spia {SPIA}R Documentation

Signaling Pathway Impact Analysis (SPIA) based on over-representation and signaling perturbations accumulation

Description

This function implements the SPIA algorithm to analyse KEGG signaling pathways.

Usage

spia(de=NULL,all=NULL,organism="hsa",nB=2000,plots=FALSE,verbose=TRUE,beta=NULL)

Arguments

de A named vector containing log2 fold-changes of the differentially expressed genes. The names of this numeric vector are Entrez gene IDs.
all A vector with the Entrez IDs in the reference set. If the data was obtained from a microarray experiment, this set will contain all genes present on the specific array used for the experiment. This vector should contain all names of the de argument.
organism A three letter character designating the organism. See a full list at ftp://ftp.genome.jp/pub/kegg/xml/organisms .
nB Number of bootstrap iterations used to compute the P PERT value. Should be larger than 100. A recommended value is 2000.
plots If set to TRUE, the function plots the gene perturbation accumulation vs log2 fold change for every gene on each pathway. The null distribution of the total net accumulations from which PPERT is computed, is plotted as well. The figures are sent to the SPIAPerturbationPlots.pdf file in the current directory.
verbose If set to TRUE, displays the number of pathways already analyzed.
beta Weights to be assigned to each type of gene/protein relation type. It should be a named numeric vector of length 23, whose names must be: c("activation","compound","binding/association","expression","inhibition","activation_phosphorylation","phosphorylation", "indirect","inhibition_phosphorylation","dephosphorylation_inhibition","dissociation","dephosphorylation","activation_dephosphorylation", "state","activation_indirect","inhibition_ubiquination","ubiquination","expression_indirect","indirect_inhibition","repression", "binding/association_phosphorylation","dissociation_phosphorylation","indirect_phosphorylation")
If set to null, beta will be by default chosen as: c(1,0,0,1,-1,1,0,0,-1,-1,0,0,1,0,1,-1,0,1,-1,-1,0,0,0).

Details

See cited documents for more details.

Value

A data frame containing the ranked pathways and various statistics: pSize is the number of genes on the pathway; NDE is the number of DE genes per pathway; tA is the observed total preturbation accumulation in the pathway; pNDE is the probability to observe at least NDE genes on the pathway using a hypergeometric model; pPERT is the probability to observe a total accumulation more extreme than tA only by chance; pG is the p-value obtained by combining pNDE and pPERT; pGFdr and pGFWER are the False Discovery Rate and respectively Bonferroni adjusted global p-values; and the Status gives the direction in which the pathway is perturbed (activated or inhibited).

Author(s)

Adi Laurentiu Tarca <atarca@med.wayne.edu>, Purvesh Khatri, Sorin Draghici

References

Adi L. Tarca, Sorin Draghici, Purvesh Khatri, et. al, A Signaling Pathway Impact Analysis for Microarray Experiments, 2008, Bioinformatics, 2009, 25(1):75-82.

Purvesh Khatri, Sorin Draghici, Adi L. Tarca, Sonia S. Hassan, Roberto Romero. A system biology approach for the steady-state analysis of gene signaling networks. Progress in Pattern Recognition, Image Analysis and Applications, Lecture Notes in Computer Science. 4756:32-41, November 2007.

Draghici, S., Khatri, P., Tarca, A.L., Amin, K., Done, A., Voichita, C., Georgescu, C., Romero, R.: A systems biology approach for pathway level analysis. Genome Research, 17, 2007.

See Also

plotP

Examples

# Example using a colorectal cancer dataset obtained using Affymetrix geneChip technology (GEO GSE4107).
# Suppose that proper preprocessing was performed and a two group moderated t-test was applied. The topTable 
# result from limma package for this data set is called "top".
#The following lines will annotate each probeset to an entrez ID identifier, will keep the most significant probeset for each 
#gene ID and retain those with FDR<0.05 as differentially expressed.
#You can run these lines if hgu133plus2.db package is available

#data(colorectalcancer)
#x <- hgu133plus2ENTREZID 
#top$ENTREZ<-unlist(as.list(x[top$ID]))
#top<-top[!is.na(top$ENTREZ),]
#top<-top[!duplicated(top$ENTREZ),]
#tg1<-top[top$adj.P.Val<0.05,]
#DE_Colorectal=tg1$logFC
#names(DE_Colorectal)<-as.vector(tg1$ENTREZ)
#ALL_Colorectal=top$ENTREZ

data(colorectalcancer)

# pathway analysis using SPIA; # use nB=2000 or higher for more accurate results
res<-spia(de=DE_Colorectal, all=ALL_Colorectal, organism="hsa",beta=NULL,nB=200,plots=FALSE, verbose=TRUE)
res
# Create the evidence plot
plotP(res)


[Package SPIA version 1.0.0 Index]