2019-01-17

## calculate correlation of multiple genes and multiple LncRNAs
lnc_Gene_Correlation<-function(difLncProfile,GeneExprofile,cutoff_cor=0.6,cutoff_p=0.01){
    ## define variables
    lnc<-c()
    gene<-c()
    corelation<-c()
    p<-c()
    
    k<-1
    ## use a for circulation to calculate all correlations of lncRNAs and mRNA
    for(i in 1:dim(difLncProfile)[1]){
        gene[k:(k+dim(GeneExprofile)[1]-1)]<-rownames(GeneExprofile)
        lnc[k:(k+dim(GeneExprofile)[1]-1)]<-rep(rownames(difLncProfile[i,]),dim(GeneExprofile)[1])
        corelation[k:(k+dim(GeneExprofile)[1]-1)]<-apply(GeneExprofile,1,function(x){cor(as.numeric(difLncProfile[i,]),as.numeric(x))})
        p[k:(k+dim(GeneExprofile)[1]-1)]<-apply(GeneExprofile,1,function(x){cor.test(as.numeric(difLncProfile[i,]),as.numeric(x))$p.value})
        k<-k+dim(GeneExprofile)[1]
    }
    ## combine the variables to a table
    lncGeneCor<-data.frame(lnc,gene,corelation,p)
    ## filter the correlations
    lncGeneCor1<-lncGeneCor[which(abs(lncGeneCor[["corelation"]])>=cutoff_cor & lncGeneCor[["p"]]<cutoff_p),]
    return(lncGeneCor1)
}


## read files
read.table("GeneExprofile.txt")->GeneExprofile
read.table("DifLncProfile.txt")->DifLncProfile
## set cutoff
cutoff_cor=0.6
cutoff_p=0.01
## function call to calculate all correlations of lncRNAs and mRNA
LncGeneCor<-lnc_Gene_Correlation(DifLncProfile,GeneExprofile)
## write files
write.table(LncGeneCor,"LncGeneCorResult.txt",row.names=F,sep="\t",quote=F)

## show some result
head(LncGeneCor)
## to get neighbor genes of one lncRNAs
lncRNA2="ENSG00000034063"
neighborGenes<-LncGeneCor[which(LncGeneCor[,1]==lncRNA2),2]
write.table(neighborGenes,"neighborGenes.txt",row.names=F,col.names=F,quote=F,sep="\t")