R/PPIConstruction.R
In fairmiracle/PPINet: Protenin-protein interaction network construction and module identification
#' PPI network from STRING
#'
#' Constructing protenin-protein interaction network from STRING database, when
#' STRING contains gene names from gene expression platform, like Saccharomyces
#' cerevisiae
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param STRINGfilename The file downloaded from STRING, like 4932.protein.links.v10.txt
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param savename The file to save filtered gene symbols and network edgelist
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING PPINet
#'
#'
PPIFromString <- function(pvalues, GeneNames, STRINGfilename, STRINGTheta = 0, savename){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from STRING...\n')
PPI=read.delim(STRINGfilename,sep=' ')
selectedrow = which(PPI[,3] >= STRINGTheta)
PPI = PPI[selectedrow,]
septag = strsplit(STRINGfilename,'.',fixed = TRUE)[[1]][1]
linkssource = gsub(paste(septag,'.',sep=''),'',PPI[,1])
linkstarget = gsub(paste(septag,'.',sep=''),'',PPI[,2])
selectedp = union(linkssource,linkstarget)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = intersect(selectedp,GeneNames)
furtherselectedrows = match(selectedgenes,GeneNames)
GeneNames = GeneNames[furtherselectedrows]
pvalues = pvalues[furtherselectedrows]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=3)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], GeneNames) && is.element(linkstarget[i], GeneNames)){
net[j,1] = match(linkssource[i],GeneNames)
net[j,2] = match(linkstarget[i],GeneNames)
net[j,3] = PPI[i,3]
j = j+1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' PPI network from STRING and Ensembl
#'
#' Constructing protenin-protein interaction network from STRING database, when
#' protein name is needed for mapping with gene name from Ensembl
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param STRINGfilename The file downloaded from STRING, like 9606.protein.links.v10.txt
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param Ensemblfilename The file downloaded from ensembl (Customise your download),
#' like mart_export.txt, first column is Associated Gene Name like MT-ND1,
#' second column is Ensembl Protein ID like ENSP00000354687
#' @param savename The file to save filtered gene symbols and network edgelist
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING Ensembl PPINet
#'
#'
PPIFromStringEnsembl <- function(pvalues, GeneNames, STRINGfilename, STRINGTheta = 0,
Ensemblfilename, savename){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from STRING...\n')
PPI=read.delim(STRINGfilename,sep=' ')
selectedrow = which(PPI[,3] >= STRINGTheta)
PPI = PPI[selectedrow,]
septag = strsplit(STRINGfilename,'.',fixed = TRUE)[[1]][1]
linkssource = gsub(paste(septag,'.',sep=''),'',PPI[,1])
linkstarget = gsub(paste(septag,'.',sep=''),'',PPI[,2])
selectedp = union(linkssource,linkstarget)
## Mapping from Ensembl.Protein.ID ENSPxxxx to Associated.Gene.Name C20orf144
cat('Mapping gene names from Ensembl...\n')
maps = read.delim(Ensemblfilename, sep='\t')
maps = maps[!(is.na(maps[,1])|maps[,2]==""), ]
selectedp = intersect(selectedp,maps[,2])
selectedmapsrow = match(selectedp,maps[,2])
maps = maps[selectedmapsrow,]
interg = intersect(GeneNames,maps[,1])
selectedmapsrow = match(interg,maps[,1])
maps = maps[selectedmapsrow,]
selectedp = maps[,2]
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = match(interg,GeneNames)
GeneNames = GeneNames[selectedgenes]
pvalues = pvalues[selectedgenes]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=3)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], selectedp) && is.element(linkstarget[i], selectedp)){
ipg1 = maps[which(maps[,2] == linkssource[i]),1]
net[j,1] = match(ipg1,GeneNames)
ipg2 = maps[which(maps[,2] == linkstarget[i]),1]
net[j,2] = match(ipg2,GeneNames)
net[j,3] = PPI[i,3]
