Paired analysis

[simleopold]

Hi,

I tried to analyse paired methylation data but the results seem weird.
I used the paired analysis because the QC suggests it, as shown below. The patients are identified with their id and the timepoints (10 or 120), we have also two measures per patients.

However, I have more significant DMP for the no-paired analysis et far better p-values too, as shown below (no paired first and paired)
(no paired analysis)

(paired analysis)

Do you have any explanations for this?

Merci de m'avoir accordé votre temps,

Simon