R/topHundredNetwork.R
In uc-bd2k/SigNetA:
#' Top Hundred Network Function
#'
#' This function allows you to analyze top 100 genes from a signature using the interactome dataset
#' @param File ? Defaults to NULL.
#' @keywords topHundredNetwork
#' @export
#' @examples
#' topHundredNetwork() will analyze network for vorinostat signature from ilincs
topHundredNetwork<-function(File=NULL,phy=FALSE,layOut=1,package=FALSE,nodeGoData=NULL,edgeGoData=NULL,proteinN=1){
sortedFile<-sortNetwork(File)
logic<-sortedFile
##STRING NETWORK ONE(GET MODULE)##
if(proteinN==1){
####Modified function used#######
ppi<-igraph::simplify(ppiGW,remove.loops = TRUE,remove.multiple = FALSE)
###Identify module using FastHeinz algorithm, nsize is fixed to 30 nodes
# names(pval)<-logic$GeneID
#module=modules_RWR_TopScores(subnet=ppi, data_vector=pval, damping_factor=0.8, nseeds=10)
module <- subNetwork(as.character(logic$GeneID), ppi,neighbors = "none")
#CONSTRUCT module only dataframe##
modIds<-V(module)$name
geninfo<- geneData[which(as.character(geneData$GeneID) %in% as.character(modIds)),]
geneLabels<-apply(geninfo,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
geninfo$diffExp<-NA
geninfo$pval<-NA
for(i in 1:nrow(geninfo)){
if(geninfo$GeneID[i] %in% logic$GeneID){
geninfo$diffExp[i]<-logic$coefficients[match(geninfo$GeneID[i],logic$GeneID)]
geninfo$pval[i]<-logic$Pvals[match(geninfo$GeneID[i],logic$GeneID)]
}
}
logFC<-as.numeric(geninfo$diffExp)
names(logFC)<-geneLabels
pval<-as.numeric(geninfo$pval)
pval<- -log10(pval)
names(pval)<-geneLabels
#END of CONSTRUCT module only dataframe##
colorNet<-plotmodule2(module, scores = pval, diff.expr = logFC)
colorNetframe<-data.frame(V(colorNet$n)$name,colorNet$sc,colorNet$d)
module<-igraph.to.graphNEL(colorNet$n) #STRING
# dev.off()
}
else if(proteinN==2){
# ppi<- rmSelfLoops(interactome)
ppi<-interactome
#ppi=decompose.graph(ppi)[[1]] #get the largest subgraph
###Identify module using FastHeinz algorithm, nsize is fixed to 30 nodes
# names(pval)<-logic$GeneID
module <- subNetwork(as.character(logic$GeneID), ppi,neighbors = "none")
modIds<-V(module)$name
geninfo<- geneData[which(as.character(geneData$GeneID) %in% as.character(modIds)),]
geneLabels<-apply(geninfo,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
geninfo$diffExp<-NA
geninfo$pval<-NA
for(i in 1:nrow(geninfo)){
if(geninfo$GeneID[i] %in% logic$GeneID){
geninfo$diffExp[i]<-logic$coefficients[match(geninfo$GeneID[i],logic$GeneID)]
geninfo$pval[i]<-logic$Pvals[match(geninfo$GeneID[i],logic$GeneID)]
}
}
logFC<-as.numeric(geninfo$diffExp)
names(logFC)<-geneLabels
pval<-as.numeric(geninfo$pval)
pval<- -log10(pval)
names(pval)<-geneLabels
colorNet<-plotmodule2(module, scores = pval, diff.expr = logFC)
module<-igraph.to.graphNEL(colorNet$n) #STRING
# dev.off()
}
##END...STRING NETWORK ONE(GET MODULE)##
##END...STRING NETWORK ONE(GET MODULE##
## IGRAPH LAYOUTS
if(layOut=="1"){
l<-layout_with_fr(colorNet$n)
