sorry, but I did;t do sampling and PCR procedures. but I know these are some different types of meat samples. I’m not sure this is helpful or not
Thank you


this is Per-sample sequence counts after I imported the data before trimming by cutadapt

@mohsen_ej, I have no experience with meat samples. If you’re expecting to see low read counts, that will significantly effect your analysis. You should probably consult the literature, or discuss with colleagues who have experience with meat sample data to figure out what a reasonable sequencing depth might look like. Let us know once you have a clearer picture.

so if it is natural to have this sample depth, could you please tell me which option is better based on the information I mentioned?
use paired samples or just use the forward read?
because the sequence center has checked the data and then sent to me so I think they expect this kind of quality of data.
Thank you

If this sampling depth is expected…

the answer to this question still depends on your study. Sequencing depth is probably more important at this level for many studies, but for others the length of the reads might be invaluable.

Regardless, this may be an unnecessary choice. You can probably improve your paired-end data recovery significantly by adjusting the DADA2 parameters you’ve chosen. There are tons of explanations about how interpret DADA2 stats like the ones you posted, and how to set DADA2 trim/truncation parameters to improve read counts on this forum. Do some searching :mag:, read up, and see if you can reduce your loss.

CK

Thank you for your explanation.
So you mean it’s better to work on both forward and reverse sequence by adjusting trim & trunc parameters and try to improve read counts and no need to ignore the reverse data.
If I am wrong, correct me please. because in some of the questions, you suggest to just continue with forward reads because the reverse read is low quality.
thank you