Hi,
I am analysing paired-end 16S rRNA amplicon data in QIIME 2 and would like to use DADA2 because my previous datasets were processed with the same pipeline, and I need the results to remain as comparable as possible.
After importing the FASTQ files and generating the demultiplexing summary, the quality plots were almost completely flat at Q38 for both forward and reverse reads. I inspected the FASTQ files directly and found that the quality strings consist almost entirely of G, with only occasional lower-quality characters. Under Phred+33 encoding, G corresponds to Q38.
For example
@read_id
TACGGGAGGCAGCAGTGGGGAATATT...
+
GGGGGGGGGGGGGGGGGGGGGGGGGG...
These scores in the files are from the sequencing company. I did not do any quality filtering yet.
Can DADA2 still be used reliably with nearly uniform Q38 scores? If so, are there any recommended parameter changes or checks I should perform to confirm that the error model, filtering, and denoising are behaving appropriately?
I have been doing this analysis for years, and I need to be consistent and not change the pipeline. Is there any way?
I would be grateful for any idea.
Thank you very much.
Kejty