We are working with paired-end 16S data from MiSeq. We are to the point of visualizing the demux.qzv files. The reverse reads are lower quality compared to the forward, which I know is something that often happens with Illumina data. So my question is moving forward how can I drop these reverse reads? Should I just go back and demux with a single read? Or can you simply "trim" everything on the reverse? I have attached the Quality Plot here to show the difference in quality.
I do not recommand you to drop the reverse sequence since the quality is actually fine in my opinion.Also there is an easy way to truncate the low quality base in the tail of both forward and reverse sequence by q2-dada2 or some other plugins provided in qiime2.
Follow the official doc “Atacama soil microbiome” tutorial and you can learn how to deal with your data in the same way.
I recently did this exact scenario. My sequence quality was even worse for the tail end of my reverse sequences and after trimming the low quality bases, there was not enough to join. If you can save the reverse reads by trimming and still get joining as suggested by @sixvable , that would definitely be preferred, but if you end up in the same situation as me, you can just process your already demultiplexed paired reads just using commands for the single reads. I did this for my own sequences with the deblur steps so I am sure this will work and it looks like paired end data is also accepted by single read version of dada2.
Thank you for the response! I was hoping to keep the quality score between 30 and 35 which is why I was so concerned with the quality of the reverse. However, I see that perhaps I can just truncate those tails and maybe a few on the left for reverse. I was going to truncate at about 140 or the forwards.