visLay<-"layout_with_fr"
}
else if(layOut=="3"){
l<-layout_on_grid(colorNet$n)
visLay<-"layout_on_grid"
}
else if(layOut=="4"){
l<-layout_with_kk(colorNet$n)
visLay<-"layout_with_kk"
}
else if(layOut=="5"){
l<-layout_on_sphere(colorNet$n)
visLay<-"layout_on_sphere"
}
else if(layOut=="6"){
l<-layout_with_graphopt(colorNet$n)
visLay<-"layout_with_graphopt"
}
###dev.off();
conNodes<-function(x){
if(ltn[x]>0)
id[x]
}
id <- nodes(module) #interactome
ltn<-unlist(lapply(edgeL(module),function(x) length(x[[1]]))) #interactome
modNodes<-unlist(lapply(1:length(ltn),conNodes )) # only get connected nodes and no single nodes
id<-modNodes
#TEST#
id <- nodes(module)
name <- id
label<-id #visNetwork and interactome
#END TEST#
geninfo2<- geneData[which(geneData$GeneID%in%as.character(id)),]
name<-apply(geninfo2,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))#STRING
label<-apply(geninfo2,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
nodeData <- data.frame(id, name, stringsAsFactors=FALSE) ##RCYTOSCAPEJS2
nodeVisData<-data.frame(id,label,stringsAsFactors = FALSE)
id<-nodes(module)
for(i in 1:nrow(geninfo2)){
if(geninfo2$GeneID[i] %in% logic$GeneID){
colorNet$d[i]<-logic$coefficients[match(geninfo2$GeneID[i],logic$GeneID)]
colorNet$sc[i]<-(-log10(logic$Pvals[match(geninfo2$GeneID[i],logic$GeneID)] ))
}
}
nodeData$color<- rep("#00FF0F",nrow(nodeData)) #changed color of nodes
nodeData$shape <- "ellipse" #default shape
nodeData$href <- paste0("http://www.ncbi.nlm.nih.gov/gene/",gsub("[\\(\\)]", "", regmatches(nodeData$name, gregexpr("\\(.*?\\)", nodeData$name))))
nodeData$geneID<-gsub("[\\(\\)]", "", regmatches(nodeData$name, gregexpr("\\(.*?\\)", nodeData$name)))
nodeData$name<-sub(" *\\(.*", "", nodeData$name)
nodeData$Diff_Exp="none"
nodeData$score="none"
nodeData$x="none" #x and y are the columns required for manual layouts. Also, use "preset" as layout mode in rcytoscapejs2.R
nodeData$y="none"
for(i in 1:length(name)){
nodeData[i,3]<-colorNet$c[i];
nodeData[i,7]<-colorNet$d[i]
nodeData[i,8]<-colorNet$sc[i]
nodeData[i,9]<-(l[i,1]+1)*100
nodeData[i,10]<-(l[i,2] +1)*100
}
statNet<<-nodeData
#statNet$df_data<<-nodeData
source<-unlist(lapply(1:length(ltn),function(x) rep(id[x],ltn[x])))
target<-unlist(lapply(edgeL(module), function(x) id[unlist(x)]))
vect<-c()
for(i in 1:length(target)) #extracting the value from the key value pair
vect[i]<-target[[i]]
edgeData <- data.frame(source, target, stringsAsFactors=FALSE)
##VisNetwork###########
nodeVisData$color.background<-rep("blue",nrow(nodeData))
nodeVisData$borderWidth <- 2
nodeVisData$color.border <- "black"
nodeVisData$title<-"<p>Hello world</p>"
nodeVisData$shape<-"dot"
nodeVisData$size<-0
nodeVisData$id<-sub(" *\\(.*", "", nodeVisData$id)
nodeVisData$label<-sub(" *\\(.*", "", nodeVisData$label)
normalize <- function(x) {
#return (((x*2 - min(x)) / (max(x) - min(x)))*50)
return ( 20*((x - min(x)) / (max(x) - min(x))) + 30 )
}
