modified: R/General_functions.R

	new file:   man/Fun.enrich_withFC.Rd

NAMESPACE

@@ -22,6 +22,7 @@ export(DepthSummary)
 export(DistObjects)
 export(DoDE)
 export(FromDist2Graph)
+export(Fun.enrich_withFC)
 export(GEM_Wrapper)
 export(GTSummary)
 export(Get_Clonal_Variants)

R/General_functions.R

@@ -462,4 +462,92 @@ pvalues<-c(pvalues,binom.md$p.value)
 fulllist<-cbind(query.peaks.motif.summary,FC=query.peaks.motif.summary$ratio/(sum(query.peaks.motif.summary$Freq.y)/sum(query.peaks.motif.summary$Freq.x)),pvalues=pvalues)
 fulllist$qvalue<-qvalue(fulllist$pvalues)$qvalues
 return(fulllist)
-}
+}
+
+#' Fun.enrich_withFC
+#'
+#' This function is to do GSEA enrichment analysis
+#' @param markergenesList is a list containing several elements each of which is a vector of gene names
+#' @param All.genes the background genes  data(all.genes.refer)
+#' @param db  =msig.db   the term database to use, default is the msig.db;  data(msig.db)
+#' @param qcutoff=0.05  the term to show with q value lower than 0.05
+#' @param top=NULL  pass a number to only show top # terms
+#' @return this will return a simple object with two elements, one is the table to show the significant terms , the other one shows the gene names involved significant terms
+#' @export
+#' @examples
+#' S7G.GSEA<-Fun.enrich_withFC(markergenesList=list(S7G=topS7G.genes),All.genes=all.genes.refer)
+Fun.enrich_withFC<-function(markergenesList,All.genes=all.genes.refer
+,db=msig.db,qcutoff=0.05,top=NULL)
+{
+	require(qvalue)
+	binomial.test.resultsList<-list()
+	resultname<-c()
+	for (markname in names(markergenesList))
+	{
+		if(length((markergenesList[[markname]]))==0)
+		{
+			next
+		}
+		print(markname)
+		binomialtest.msig.result<-c()
+		for (name in names(db))
+		{
+			binomialtest.msig.result<-rbind(binomialtest.msig.result,binomialtest.msig.enrch_deplet(as.character(markergenesList[[markname]]),All.genes,name,thedatabase=db))
+		}
+		fdr.enrich<-qvalue(binomialtest.msig.result[,2])$qvalues
+		fdr.deplete<-qvalue(binomialtest.msig.result[,3])$qvalues
+		FDR.combined<-c()
+		FDR.combined[which(is.na(binomialtest.msig.result$ratio))]<-NA
+		FDR.combined[which(binomialtest.msig.result$ratio>1)]<-fdr.enrich[which(binomialtest.msig.result$ratio>1)]
+		FDR.combined[which(binomialtest.msig.result$ratio<=1)]<-fdr.deplete[which(binomialtest.msig.result$ratio<=1)]
+		if(is.null(FDR.combined))
+		{
+			next
+		}
+		binomialtest.msig.result<-cbind(binomialtest.msig.result,FDR=FDR.combined)
+		binomialtest.msig.result<-binomialtest.msig.result[order(binomialtest.msig.result$FDR),]
+		binomial.test.resultsList<-c(binomial.test.resultsList,list(binomialtest.msig.result))
+		resultname<-c(resultname,markname)
+
+	}
+
+	names(binomial.test.resultsList)=resultname
+	DRscombo<-data.frame()
+	for (markname in names(binomial.test.resultsList))
+	{
+		tmp<-subset(binomial.test.resultsList[[markname]],FDR<qcutoff)
+		row.names(tmp)<-tmp[,"name"]
+		colnames(tmp)[c(4,5)]<-paste(markname,colnames(tmp)[c(4,5)],sep="_")
+		for (othermarkname in setdiff(names(binomial.test.resultsList),markname))
+		{
+			tmp2<-binomial.test.resultsList[[othermarkname]]
+			row.names(tmp2)<-tmp2[,"name"]
+			tmp2<-tmp2[row.names(tmp),c("ratio","FDR")]
+			colnames(tmp2)<-paste(othermarkname,colnames(tmp2),sep="_")
+			tmp<-cbind(tmp,tmp2)
+		}
+		tmp<-tmp[,c(-1,-2,-3)]
+		if (!is.null(top))
+		{
+			tmp<-head(tmp,n=top)
+		}
+		DRscombo<-Tomerge.col(DRscombo,tmp)
+	}
+	all.list<-c()
+	for (term in row.names(DRscombo))
+	{
+		cur.genelist<-c()
+		for (i in 1:length(markergenesList))
+		{
+			cur.genes<-markergenesList[[i]][markergenesList[[i]] %in% as.character(db[[term]])]
+			cur.genelist<-c(cur.genelist,list(cur.genes))
+		}
+		names(cur.genelist)<-names(markergenesList)
+		all.list<-c(all.list,list(cur.genelist))
+
+	}
+	names(all.list)<-row.names(DRscombo)
+#)
+#DRscombo<-DRscombo[tmpcondition,]
+	return(list(DRscombo,all.list=all.list))
+}