Hi, I'm looking at some 16s microbial data and got the following set of quality scores. Initially I tried trimming these samples at 200 (forward) and 180 (reverse), but my feature table frequency was an average of 25 counts with a peak of 750 (and many, many counts at 1).
P.s. I presume I over-filtered my data, so I want to ask directly how best to interpret these quality scores.
These indeed look like "binned" quality scores. If you used dada2 to generate the feature table, I would recommend looking at the dada2 stats visualization to gauge the impact of your truncation parameters. Try rerunning with other truncation parameters and then compare. If I were using dada2 on these data I wouldn't truncate at all. There is really no definition of what constitutes a "good" quality score. Different people have different rules of thumb. Whatever parameters (within reason) to dada2 give you the highest percent of reads retained is what I would move forward with.
