Hi there,
I found it very helpful using ImpulseDE2 over longitudinal data while controlling for several other confounding variables, but I just found maybe these confounding variables that are given to the paramter 'vecConfounders ' should be categorical variables only. If we also want to remove the confounding effect of some continuous variable, is it possible? or is it reasonable?
The reason why I am asking this question is that when I tried removing a PMI (post mortem interval) effect using ImpulseDE2 to identify DEGs over time course in a human AD RNAseq project, note that PMI is different for each sample, and got nothing from p-adj and p-nominal DEGs. And I hypothesize that if we treat PMI as a confounding variable, then each sample is an independent 'batch', or single time point, thus cannot form a curve over the time course.
I was wondering if my understanding is correct, or do you have any suggestions over the issue of controlling confounding effects from continuous variables, thanks a lot!
Hi there,
I found it very helpful using ImpulseDE2 over longitudinal data while controlling for several other confounding variables, but I just found maybe these confounding variables that are given to the paramter 'vecConfounders ' should be categorical variables only. If we also want to remove the confounding effect of some continuous variable, is it possible? or is it reasonable?
The reason why I am asking this question is that when I tried removing a PMI (post mortem interval) effect using ImpulseDE2 to identify DEGs over time course in a human AD RNAseq project, note that PMI is different for each sample, and got nothing from p-adj and p-nominal DEGs. And I hypothesize that if we treat PMI as a confounding variable, then each sample is an independent 'batch', or single time point, thus cannot form a curve over the time course.
I was wondering if my understanding is correct, or do you have any suggestions over the issue of controlling confounding effects from continuous variables, thanks a lot!