Chris Fields

@EllaMae-O what I meant when using the term "shotgun" is whether the library preparation used a fragmentation step. Similar to a DNA shotgun library preparation from an environmental samples for...

@trawler-crawler do you know if this data originates from a PacBio Revio? The quality profile and error plot suggests this. The default outputs from a Revio have binned quality scores....

@schyells The MiSeq i100 produces FASTQ with binned quality scores. Internally we did some checks and found it used 2, 12, 24, and 38, but there are very few Q2...

@schyells that error model for the R2 reads looks like the issue where we see the 'dip' when applying the default Illumina error correction to binned data. Can you post...

@benjjneb for what it's worth, I expect this will become more of an issue when the NovaSeqX adds 2x300 reads (I expect probably within a year).

A small update. The newer `makeBinnedQualErrfun` with quality scores of `2,12,24,38` worked for me on a very small MiSeq i100 data set (`123534753 total bases in 539696 reads from 5...

That makes sense, we see very few Q2's in our data overall.

Odd, this is the first time I've seen this. Just out of curiousity, I went back to check a few of our 'worst-case' runs. The worst so far was a...

@dgiovannelli there are a few key steps missing, namely how you ran `learnErrors` and `dada`; these data look binned, so it may be worth checking on the updated methods for...

> Hi all, i've searched around and found [this post](https://github.com/benjjneb/dada2/issues/938#issuecomment-774051061) . [@cjfields](https://github.com/cjfields) I assume this is what you were referring to. Please let me know if I am mistaken. @benjjneb...