j = j+1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' PPI network from BioGRID
#'
#' Constructing protenin-protein interaction network from BioGRID database
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param BioGRIDfilename The file downloaded from BioGRID, like
#' BIOGRID-ORGANISM-Homo_sapiens-3.4.138.tab.txt
#' @param savename The file to save filtered gene symbols and network edgelist
#' @param column1 Source column in BioGRID file, for Saccharomyces_cerevisiae
#' column1=1 and for Homo_sapiens column1=3
#' @param column2 Target column in BioGRID file, for Saccharomyces_cerevisiae
#' column2=2 and for Homo_sapiens column3=4
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords BioGRID PPINet
#'
#'
PPIFromBioGRID <- function(pvalues, GeneNames, BioGRIDfilename, savename,
column1,column2){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from BioGRID...\n')
PPI=read.delim(BioGRIDfilename,skip=35)
links = PPI[,c(column1,column2)]
linkssource = PPI[,column1]
linkstarget = PPI[,column2]
selectedp = union(links[,1],links[,2])
selectedgenes = intersect(selectedp,GeneNames)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = match(selectedgenes,GeneNames)
GeneNames = GeneNames[selectedgenes]
pvalues = pvalues[selectedgenes]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=2)
j = 1
# edges <- links[,1] %in% GeneNames & links[,2] %in% GeneNames
# net <- links[edges,]
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], GeneNames) && is.element(linkstarget[i], GeneNames)){
net[j,1] = match(linkssource[i],GeneNames)
net[j,2] = match(linkstarget[i],GeneNames)
j = j + 1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' Dynamic PPI network from BioGRID
#'
#' Constructing Dynamic protenin-protein interaction network from BioGRID database,
#' where the structures of each layer are also determined by vertecies weights.
#'
#' @param datExpr A TxN expression matrix, each column is N-dimensional vector
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param BioGRIDfilename The file downloaded from BioGRID, like
#' BIOGRID-ORGANISM-Homo_sapiens-3.4.138.tab.txt
#' @param column1 Source column in BioGRID file, for Saccharomyces_cerevisiae
#' column1=1 and for Homo_sapiens column1=3
#' @param column2 Target column in BioGRID file, for Saccharomyces_cerevisiae
#' column2=2 and for Homo_sapiens column2=4
#' @param listz a list of node score vectors, each layer has a n-length vector
#' @param pr The proportion of expressed genes selected for each layer
#' @return A list of net from \code{\link{PPIFromBioGRID}}
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords BioGRID PPINet
#'
#'
DynamicPPIFromBioGRID <- function(datExpr, GeneNames, BioGRIDfilename,
column1,column2,listz,pr=0.6){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from BioGRID...\n')
PPI <- read.delim(BioGRIDfilename,skip=35)
links <- PPI[,c(column1,column2)]
linkssource <- PPI[,column1]
linkstarget <- PPI[,column2]
selectedp <- union(links[,1],links[,2])
selectedgenes = intersect(selectedp,GeneNames)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes <- match(selectedgenes,GeneNames)
GeneNames <- GeneNames[selectedgenes]
datExpr <- datExpr[,selectedgenes]
for(iz in 1:length(listz)){
pvalues <- listz[[iz]]
pvalues <- pvalues[selectedgenes]
names(pvalues) <- GeneNames
filteridx <- sort(pvalues,decreasing=TRUE,index.return=TRUE)$ix[1:ceiling(length(pvalues)*pr)]
##Network construction
pvalues <- pvalues[filteridx]
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=2)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], names(pvalues)) && is.element(linkstarget[i], names(pvalues))){
net[j,1] <- match(linkssource[i],names(pvalues))
net[j,2] <- match(linkstarget[i],names(pvalues))
j = j + 1
}
}
net <- net[1:(j-1),]
cat('Saving files...\n')
write.table(pvalues, paste(iz,'_Pvalues.dat',sep=''),row.names=TRUE,
sep=",",col.names=FALSE,quote=FALSE)
