GSEA.plot: GSEA.plot performs the GSEA and plots the running sum for one...

In galder-max/GSEAlite: GSEA lite performs simple GSEA analysis with few useful functions

Usage

GSEA.plot(m, geneset, o = NULL, ga = NULL, nperms = 1000, metric = c("SNR", "FC"), col = rgb(0.3, 0.73, 0.3))

Arguments

m	matrix of expressions. columns represent samples, rows are genes.
geneset	vector of genes symbols or ID present in rownames(m) make sure the geneset is not too small (typically 15) or too big (typically 500)
o	2-class factor where length(o) equals ncol(m). If left to NULL, m is considered as a paired difference between two classes.
nperms	number of permutations; the higher, the better the NES estimation; default to 1000
metric	metric used SNR to use Signal-to-Noise (snr.SNR function) FC to use Fold changes (snr.FC function)
col	color of the running sum curve. Matter of taste...

Value

returns NULL value since it is primarily used for the plot

Author(s)

[e-mail]

See Also

GSEA.listofgenes

Examples

## creating synthetic inputs
m<-matrix(rnorm(10*10000),10000,10)
rownames(m)<-paste("r",1:10000)
o<-as.factor(c(rep(1,5),rep(2,5)))

## creating random geneset of size 100
geneset<-paste("r",sample(1:10000,100,rep=F))

## plotting the GSEA running sum
GSEA.plot(m,geneset,o,nperms=100)

## adding some signal to the synthetic data and taking geneset of size
## 100 among the 1000 genes with signal
geneset<-paste("r",sample(1:1000,100,rep=F))
m[1:1000,1:5]<-m[1:1000,1:5]+rnorm(1000*5,mean=2)

## performing the GSEA plot again
##note that the enrichment is negative since class2 is less expressed
##than class 1
GSEA.plot(m,geneset,o,nperms=100)


##PAIRED TEST
##if no factor is provided m is considered as a paired difference
##between class1 & class2
GSEA.plot(m[,1:5]-m[,6:10],geneset,nperms=100)

galder-max/GSEAlite documentation built on Jan. 30, 2021, 10:50 p.m.