You're welcome :) Just a warning: I'm going on vacation next week, so if you need some help soon, then let me know before tomorrow afternoon, please :)

Hallo Giulia :) Sorry for the late reply. Yes, I mean the relationship between the predictors (independent variables) and the response (dependent variable). Georg made a demonstration of how to check it in the workshop, and you can look through it in one of the documents from there, the 'NDT_models' :) If you don't have the document, I can also send it to you. If you don't understand it, then just let me know. We can also make a video call about it, if you need it! :) Best regards,

Jeppe

Hi Jeppe!

Thank you for your answer! How do I know if the assumption of linearity holds? I know limma can also be performed on non-normally distributed data, but here I think you are referring more at the assumption regarding the relation and not the distribution of the data, right?

I hope my question is clear :)

Hallo :)

First I will just say that I don't know anything about limma in general, but a quick google search told me that it is about linear models. My best guess would then be that there is not a linear relationship between your dependent and independent variables before you do the transformation (or there was and there no longer is after transformation). So the most 'reliable' result would then the one, where the linearity assumption holds (if it does in any of them). I'm not sure, if that helps at all or if it even makes sense for your situation, I'm just speaking of linear models in general.

If nobody else can offer more help, you can also give me some more information about what exactly you're trying to do and I can look more into it :)

Have a nice week as well! :) Jeppe