write.table(net, paste(iz,'_Net.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
}
#return (list(net = net,p.values = pvalues))
}
#' PPI network simplification
#'
#' Simplify protenin-protein interaction network, for module identification
#'
#' @param net Edgelist from \code{\link{PPIFromString}},
#' \code{\link{PPIFromStringEnsembl}} and \code{\link{PPIFromBioGRID}}
#' @param p.values The p-values vector
#' @param savename Name for saving PPI edge list as plain file and also saving
#' nodes in largest connected component and figure as eps format by default
#'
#' @return a list containing module size, best score, module as a list of nodes
#'
#' @seealso \code{\link{PPIFromString}}, \code{\link{PPIFromStringEnsembl}}
#' and \code{\link{PPIFromBioGRID}}
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING CONSTRUCTION
#'
#' @examples
#' require(genalg)
#' Sp <- NetSimplify(net,0)
#' GA_result <- GA_search_connected(lambda=0.5,Sp$node_score,Sp$edge_score,
#' Sp$Edgelist,num_iter=1000, muCh=0.05, zToR=10, minsize=10)
#'
NetSimplify <- function(net, p.values, STRINGTheta = 0){
if (dim(net)[2] > 2 & STRINGTheta > 0)
net = net[which(net[,3] > STRINGTheta),]
GeneNames = names(p.values)
el = cbind(GeneNames[net[,1]],GeneNames[net[,2]])
#el = apply(net[,1:2], 2, as.character)
require(igraph)
g <- graph.edgelist(el, directed = FALSE)
sg = simplify(g)
comps <- decompose.graph(sg,min.vertices=3)
comps_edgelist = get.edgelist(comps[[1]], names=TRUE)
Escores = numeric(length = dim(comps_edgelist)[1])
if (dim(net)[2] == 3){
edges <- comps_edgelist[,1] %in% GeneNames & comps_edgelist[,2] %in% GeneNames
Escores = net[edges,3]
Escores = Escores/max(Escores)
}
compsids = match(V(comps[[1]])$name,GeneNames)
Nscores = p.values[compsids]
names(Nscores) = GeneNames[compsids]
return (list(node_score = Nscores, edge_score = Escores,
Edgelist = comps_edgelist))
}
#' PPI network visulization
#'
#' Plot protenin-protein interaction network by igraph
#'
#' @param net Edgelist from \code{\link{PPIFromString}},
#' \code{\link{PPIFromStringEnsembl}} and \code{\link{PPIFromBioGRID}}
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param savename Name for saving PPI edge list as plain file and also saving
#' nodes in largest connected component and figure as eps format by default
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING CONSTRUCTION
#'
#'
PPIplot <- function(net, STRINGTheta = 0, savename){
if (dim(net)[2] > 2 & STRINGTheta > 0)
net = net[which(net[,3] > STRINGTheta),1:2]
el = apply(net[,1:2], 2, as.character)
require(igraph)
g <- graph.edgelist(el, directed = FALSE)
sg = simplify(g)
comps <- decompose.graph(sg,min.vertices=3)
comps_edgelist = get.edgelist(sg, names=TRUE)
write.table(comps_edgelist, file = paste(savename,'Edgelist.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
write.table(V(comps[[1]])$name, file = paste(savename,'Compnent.dat',sep=''),
row.names = FALSE, col.names = FALSE, sep="\t", quote = FALSE)
postscript(paste(savename,'.eps',sep=''), fonts=c("serif", "Palatino"))
plot(comps[[1]],edge.width=0.1,vertex.label = NA,vertex.size=0.5)
dev.off()
}
fairmiracle/PPINet documentation built on May 16, 2019, 9:59 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#' PPI network from STRING
#'
#' Constructing protenin-protein interaction network from STRING database, when
#' STRING contains gene names from gene expression platform, like Saccharomyces
#' cerevisiae
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param STRINGfilename The file downloaded from STRING, like 4932.protein.links.v10.txt
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param savename The file to save filtered gene symbols and network edgelist
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING PPINet
#'
#'
PPIFromString <- function(pvalues, GeneNames, STRINGfilename, STRINGTheta = 0, savename){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from STRING...\n')
PPI=read.delim(STRINGfilename,sep=' ')
