DADA2 parameters, Data Quality is so poor

Dear Colleagues
We ran a metagenomics 301 bp long paired end sequencing on MiSeq using Nextera XT kit.
Our amplicon size was about 550 bp (V3, V4).
Our student said that MiSeq stopped at 250th cycle but then she got result at the end.
Anyways.
We got data and surprisingly Forward reads are 301 bp while reverse reads are 184 bp long.
Also the quality of the sequencing is not good.
I am attaching here demux.qzv file to ask you suggestion for dada2 analysis.
demux.qzv (291.8 KB)

Thanks for the help.

Hi @drmusk,
That’s too bad about the run!
The only option I see with this run is to discard your reverse reads and simply use the forward reads only. Since with your primers you would only have a 50bp overlap between your F and R reads and your reverse reads are already interrupted at 180bp that means you would have no overlap and therefore DADA2 (or any other tool for that matter) would be able to merge those reads together.

So just run q2-dada2 denoise single on your forward reads. As for the parameters, trim the first 10 bases as they seem rather poor but for the truncate parameters you may have to try a few different ones. You could try somewhere high first like 270bp, see if enough reads are retained across all your samples. If this tosses out too many reads then you can start dialing it down to say 240 then 190 and etc until you are satisfied with your coverage, though of course the trade-off is loss of resolution as you go lower.
If you do plan on doing a few of these runs you can always also use deblur, I find it when dealing with shorter reads such as your case here the results between deblur and dada2 are very similar but deblur will complete much faster.
Hope this helps. Good luck!

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