Benjamin Callahan

> Do you have any idea why sampling 10^10 bases leads to the Q37 point being lost/dropped for these few (G2C/G2T & C2A/C2G) instances? Sorry, but no. I don't fully...

@wangjiawen2013 See this comment: https://github.com/benjjneb/dada2/issues/791#issuecomment-502256869

This is coming up enough that we ought to at least implement an enforced monotonicity option in the main package. We are still waiting for good test data on qual-binned...

> > This is coming up enough that we ought to at least implement an enforced monotonicity option in the main package. > > We are still waiting for good...

You're OK to proceed. It looks ugly on binned data, but still seems to work reasonably well. You can read much more at the main thread for DADA2 and binned...

It's probably a memory issue, or a databse format issue (as that doesn't look like one of our standard databases), or a combination of the two. How was that reference...

Can you post the full ID lines, not just the beginnings? When you say it "comes from the MIDORI database" -- this is released by the MIDORI database folks? Or,...

Format looks OK. Is there a link to this file? So I can try it out on my own machine. As for the memory issue, I would reach out to...

First you need to identify how much overlap there is between. your forward and reverse reads. This depends on your primer set and how your library is prepared, in particular...

> So is it possible that the problem is related to the small number of sequences originating from jellyfish bacteria? Yes, that is likely. `assignTaxonomy` is implementing the naive Bayesian...