Arthur Rand

Hello @Ge0rges, Thank you for the suggestion. Back when I started looking into DMR I considered the KS-test, but I discarded it since it doesn't easily handle the multivariate case...

Hello @Ge0rges, I don't have that data at hand. If you need a tool that calculates the DMR score or MAP-based p-value, I can whip something up if it helps.

I'd certainly be interested to hear how it works!

Hello @nchernia, I haven't seen this before, but I may not have been looking for this error mode exactly. What is probably happening is that the model hasn't been trained...

Hello @nchernia, You could also increase `--min-coverage`, it's a bit of wack-a-mole since you've already shown that you could have regions of median-depth coverage but high error reads. I think...

Hello @ralanany, Which version of `modkit` do you have installed. The `bedmethyl merge` command was introduced in [0.4.2](https://github.com/nanoporetech/modkit/releases/tag/v0.4.2).

Hello @Raya-Faigenbaum-Romm, The schema for the output segments table can be found in the [documentation](https://nanoporetech.github.io/modkit/intro_dmr.html#segmenting-on-differential-methylation). The argument is a file to write the data to, simply adding the file is...

@dawe thanks for bringing this to my attention!

Hello @DepledgeLab, > I interpret all this as the model simply being unable to classify a given basecall as an A or m6A and reporting it as a third state...

Hello @eesiribloom, The Cohen's h metric is directional, but the CIs are absolute value. You can see this in the test [here](https://github.com/nanoporetech/modkit/blob/5120ef765ea7a29904ca5b2e1ddf1f22ae7f052a/modkit-core/src/dmr/util.rs#L630) actually, reproduced and annotated below: ```rust fn test_cohen_h_signed()...