Search
Question: limma; Only fit list of certain genes/ filter eSet
0
23 months ago by
bi_Scholar0 wrote:

Hello,

I got a list of genes (<1000) for which I want to perform DEG-Analysis. Since I'm only interested in certain genes, I want to filter out the rest to reduce false-negative results and get more information about the genes I'm actually interested in.

Is there a simple way to do this with limma, or can I somehow filter genes from a expression set? If I try to fit a exprs-Matrix instead of a exprs-Set, I'm excluding genenames, GO annotations, etc. which is very unpleasant.

modified 23 months ago by Gordon Smyth35k • written 23 months ago by bi_Scholar0
4
23 months ago by
Gordon Smyth35k
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
Gordon Smyth35k wrote:

Yes, there's a simple way to do it. Just do the usual linear model analysis using all the expressed genes, not just your focus set of interest. Then make the topTable just for your set of interest:

topTable( fit[mygenes, ], coef=whatever)

where mygenes selects your genes of interest and "whatever" is the contrast you want to test for. mygenes can be a vector of row names, or a TRUE/FALSE vector, or a vector of row numbers.