for(i in 1:length(label)){
nodeVisData[i,3]<-colorNet$c[i];
nodeVisData[i,6]<-paste0("<p><b>Gene name:</b>",statNet$name[i],"</p><br><p><b>Gene ID:</b>",statNet$geneID[i],"</p><br><p><b>Differential Expression:</b>",statNet$Diff_Exp[i],"</p><br><p><b>Score:</b>",statNet$score[i],"</p><p><b>NCBI link:</b><a href='",statNet$href[i],"' target='_blank'>",statNet$href[i],"</a></p>")
# nodeVisData[i,8]<-(scoreLength - colorNet$sc[i])
nodeVisData[i,8]<-colorNet$sc[i]
}
nodeVisData<-data.frame(nodeVisData[1:7], apply(nodeVisData["size"],2, normalize) )
for( l in 1:length(nodeVisData$size)){
if(nodeVisData$size[l]<1)
{
nodeVisData$size[l]<-1
}
}
ltn<-unlist(lapply(edgeL(module),function(x) length(x[[1]])))
sourceVis<-unlist(lapply(1:length(ltn),function(x) rep(id[x],ltn[x])))
targetVis<-unlist(lapply(edgeL(module), function(x) id[unlist(x)]))
vect<-c()
for(i in 1:length(targetVis)) #extracting the value from the key value pair
vect[i]<-targetVis[[i]]
sourceVis<-sub(" *\\(.*", "", sourceVis)
targetVis<-sub(" *\\(.*", "", targetVis)
edgeVisData <- data.frame(from=sourceVis, to=targetVis, stringsAsFactors=FALSE)
for (i in 1:nrow(edgeVisData))
{
edgeVisData[i, ] = sort(edgeVisData[i, ])
}
edgeVisData = edgeVisData[!duplicated(edgeVisData),]
##STRING TWO(ADDING EDGE VALUES)#
if(proteinN=="1"){
edgeVisData$title<-"ppi"
# edgeVisDataMod<-merge(edgeVisData,s,by.x = c("from","to"),by.y =c("a","b"),all.x = TRUE)
edgeVisDataMod<-igraph::as_data_frame(igraph.from.graphNEL(module),what="edges")
for(i in 1:nrow(edgeVisDataMod)){
edgeVisDataMod$title[i]<-paste0("<p><b>Neighborhood score:</b>",edgeVisDataMod$f.neighborhood[i],"</p><b>Fusion score:</b>",edgeVisDataMod$f.fusion[i],"</p><b>Cooccurence score:</b>",edgeVisDataMod$f.cooccurence[i],"</p><b>Coexpression score:</b>",edgeVisDataMod$f.coexpression[i],"</p><b>Experimental score:</b>",edgeVisDataMod$f.experimental[i],"</p><b>Database score:</b>",edgeVisDataMod$f.database[i],"</p><b>Textmining score:</b>",edgeVisDataMod$f.textmining[i],"</p><b>Combined score:</b>",edgeVisDataMod$f.combined_score[i],"</p>")
}
edgeVisData<-edgeVisDataMod
}
##END...STRING TWO (ADDING EDGE VALUES)##
if(is.null(nodeGoData) & is.null(edgeGoData))
{
visObj<- visNetwork(nodeVisData, edgeVisData,height="800px",width="900px")
}
else{
visObj<- visNetwork(nodeGoData, edgeGoData,height="800px",width="900px")
# visObj<-enrichObj
}
visObj<-visExport(visObj,type ="png", name="network",float="right")
if(phy){
#print("physics active")
}
else{
visObj<-visIgraphLayout(visObj,layout=visLay)
}
if(package==TRUE)
{
visNetwork(nodeVisData, edgeVisData,height="800px",width="900px")
}
else{
visObj<-visInteraction(visObj,navigationButtons = TRUE)
visObj<-visOptions(visObj,manipulation = TRUE)
}
return(list("networkObj"=visObj,"networkData"=statNet,"edgeData"=edgeVisData,"nodeData"=nodeVisData))
}
uc-bd2k/SigNetA documentation built on June 7, 2019, 7:55 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.
