Question

How to return probelist,FC and p-value column by putting a p-value threshold in Limma

0

Entering edit mode

Agaz Hussain Wani ▴ 260

@agaz-hussain-wani-7620

Last seen 6.1 years ago

India

Hi,

I am using limma package to perform differential expression of Illumina data. I am wondering if i can get columns of probelist, foldchange and p-value by putting a threshold of say p-value <0.01. I can get probelist by using

sigGene <- probeList[ fit2$p.value[,2] < 0.01]

which returns the probelist only, however i am interested to get all the three columns mentioned above where the p-values is of my interest. Can i get this by using limma. Thanks

r limma differential expression • 1.2k views

ADD COMMENT • link updated 9.0 years ago by Gordon Smyth 50k • written 9.0 years ago by Agaz Hussain Wani ▴ 260

score 1 · Answer 1 · 2015-06-07

1

Entering edit mode

Gordon Smyth 50k

@gordon-smyth

Last seen 10 hours ago

WEHI, Melbourne, Australia

Yes, the topTable() function does this. For example

topTable(fit2, coef=2, n=Inf, p=0.01, adjust.method="none")

However I would strongly suggest that you consider using a multiple testing adjustment as well. Benjami-Hochberg is the default.

ADD COMMENT • link 9.0 years ago Gordon Smyth 50k

0

Entering edit mode

Thanks for your answer.

ADD REPLY • link 9.0 years ago Agaz Hussain Wani ▴ 260

0

Entering edit mode

Is there something to set the value for number of studies to search in for significant gene. Say i have 12 studies with two treatment conditions, can i find a gene which is significant at-least in four studies or so depending on the p-value.

ADD REPLY • link 9.0 years ago Agaz Hussain Wani ▴ 260

0

Entering edit mode

There's nothing off the shelf to do this, and it's not trivial to implement. The multiple testing situation is even more complicated here, as you'd have to account for multiple studies as well as multiple genes.

ADD REPLY • link 8.9 years ago Aaron Lun ★ 28k