Arthur Rand

I'm going to keep this open to track the work. Thanks again for noticing the inconsistency.

Hello @Yang990-sys, Could you attach the log file for the command that is failing to this thread? Also could you run ```bash tabix -l ${sample} ``` for each of the...

Hello @Yang990-sys, I understand. It would still be helpful for me to debug what's going on if you can provide the logfile `dmr_multi.log`. Modkit never actually merges the files, it...

@Yang990-sys Just to be sure. All of the pileups use the same reference, which in this case is `all_transcripts.fa`, correct?

@Yang990-sys would you be willing to share the data that exposes the problem with me? We can do it through email so that it's not publicly accessible. My email is...

Hello @Macdot3, Could you do two things for me to help diagnose this issue? 1. Tell me the exact dorado basecalling command you used so I know the basecalling model...

Hello @faulk-lab, @Macdot3, and @DHmeduni, Sorry for the delay following up. We have released updates to the 5hmC/5mC models and they should have generally more confident canonical calls. We have...

Hello @billytcl, If you're using Dorado >=v1.0.0 you need to use Modkit v0.5.0. But, sadly, it gets a little worse. It looks like you're using the all-context cytosine modification model...

Hello @billytcl, `modkit extract calls --cpg` will only use CG sites when calculating the threshold value, `pileup` aggregates probabilities from the reads without filtering to aligned sequence context or motif....

> It's too bad there's no way to just re-call the methylated bases without going back to the pod5s... You can do this with remora (as you probably know), but...