#' Top Hundred Network Function
#'
#' This function allows you to analyze top 100 genes from a signature using the interactome dataset
#' @param File ? Defaults to NULL.
#' @keywords topHundredNetwork
#' @export
#' @examples
#' topHundredNetwork() will analyze network for vorinostat signature from ilincs
topHundredNetwork<-function(File=NULL,phy=FALSE,layOut=1,package=FALSE,nodeGoData=NULL,edgeGoData=NULL,proteinN=1){
sortedFile<-sortNetwork(File)
logic<-sortedFile
##STRING NETWORK ONE(GET MODULE)##
if(proteinN==1){
####Modified function used#######
ppi<-igraph::simplify(ppiGW,remove.loops = TRUE,remove.multiple = FALSE)
###Identify module using FastHeinz algorithm, nsize is fixed to 30 nodes
# names(pval)<-logic$GeneID
#module=modules_RWR_TopScores(subnet=ppi, data_vector=pval, damping_factor=0.8, nseeds=10)
module <- subNetwork(as.character(logic$GeneID), ppi,neighbors = "none")
#CONSTRUCT module only dataframe##
modIds<-V(module)$name
geninfo<- geneData[which(as.character(geneData$GeneID) %in% as.character(modIds)),]
geneLabels<-apply(geninfo,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
geninfo$diffExp<-NA
geninfo$pval<-NA
for(i in 1:nrow(geninfo)){
if(geninfo$GeneID[i] %in% logic$GeneID){
geninfo$diffExp[i]<-logic$coefficients[match(geninfo$GeneID[i],logic$GeneID)]
geninfo$pval[i]<-logic$Pvals[match(geninfo$GeneID[i],logic$GeneID)]
}
}
logFC<-as.numeric(geninfo$diffExp)
names(logFC)<-geneLabels
pval<-as.numeric(geninfo$pval)
pval<- -log10(pval)
names(pval)<-geneLabels
#END of CONSTRUCT module only dataframe##
colorNet<-plotmodule2(module, scores = pval, diff.expr = logFC)
colorNetframe<-data.frame(V(colorNet$n)$name,colorNet$sc,colorNet$d)
module<-igraph.to.graphNEL(colorNet$n) #STRING
# dev.off()
}
else if(proteinN==2){
# ppi<- rmSelfLoops(interactome)
ppi<-interactome
#ppi=decompose.graph(ppi)[[1]] #get the largest subgraph
###Identify module using FastHeinz algorithm, nsize is fixed to 30 nodes
# names(pval)<-logic$GeneID
module <- subNetwork(as.character(logic$GeneID), ppi,neighbors = "none")
modIds<-V(module)$name
geninfo<- geneData[which(as.character(geneData$GeneID) %in% as.character(modIds)),]
geneLabels<-apply(geninfo,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
geninfo$diffExp<-NA
geninfo$pval<-NA
for(i in 1:nrow(geninfo)){
if(geninfo$GeneID[i] %in% logic$GeneID){
geninfo$diffExp[i]<-logic$coefficients[match(geninfo$GeneID[i],logic$GeneID)]
geninfo$pval[i]<-logic$Pvals[match(geninfo$GeneID[i],logic$GeneID)]
}
}
logFC<-as.numeric(geninfo$diffExp)
names(logFC)<-geneLabels
pval<-as.numeric(geninfo$pval)
pval<- -log10(pval)
names(pval)<-geneLabels
colorNet<-plotmodule2(module, scores = pval, diff.expr = logFC)
module<-igraph.to.graphNEL(colorNet$n) #STRING
# dev.off()
}
##END...STRING NETWORK ONE(GET MODULE)##
##END...STRING NETWORK ONE(GET MODULE##
## IGRAPH LAYOUTS
if(layOut=="1"){
l<-layout_with_fr(colorNet$n)
