GSReg (version 1.6.0)

GSReg.GeneSets.DIRAC: Performs DIRAC for gene set analysis from the paper Eddy et al (2010).

Description

GSReg.GeneSets.DIRAC performs DIRAC for gene set analysis from the paper Eddy et al (2010). In fact, the Null hypothesis is that the conservation index is not significantly different under two phenotypes. The function calculates the p-value using permutation test; hence, extremely low p-value cannot be reached.

Usage

GSReg.GeneSets.DIRAC(geneexpres, pathways, phenotypes, Nperm = 1000,minGeneNum=5)

Arguments

geneexpres

the matrix of gene expressions. The rownames must represent gene names and the columns represent samples. There must not be any missing values. Please use imputation or remove the genes with missing values.

pathways

a list containing pathway information. Each element represents a pathway as a character vector. The genes shown in the pathway must be present in geneexpres. Please prune the genes in pathways using GSReg.Prune before applying this function.

phenotypes

a binary factor containing the phenotypes for samples in geneexpres; hence, the column number of geneexpres and the length of phenotypes must be equal.

Nperm

The number of permutation tests required for p-value calculation. If Nperm==0 then no p-value is calculated.

minGeneNum

the minimum number of genes required in a pathway.

Value

IThe output is a list with three elements. Each element of the output list is a vector are named according to the pathway.

$mu1

a vector containing the variability in DIRAC sense (1- conservation indices in Eddy et al (2010) paper) for all pathways in phenotype == levels(phenotypes)[1].

$mu2

a vector containing the variability in DIRAC sense (1- conservation indices in Eddy et al (2010) paper) for all pathways in phenotype == levels(phenotypes)[2].

$pvalues

a vector containing p-values for each pathway. Low p-values means that the gene expressions have different orderings under different phenotypes.

References

Eddy, James A., et al. "Identifying tightly regulated and variably expressed networks by Differential Rank Conservation (DIRAC)." PLoS computational biology 6.5 (2010): e1000792.

See Also

GSReg.GeneSet.VReg

Examples

library(GSBenchMark)
### loading and pruning the pathways
data(diracpathways)
### loading the data
data(leukemia_GSEA)


### extracting gene names
genenames = rownames(exprsdata);


### DIRAC analysis
DIRAna = GSReg.GeneSets.DIRAC(pathways=diracpathways,geneexpres=exprsdata,Nperm=10,phenotypes=phenotypes)
dysregulatedpathways = rbind(DIRAna$mu1[which(DIRAna$pvalues<0.05)],
DIRAna$mu2[which(DIRAna$pvalues<0.05)],DIRAna$pvalues[which(DIRAna$pvalues<0.05)]);
rownames(dysregulatedpathways)<-c("mu1","mu2","pvalues");
print(dysregulatedpathways[,1:5])
plot(x=dysregulatedpathways["mu1",],y=dysregulatedpathways["mu2",],
xlim=range(dysregulatedpathways[1:2,]),ylim=range(dysregulatedpathways[1:2,]))
lines(x=c(min(dysregulatedpathways[1:2,]),max(dysregulatedpathways[1:2,])),
y=c(min(dysregulatedpathways[1:2,]),max(dysregulatedpathways[1:2,])),type="l")