Dealing with AN and Inf in a dataframe
0
0
Entering edit mode
@bernal-mizrachi-leon-2861
Last seen 5.2 years ago

I have not use R in a while and am trying to get a heatmap for a study we are doing. But I a encountering a problem that have not been able to overcome. I know it is a silly question but can figure out. I guess am a little rusty.

I am using the following data set from GEO GSE18817

And this is my complete code:

library("affy")
library("limma")
library("made4")
library("marray")
library("Heatplus")

logData<- read.table(dataFileName0, sep="\t", header = TRUE, skip = 0, na.strings = "NA", blank.lines.skip = FALSE, stringsAsFactors = default.stringsAsFactors())
dim(logData)

 

GeneID <- logData[,1]
GeneName <- logData[,3]
GeneData <- log2(logData[,4:dim(logData)[2]])
dim(GeneData)
boxplot(GeneData)
apropos("quantile")
Normalized<-normalizeQuantiles(GeneData)
boxplot(Normalized)

type1 <- c("NFE2L1","TRAM1","JUNB","RASSF4","SOCS1","PTPRF","SPSB3","ADAM8","C2","NPAS1","PLXND1","ZNF646","DENND3","PPARD")
type2<- c("HLA-F","HLA-E","ICAM1","ZFAND3","TNS4","EXD3","LMNA","BIRC7","FAM65A","SYNPO")
type3 <- c("USP1","HDJC9","PITX2","CCDC99","FAM29A","MCM10","C12orf48","PBK","ECT2","MSH2","DHFR","CCNA2","MAD2L1","KIF11")
type4<- c("AGPAT5","BRIX1","EIF4E","HSPD1","SFRS3","COX11","SEH1L","CPSF6","SETMAR","NIP7","PDSS1") #"RBM15",
#targetGenes <- c(type1,type2,type3,type4)

Agenes <-c(type1,type2)
Bgenes <- c(type3, type4)

targetGenes <- c(Bgenes)  #this is for calculating how many probes available
#chosen normalized data for target gene symbs
targetLines <- "NA"     #initial
for (i in 1:length(targetGenes)) {
  if ((targetGenes[i])%in%(GeneName)){
    if (length(which(GeneName == targetGenes[i])) > 0) {
      targetLines <- c(targetLines, which(GeneName == targetGenes[i]))
      cat(i, " : ", which(GeneName == targetGenes[i]), " --- ", targetGenes[i], "\n")
    }
  } #else {targetGenes[i] <- "NA"}
}
targetLines <- targetLines[targetLines != "NA"]     #remove the "NA" values
numPart1 <-  length(   targetLines)


targetGenes <- c(Bgenes, Agenes)

#chosen normalized data for target gene symbs
targetLines <- "NA"     #initial
for (i in 1:length(targetGenes)) {
  if ((targetGenes[i])%in%(GeneName)){
    if (length(which(GeneName == targetGenes[i])) > 0) {
      targetLines <- c(targetLines, which(GeneName == targetGenes[i]))
      cat(i, " : ", which(GeneName == targetGenes[i]), " --- ", targetGenes[i], "\n")
    }
  } #else {targetGenes[i] <- "NA"}
}
targetLines <- targetLines[targetLines != "NA"]
#targetLines <-     #remove the "NA" values
numPart2 <-  length(   targetLines) -  numPart1

#To select those genes for heatmap I do

gene.symbs <- GeneName[as.numeric(targetLines)]
gene.data <- Normalized[targetLines, ]

genevec<- c(rep(numPart1),rep(numPart2)  )
gene.class <- colnames(gene.data)

gene.data[ is.na(gene.data) ] <-0 #######Which I think is not working because when I put the function I get:

 

heatplot(gene.data, labRow = gene.symbs, classvec=genevec,cexRow = 1, cexCol = 1)

Data (original) range:  -9.54 0"
[1] "Data (scale) range:  NaN NaN"
[1] "Data scaled to range:  NaN NaN"
Error in hclust(distf(t(data)), method = method) : 
  NA/NaN/Inf in foreign function call (arg 11)

I have tried

using 

gene.data [!is.Infinite(gene.data)] <- 0 ### and has not worked.

Any suggestions?

thanks

Leon 

 

heatmap heatplus • 642 views
ADD COMMENT

Login before adding your answer.

Traffic: 482 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6