selectedrow = which(PPI[,3] >= STRINGTheta)
PPI = PPI[selectedrow,]
septag = strsplit(STRINGfilename,'.',fixed = TRUE)[[1]][1]
linkssource = gsub(paste(septag,'.',sep=''),'',PPI[,1])
linkstarget = gsub(paste(septag,'.',sep=''),'',PPI[,2])
selectedp = union(linkssource,linkstarget)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = intersect(selectedp,GeneNames)
furtherselectedrows = match(selectedgenes,GeneNames)
GeneNames = GeneNames[furtherselectedrows]
pvalues = pvalues[furtherselectedrows]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=3)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], GeneNames) && is.element(linkstarget[i], GeneNames)){
net[j,1] = match(linkssource[i],GeneNames)
net[j,2] = match(linkstarget[i],GeneNames)
net[j,3] = PPI[i,3]
j = j+1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' PPI network from STRING and Ensembl
#'
#' Constructing protenin-protein interaction network from STRING database, when
#' protein name is needed for mapping with gene name from Ensembl
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param STRINGfilename The file downloaded from STRING, like 9606.protein.links.v10.txt
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param Ensemblfilename The file downloaded from ensembl (Customise your download),
#' like mart_export.txt, first column is Associated Gene Name like MT-ND1,
#' second column is Ensembl Protein ID like ENSP00000354687
#' @param savename The file to save filtered gene symbols and network edgelist
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING Ensembl PPINet
#'
#'
PPIFromStringEnsembl <- function(pvalues, GeneNames, STRINGfilename, STRINGTheta = 0,
Ensemblfilename, savename){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from STRING...\n')
PPI=read.delim(STRINGfilename,sep=' ')
selectedrow = which(PPI[,3] >= STRINGTheta)
PPI = PPI[selectedrow,]
septag = strsplit(STRINGfilename,'.',fixed = TRUE)[[1]][1]
linkssource = gsub(paste(septag,'.',sep=''),'',PPI[,1])
linkstarget = gsub(paste(septag,'.',sep=''),'',PPI[,2])
selectedp = union(linkssource,linkstarget)
## Mapping from Ensembl.Protein.ID ENSPxxxx to Associated.Gene.Name C20orf144
cat('Mapping gene names from Ensembl...\n')
maps = read.delim(Ensemblfilename, sep='\t')
maps = maps[!(is.na(maps[,1])|maps[,2]==""), ]
selectedp = intersect(selectedp,maps[,2])
selectedmapsrow = match(selectedp,maps[,2])
maps = maps[selectedmapsrow,]
interg = intersect(GeneNames,maps[,1])
selectedmapsrow = match(interg,maps[,1])
maps = maps[selectedmapsrow,]
selectedp = maps[,2]
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = match(interg,GeneNames)
GeneNames = GeneNames[selectedgenes]
pvalues = pvalues[selectedgenes]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=3)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], selectedp) && is.element(linkstarget[i], selectedp)){
ipg1 = maps[which(maps[,2] == linkssource[i]),1]
net[j,1] = match(ipg1,GeneNames)
ipg2 = maps[which(maps[,2] == linkstarget[i]),1]
net[j,2] = match(ipg2,GeneNames)
net[j,3] = PPI[i,3]
j = j+1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' PPI network from BioGRID
#'
#' Constructing protenin-protein interaction network from BioGRID database
#'
#' @param pvalues A N-dimensional vector of gene expression level, processed by limma
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param BioGRIDfilename The file downloaded from BioGRID, like
#' BIOGRID-ORGANISM-Homo_sapiens-3.4.138.tab.txt
#' @param savename The file to save filtered gene symbols and network edgelist
#' @param column1 Source column in BioGRID file, for Saccharomyces_cerevisiae
#' column1=1 and for Homo_sapiens column1=3
#' @param column2 Target column in BioGRID file, for Saccharomyces_cerevisiae
#' column2=2 and for Homo_sapiens column3=4
#'
#' @return net two or three column edge list for \code{\link{NetSimplify}} or
#' \code{\link{PPIplot}} and p-values vector.
