R/ROCAreas.R

#' @title ROCs AREAS
#' @description ROCAreas is a functions which allow plotting Area under ROCs for several hypothesis testing methods, shuch a several Differential Expression analysis.
#' @author Enrique Perez_Riesgo
#' @param dataRoc a data.frame where each column stores p values from each of the hypothesis test methods which are wanted to compare. The last column store the information related to real differential expression, where "0" means no differential expression and "1" means differential expression.
#' @param main the title of the plot
#' @param resample The number of resamplings of genes from data set in order to construct a boxplot from several samples from just one data set
#' @param DE proportion of differential expressed genes
#' @return a box plot with the AREA under ROC for every and each one of hypothesis test methods
#' @export ROCAreas
#' @examples
#' library(compcodeR)
#' library(edgeR)
#' set.seed(123456987)
#' datasCI<-generateSyntheticData(dataset = "EMOresults", n.vars = 15000,samples.per.cond = 6, n.diffexp = 1500,repl.id = 1, seqdepth = 1e7,fraction.upregulated = 0.5,between.group.diffdisp = FALSE,filter.threshold.total = 1,filter.threshold.mediancpm = 0,fraction.non.overdispersed = 0)
#' expressiondata<-datasCI@count.matrix
#' TMMfac<-calcNormFactors.default(expressiondata,method = "TMM")
#' exprT<-t(t(expressiondata)*TMMfac)
#' conditions<-(datasCI@sample.annotations$condition-1)
#' testEMOTMM<-EMObuODlmTest(expr=exprT, condition = conditions,original = exprT,originalm=exprT)
#' edgeR.dgelist = DGEList(counts = expressiondata, group = factor(conditions))
#' edgeR.dgelist = calcNormFactors(edgeR.dgelist, method = "TMM")
#' edgeR.dgelist = estimateCommonDisp(edgeR.dgelist)
#' edgeR.dgelist = estimateTagwiseDisp(edgeR.dgelist, trend = "movingave")
#' edgeR.test = exactTest(edgeR.dgelist)
#' edgeR.pvalues = edgeR.test$table$PValue
#' edgeR.adjpvalues = p.adjust(edgeR.pvalues, method = "BH")
#' testedgeRTMM<-data.frame(FC=edgeR.test$table$logFC, PVALUE=edgeR.test$table$PValue, ADJ.PVAL=edgeR.adjpvalues)
#' dataRoc<-data.frame(EMO=testEMOTMM$ADJ.PVAL,edgeR=testedgeRTMM$ADJ.PVAL, DE=datasCI@variable.annotations$differential.expression)
#' ROCAreas(dataRoc = dataRoc, main="ROC" Area,resample = 10,DE = 0.1)


ROCAreas<-function(dataRoc, resample=10,main, DE=0.1){
  AREAS<-data.frame(matrix(ncol=(dim(dataRoc)[2]-1), nrow=resample))
  colnames(AREAS)<-colnames(dataRoc)[1:(dim(dataRoc)[2]-1)]
  for(j in 1:resample){
    indexNDE<-sample(rownames(dataRoc[dataRoc[,"DE"]==0,]), size = 1000*(1-DE), replace = FALSE)
    indexDE<-sample(rownames(dataRoc[dataRoc[,"DE"]==1,]), size = 1000*DE, replace = FALSE)
    secuencians<-sort(c(0.05,unique(dataRoc[c(indexNDE, indexDE),(dim(dataRoc)[2]-1)])),decreasing = FALSE)
    FPr<-data.frame(matrix(nrow = length(secuencians), ncol = (dim(dataRoc)[2]-1)))
    TPr<-data.frame(matrix(nrow = length(secuencians), ncol = (dim(dataRoc)[2]-1)))
    areas<-data.frame(t(as.matrix(rep(0,(dim(dataRoc)[2]-1)))))

    for(i in (1:length(secuencians))){
      FPr[i,]<-apply(dataRoc[c(indexNDE,indexDE),1:(dim(dataRoc)[2]-1)]<=(secuencians[i])&dataRoc[c(indexNDE,indexDE),"DE"]==0, 2,function(x){sum(x,na.rm = TRUE)})/sum(dataRoc[c(indexNDE,indexDE),"DE"]==0,na.rm = TRUE)
      TPr[i,]<-apply(dataRoc[c(indexNDE,indexDE),1:(dim(dataRoc)[2]-1)]<=(secuencians[i])&dataRoc[c(indexNDE,indexDE),"DE"]==1, 2,function(x){sum(x,na.rm = TRUE)})/sum(dataRoc[c(indexNDE,indexDE),"DE"]==1,na.rm = TRUE)
    }
    for(k in 1:(dim(dataRoc)[2]-1)){
      areas[,k] <- pracma::trapz(y = TPr[,k],x=FPr[,k])
    }
    AREAS[j,]<-data.frame(areas)
  }
  par(mar=c(8,7,5,0.5))
  boxplot(AREAS, main=paste("ROC area \n", main), ylim=c(0,1), cex.axis=2, las=2, cex.lab=2, ylab="Areas\n ")

}
emodoro/dEMO documentation built on May 28, 2019, 12:57 p.m.