filterAndTrim with ITS data

[mya-darsan]

I work with ITS data and I am running into some quality issues.

My usual filterAndTrim parameters are as follows: filterAndTrim(fnFs, filtFs, fnRs, filtRs, maxN = 0, maxEE = c(2, 2), truncQ = 2, minLen = 50, rm.phix = TRUE, compress = TRUE, multithread = TRUE)

This is an image of the forward reads pre and post trim (which I think is fine):

This is an image of the reverse reads pre and post trim (which i think is not great especially with that weird bit around 200):

Any ideas on how to improve the reverse reads without using truncLen (not advised for ITS)? I tried having a stricter expected error and higher quality threshold for reverses and it still didn't address that big dip around 200.

(Additional context: I receive the files demultiplexed and without primers so I do not use cutadapt prior to this)

[benjjneb]

On the quality profile I wouldn't worry about it too much. DADA2 models error rates as a function of quality score, so it will handle the dip in quality in the reverse reads fairly well.

Additional context: I receive the files demultiplexed and without primers so I do not use cutadapt prior to this

The issue with ITS is that short ITS regions lead to reads that read-through to the opposite primer: forward reads that read through to the reverse primer, and reverse reads that read-through to the forward primer. I recommend checking if this might be an issue in your data. Our ITS tutorial describes one way to do this: https://benjjneb.github.io/dada2/ITS_workflow.html