visLay<-"layout_with_fr"
}
else if(layOut=="3"){
l<-layout_on_grid(colorNet$n)
visLay<-"layout_on_grid"
}
else if(layOut=="4"){
l<-layout_with_kk(colorNet$n)
visLay<-"layout_with_kk"
}
else if(layOut=="5"){
l<-layout_on_sphere(colorNet$n)
visLay<-"layout_on_sphere"
}
else if(layOut=="6"){
l<-layout_with_graphopt(colorNet$n)
visLay<-"layout_with_graphopt"
}
###dev.off();
conNodes<-function(x){
if(ltn[x]>0)
id[x]
}
id <- nodes(module) #interactome
ltn<-unlist(lapply(edgeL(module),function(x) length(x[[1]]))) #interactome
modNodes<-unlist(lapply(1:length(ltn),conNodes )) # only get connected nodes and no single nodes
id<-modNodes
#TEST#
id <- nodes(module)
name <- id
label<-id #visNetwork and interactome
#END TEST#
geninfo2<- geneData[which(geneData$GeneID%in%as.character(id)),]
name<-apply(geninfo2,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))#STRING
label<-apply(geninfo2,1,function(x) paste(x[2],"(",as.integer(x[1]),")",sep=""))
nodeData <- data.frame(id, name, stringsAsFactors=FALSE) ##RCYTOSCAPEJS2
nodeVisData<-data.frame(id,label,stringsAsFactors = FALSE)
id<-nodes(module)
for(i in 1:nrow(geninfo2)){
if(geninfo2$GeneID[i] %in% logic$GeneID){
colorNet$d[i]<-logic$coefficients[match(geninfo2$GeneID[i],logic$GeneID)]
colorNet$sc[i]<-(-log10(logic$Pvals[match(geninfo2$GeneID[i],logic$GeneID)] ))
}
}
nodeData$color<- rep("#00FF0F",nrow(nodeData)) #changed color of nodes
nodeData$shape <- "ellipse" #default shape
nodeData$href <- paste0("http://www.ncbi.nlm.nih.gov/gene/",gsub("[\\(\\)]", "", regmatches(nodeData$name, gregexpr("\\(.*?\\)", nodeData$name))))
nodeData$geneID<-gsub("[\\(\\)]", "", regmatches(nodeData$name, gregexpr("\\(.*?\\)", nodeData$name)))
nodeData$name<-sub(" *\\(.*", "", nodeData$name)
nodeData$Diff_Exp="none"
nodeData$score="none"
nodeData$x="none" #x and y are the columns required for manual layouts. Also, use "preset" as layout mode in rcytoscapejs2.R
nodeData$y="none"
for(i in 1:length(name)){
nodeData[i,3]<-colorNet$c[i];
nodeData[i,7]<-colorNet$d[i]
nodeData[i,8]<-colorNet$sc[i]
nodeData[i,9]<-(l[i,1]+1)*100
nodeData[i,10]<-(l[i,2] +1)*100
}
statNet<<-nodeData
#statNet$df_data<<-nodeData
source<-unlist(lapply(1:length(ltn),function(x) rep(id[x],ltn[x])))
target<-unlist(lapply(edgeL(module), function(x) id[unlist(x)]))
vect<-c()
for(i in 1:length(target)) #extracting the value from the key value pair
vect[i]<-target[[i]]
edgeData <- data.frame(source, target, stringsAsFactors=FALSE)
##VisNetwork###########
nodeVisData$color.background<-rep("blue",nrow(nodeData))
nodeVisData$borderWidth <- 2
nodeVisData$color.border <- "black"
nodeVisData$title<-"<p>Hello world</p>"
nodeVisData$shape<-"dot"
nodeVisData$size<-0
nodeVisData$id<-sub(" *\\(.*", "", nodeVisData$id)
nodeVisData$label<-sub(" *\\(.*", "", nodeVisData$label)
normalize <- function(x) {
#return (((x*2 - min(x)) / (max(x) - min(x)))*50)
return ( 20*((x - min(x)) / (max(x) - min(x))) + 30 )
