View Single Post
Old 04-27-2011, 01:57 AM   #4
palmy
Junior Member
 
Location: Germany

Join Date: Sep 2010
Posts: 1
Default

Quote:
Originally Posted by Simon Anders View Post

Of course you can compare as many conditions as you like with DESeq -- you just have to do so pair-wise.

If you have two conditions A and B, you may ask: Is the expression of my favorite gene larger in A or in B? How would you phrase such a "larger than" question with more than two conditions?
I’m not sure but I think the experiment, which I currently try to analyse, deals with such a “larger than” question.

The experiment is about analysing circadian or rhythmic expression of genes at different timepoints of the day. Generally in such experiments it is not clear which conditions you have to compare pair wise. As there is no condition which can be seen as untreated or treated.
E.g. you have RNA samples of 4 different timepoints of a healthy patient and 4 RNA samples of the same timepoints of a patient with a certain health condition. If you now ask which genes were differentially expressed within the 4 different timepoints in the healthy patient but not in the sick patient, I think you have to compare all 4 conditions of the healthy patient at the same time.

I initially thought I could solve the problem; when I do pairwise comparisons between the different timepoints (e.g. timepoint A with B, timepoint B with C, timepoint C with D and timepoint D with A) but I think one get many false positives with such an analysis because of the multiple testing problem.

Could DESeq cope with such a problem?
palmy is offline   Reply With Quote