Hello,

I want to make a transcriptomic analysis with Mapman (Wilcoxon) and PageMan (Ora Fisher, Wilcoxon).
From the total of 25000 genes there are 3000 interesting (threshold >=1, p-val<0,05). If I preselect the 3000 genes via Excel and give them to mapman I get different results compared to the 25000 with the specific filter configurations. I have a lot of preselected genes of interest – lists. What is the correct procedure. Should all my lists be embedded in a whole genome list? And than I run the program with the respective filter configurations?

In my case the Wilcoxon test must use the information of the 22000 other genes (that I wanted to filter out) how does it do this?
What is the problem when I give the program only the 3000 genes?

Thank you in advance for your time.

Hi

please do NOT preselect genes but use the filter.
The reason is that for the statistics (fisher's test at the very least) to work you need to compare your list to a reference set. MapMan PageMan assumes this to be the full set that you provide in your experiment file.
The reason is that regardless of which technology you use, there are always some genes which you might not be able to measuere which depends on your platform and potentially on your pretreament, thus MapMan can not make strong assumptions there.

So to sum it up:
Take th lsit of genes and mark them in different columns accordingly.

The Wilcxon test by the way doesn't use filters at all it proceeds bin-wise and compares the value. E.g. is one bin particularly uprgeulated.

Best Wishes
Björn

Hello Björn,

I was suspecting something like that ^^
Thank you so much for the fast answer.

Best wishes to you too.

Björn G.