}
for(i in 1:length(label)){
nodeVisData[i,3]<-colorNet$c[i];
nodeVisData[i,6]<-paste0("<p><b>Gene name:</b>",statNet$name[i],"</p><br><p><b>Gene ID:</b>",statNet$geneID[i],"</p><br><p><b>Differential Expression:</b>",statNet$Diff_Exp[i],"</p><br><p><b>Score:</b>",statNet$score[i],"</p><p><b>NCBI link:</b><a href='",statNet$href[i],"' target='_blank'>",statNet$href[i],"</a></p>")
# nodeVisData[i,8]<-(scoreLength - colorNet$sc[i])
nodeVisData[i,8]<-colorNet$sc[i]
}
nodeVisData<-data.frame(nodeVisData[1:7], apply(nodeVisData["size"],2, normalize) )
for( l in 1:length(nodeVisData$size)){
if(nodeVisData$size[l]<1)
{
nodeVisData$size[l]<-1
}
}
ltn<-unlist(lapply(edgeL(module),function(x) length(x[[1]])))
sourceVis<-unlist(lapply(1:length(ltn),function(x) rep(id[x],ltn[x])))
targetVis<-unlist(lapply(edgeL(module), function(x) id[unlist(x)]))
vect<-c()
for(i in 1:length(targetVis)) #extracting the value from the key value pair
vect[i]<-targetVis[[i]]
sourceVis<-sub(" *\\(.*", "", sourceVis)
targetVis<-sub(" *\\(.*", "", targetVis)
edgeVisData <- data.frame(from=sourceVis, to=targetVis, stringsAsFactors=FALSE)
for (i in 1:nrow(edgeVisData))
{
edgeVisData[i, ] = sort(edgeVisData[i, ])
}
edgeVisData = edgeVisData[!duplicated(edgeVisData),]
##STRING TWO(ADDING EDGE VALUES)#
if(proteinN=="1"){
edgeVisData$title<-"ppi"
# edgeVisDataMod<-merge(edgeVisData,s,by.x = c("from","to"),by.y =c("a","b"),all.x = TRUE)
edgeVisDataMod<-igraph::as_data_frame(igraph.from.graphNEL(module),what="edges")
for(i in 1:nrow(edgeVisDataMod)){
edgeVisDataMod$title[i]<-paste0("<p><b>Neighborhood score:</b>",edgeVisDataMod$f.neighborhood[i],"</p><b>Fusion score:</b>",edgeVisDataMod$f.fusion[i],"</p><b>Cooccurence score:</b>",edgeVisDataMod$f.cooccurence[i],"</p><b>Coexpression score:</b>",edgeVisDataMod$f.coexpression[i],"</p><b>Experimental score:</b>",edgeVisDataMod$f.experimental[i],"</p><b>Database score:</b>",edgeVisDataMod$f.database[i],"</p><b>Textmining score:</b>",edgeVisDataMod$f.textmining[i],"</p><b>Combined score:</b>",edgeVisDataMod$f.combined_score[i],"</p>")
}
edgeVisData<-edgeVisDataMod
}
##END...STRING TWO (ADDING EDGE VALUES)##
if(is.null(nodeGoData) & is.null(edgeGoData))
{
visObj<- visNetwork(nodeVisData, edgeVisData,height="800px",width="900px")
}
else{
visObj<- visNetwork(nodeGoData, edgeGoData,height="800px",width="900px")
# visObj<-enrichObj
}
visObj<-visExport(visObj,type ="png", name="network",float="right")
if(phy){
#print("physics active")
}
else{
visObj<-visIgraphLayout(visObj,layout=visLay)
}
if(package==TRUE)
{
visNetwork(nodeVisData, edgeVisData,height="800px",width="900px")
}
else{
visObj<-visInteraction(visObj,navigationButtons = TRUE)
visObj<-visOptions(visObj,manipulation = TRUE)
}
return(list("networkObj"=visObj,"networkData"=statNet,"edgeData"=edgeVisData,"nodeData"=nodeVisData))
}
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.