Ok, thanks for checking there first.
The V2 (version 2) in the import type you see corresponds to the style of the manifest file and not the Phred scores.
You certainly don't want to do such a thing even if you could. The quality scores there are crucial for DADA2 to build an error model which is then used to denoise your reads. Manually changing those is going to mess up with that process, so the best way forward is to sort this out right from the beginning, hopefully we can help you there!
A couple of follow-up that will help us diagnose:
- What is the sequencing technology used here
- Has there been any other quality control applied to your reads prior to importing them into Q2 for DADA2? Anything performed by you or your sequencing facility that may modify the quality scores?
- Can you manually look through some of the quality scores in Sample 1's raw fastq file? We're expecting to only see characters corresponding to the Phred 33 ASCII characters (you can paste a few of the lines here if you think they look odd)