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords BioGRID PPINet
#'
#'
PPIFromBioGRID <- function(pvalues, GeneNames, BioGRIDfilename, savename,
column1,column2){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from BioGRID...\n')
PPI=read.delim(BioGRIDfilename,skip=35)
links = PPI[,c(column1,column2)]
linkssource = PPI[,column1]
linkstarget = PPI[,column2]
selectedp = union(links[,1],links[,2])
selectedgenes = intersect(selectedp,GeneNames)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes = match(selectedgenes,GeneNames)
GeneNames = GeneNames[selectedgenes]
pvalues = pvalues[selectedgenes]
##Network construction
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=2)
j = 1
# edges <- links[,1] %in% GeneNames & links[,2] %in% GeneNames
# net <- links[edges,]
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], GeneNames) && is.element(linkstarget[i], GeneNames)){
net[j,1] = match(linkssource[i],GeneNames)
net[j,2] = match(linkstarget[i],GeneNames)
j = j + 1
}
}
net = net[1:(j-1),]
cat('Saving files...\n')
#save(net,GeneNames,pvalues,file = paste(savename,'.RData',sep=''))
write.table(pvalues, paste(savename,'Pvalues.dat',sep=''), row.names = FALSE,
col.names = FALSE, sep="\t")
write.table(GeneNames, paste(savename,'GeneNames.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
names(pvalues) = GeneNames
return (list(net = net,p.values = pvalues))
}
#' Dynamic PPI network from BioGRID
#'
#' Constructing Dynamic protenin-protein interaction network from BioGRID database,
#' where the structures of each layer are also determined by vertecies weights.
#'
#' @param datExpr A TxN expression matrix, each column is N-dimensional vector
#' @param GeneNames A N-dimensional vector of gene symbols, pre-filtering by expression profiles
#' @param BioGRIDfilename The file downloaded from BioGRID, like
#' BIOGRID-ORGANISM-Homo_sapiens-3.4.138.tab.txt
#' @param column1 Source column in BioGRID file, for Saccharomyces_cerevisiae
#' column1=1 and for Homo_sapiens column1=3
#' @param column2 Target column in BioGRID file, for Saccharomyces_cerevisiae
#' column2=2 and for Homo_sapiens column2=4
#' @param listz a list of node score vectors, each layer has a n-length vector
#' @param pr The proportion of expressed genes selected for each layer
#' @return A list of net from \code{\link{PPIFromBioGRID}}
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords BioGRID PPINet
#'
#'
DynamicPPIFromBioGRID <- function(datExpr, GeneNames, BioGRIDfilename,
column1,column2,listz,pr=0.6){
##PPI pairs from STRING, protein name start with ENSPxxxx
cat('Reading edge list from BioGRID...\n')
PPI <- read.delim(BioGRIDfilename,skip=35)
links <- PPI[,c(column1,column2)]
linkssource <- PPI[,column1]
linkstarget <- PPI[,column2]
selectedp <- union(links[,1],links[,2])
selectedgenes = intersect(selectedp,GeneNames)
##Matching gene names for further gene filtering
cat('Filtering genes...\n')
selectedgenes <- match(selectedgenes,GeneNames)
GeneNames <- GeneNames[selectedgenes]
datExpr <- datExpr[,selectedgenes]
for(iz in 1:length(listz)){
pvalues <- listz[[iz]]
pvalues <- pvalues[selectedgenes]
names(pvalues) <- GeneNames
filteridx <- sort(pvalues,decreasing=TRUE,index.return=TRUE)$ix[1:ceiling(length(pvalues)*pr)]
##Network construction
pvalues <- pvalues[filteridx]
cat('Matching gene names and constructing...\n')
net = matrix(0,nrow = dim(PPI)[1], ncol=2)
j = 1
for (i in 1:dim(PPI)[1]) {
if (is.element(linkssource[i], names(pvalues)) && is.element(linkstarget[i], names(pvalues))){
