Feature request - sample sheet

[adbeggs]

Firstly - thank you - a great tool

A quick feature request - would be great to have the ability to import a "a vs b" .csv sample sheet or similar? Rather than having to plug the entire lot into the command line.

And also, a limma style Bayes model would be great for dmr pair as well, but appreciate this would be more of a challenge.

Thanks!

Andrew

[ArtRand]

Hello @adbeggs,

Yes, I can work on adding those things. As for the sample-sheet, are you thinking of something like modkit dmr multi but with a text file as input or something more complicated? Other people have asked for linear models as well, so I'm investigating those as well. If you're having good (or bad) experiences with the current DMR functionality please let me know.

[adbeggs]

Hi @ArtRand - yes with a text file as well, because we have a list of samples with "condition bad" and we want to compare them to "condition good". When n>10 really it gets very unwieldy placing it all on the command line.

Modkit is very good, we have used the DMR and DMP functionality with no issues at all has given us plenty of insights. Its just with smaller sample sets, frequentist approaches can't cope with low numbers of inputs given the size of the genome, and want to move towards something more Bayesian approach (like with limma)

Thank you!

[ArtRand]

@adbeggs Got it. I can get you a development build with the sample sheet to test soon. I agree that a Bayesian (❤) approach would be a nice add (especially if I can get it to work well with multiple modifications and RNA). I'll keep you posted!

[adbeggs]

Great. @ArtRand always nice to meet a fellow fan of Bayes...