Concerning FDR calculation using Limma

Question

[sebastien200109]

I'm new to microarray and I'm trying to analyze several single channel agilent microarray using Limma package.
I have a question concerning the calculation of FDR q value using eBayes/topTable.
If I understood well, q value is an adjustment of p value linked to gene ranking. So, should the q value be calculated on an entire dataset or should it be calculated after filtering (removing non-coding RNA for instance)? As the number of variables will be different, the ranking should also be, will this impact the q value?

Thanks!

[sebastien200109]

Just what I needed to know. Thank!

[Deepak Tanwar]

If your problem got solved, please accept the answer by clicking on the green tick, which only you could see below the answer.

Answer 1

[Devon Ryan]

You typically want to calculate q-values and adjusted p-values (they're different, though related) from your filtered dataset (assuming you're filtering by things like coding status or other factors independent from the p-value).