net[j,1] <- match(linkssource[i],names(pvalues))
net[j,2] <- match(linkstarget[i],names(pvalues))
j = j + 1
}
}
net <- net[1:(j-1),]
cat('Saving files...\n')
write.table(pvalues, paste(iz,'_Pvalues.dat',sep=''),row.names=TRUE,
sep=",",col.names=FALSE,quote=FALSE)
write.table(net, paste(iz,'_Net.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
}
#return (list(net = net,p.values = pvalues))
}
#' PPI network simplification
#'
#' Simplify protenin-protein interaction network, for module identification
#'
#' @param net Edgelist from \code{\link{PPIFromString}},
#' \code{\link{PPIFromStringEnsembl}} and \code{\link{PPIFromBioGRID}}
#' @param p.values The p-values vector
#' @param savename Name for saving PPI edge list as plain file and also saving
#' nodes in largest connected component and figure as eps format by default
#'
#' @return a list containing module size, best score, module as a list of nodes
#'
#' @seealso \code{\link{PPIFromString}}, \code{\link{PPIFromStringEnsembl}}
#' and \code{\link{PPIFromBioGRID}}
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING CONSTRUCTION
#'
#' @examples
#' require(genalg)
#' Sp <- NetSimplify(net,0)
#' GA_result <- GA_search_connected(lambda=0.5,Sp$node_score,Sp$edge_score,
#' Sp$Edgelist,num_iter=1000, muCh=0.05, zToR=10, minsize=10)
#'
NetSimplify <- function(net, p.values, STRINGTheta = 0){
if (dim(net)[2] > 2 & STRINGTheta > 0)
net = net[which(net[,3] > STRINGTheta),]
GeneNames = names(p.values)
el = cbind(GeneNames[net[,1]],GeneNames[net[,2]])
#el = apply(net[,1:2], 2, as.character)
require(igraph)
g <- graph.edgelist(el, directed = FALSE)
sg = simplify(g)
comps <- decompose.graph(sg,min.vertices=3)
comps_edgelist = get.edgelist(comps[[1]], names=TRUE)
Escores = numeric(length = dim(comps_edgelist)[1])
if (dim(net)[2] == 3){
edges <- comps_edgelist[,1] %in% GeneNames & comps_edgelist[,2] %in% GeneNames
Escores = net[edges,3]
Escores = Escores/max(Escores)
}
compsids = match(V(comps[[1]])$name,GeneNames)
Nscores = p.values[compsids]
names(Nscores) = GeneNames[compsids]
return (list(node_score = Nscores, edge_score = Escores,
Edgelist = comps_edgelist))
}
#' PPI network visulization
#'
#' Plot protenin-protein interaction network by igraph
#'
#' @param net Edgelist from \code{\link{PPIFromString}},
#' \code{\link{PPIFromStringEnsembl}} and \code{\link{PPIFromBioGRID}}
#' @param STRINGTheta The threshold for PPI edges selections when using STRING
#' @param savename Name for saving PPI edge list as plain file and also saving
#' nodes in largest connected component and figure as eps format by default
#'
#' @author Dong Li, \email{dxl466@cs.bham.ac.uk}
#' @keywords STRING CONSTRUCTION
#'
#'
PPIplot <- function(net, STRINGTheta = 0, savename){
if (dim(net)[2] > 2 & STRINGTheta > 0)
net = net[which(net[,3] > STRINGTheta),1:2]
el = apply(net[,1:2], 2, as.character)
require(igraph)
g <- graph.edgelist(el, directed = FALSE)
sg = simplify(g)
comps <- decompose.graph(sg,min.vertices=3)
comps_edgelist = get.edgelist(sg, names=TRUE)
write.table(comps_edgelist, file = paste(savename,'Edgelist.dat',sep=''),
row.names = FALSE, col.names = FALSE,sep="\t",quote = FALSE)
write.table(V(comps[[1]])$name, file = paste(savename,'Compnent.dat',sep=''),
row.names = FALSE, col.names = FALSE, sep="\t", quote = FALSE)
postscript(paste(savename,'.eps',sep=''), fonts=c("serif", "Palatino"))
plot(comps[[1]],edge.width=0.1,vertex.label = NA,vertex.size=0.5)
dev.off()
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.