Benjamin Callahan

What version of the UNITE release are you using? To be comparible with `assignTaxonomy` you should be using the "General Fasta" releases, probably "All Eukarotyoes". See the dada2 training data...

> such that each is now a derep class object. Is it then possible to filter and trim (using fastqFilter?). If so, should this be done to each derep class...

First, I'd just leave out the explicit `derepFastq` calls, they aren't needed. i.e. ``` errF

There is not. I would recommend using a different tool to get best hits in a reference database to move forward in this fashion. Note that there has been a...

> pretty much stick with 99.9% consensus CCS call That's what we start with before running the DADA2 workflow. Usually also include a loose `maxEE` filter and a length window...

The learned error rates look really high for HiFi sequencing. What chemistry/instrument version are you using? What is the environment/sample-type you are sequencing?

Huh... I don't know, that should be handled pretty well by `learnErrors` from my experience. A couple more things to try: `plotComplexity` on a couple sample fastqs after after primer...

> Is this expected? Yes, the V3V4 region in bacteria has a bimodal length distribution that differs by about 20 nts.

There is a poorly documented, but intentional, behavior by decontam that when using "combined" methods (or other methods that use both frequency and prevalence) that the negative controls are excluded...

> So, my question is whether I should run this analysis on all samples or only on the subset? I'm assuming the analysis you are referring to is decontam. In...