I was looking for info about the best way to remove singletons (and best time to do it in the workflow) but I saw this in this thread:
Is that the general idea now or is it only applicable for the rarefaction calculation in that question?
I completely agree with Pat on this. There is no reason that singletons are more likely to be bad sequences than common sequences (remember we are PCRing to generate these data). Removing singletons to improve read quality is a random sledge hammer-maybe you’ll get rid of bad sequences, for sure you’ll get rid of good sequences.