dupMetrics2pdf.R

#!/usr/bin/env Rscript
suppressPackageStartupMessages(library("optparse"))
# specify our desired options in a list
# by default OptionParser will add an help option equivalent to
# make_option(c("-h", "--help"), action="store_true", default=FALSE,
# help="Show this help message and exit")
option_list <- list(
                    make_option(c("-m", "--metrics"), default=".dupmetrics",
                                help="The file pattern to match (in dir) for finding hsmetrics files [default %default]",
                                metavar="metrics"),
                    make_option(c("-o","--outprefix"), default="dupmetrics",
                                help="The output file prefix [default %default]; pdf, a bunch of pngs, and a text file of all samples will be generated with this prefix")
                    )
# get command line options, if help option encountered print help and exit,
# otherwise if options not found on command line then set
parser <- OptionParser(usage = "%prog [options] directory_for_dupmetrics", option_list=option_list)
args <- parse_args(parser,positional_arguments=TRUE)
opts <- args$options

if(length(args$args)!=1) {
  stop("We require the directory of the dupmetrics files as an argument")
}

metdirectory=args$args[1]
fnames=dir(metdirectory,pattern=opts$metrics)

tmp=do.call('rbind',sapply(dir(metdirectory,pattern=opts$metrics),
  function(x) {
    tmp1=read.table(file.path(metdirectory,x),nrows=1,header=TRUE,sep="\t")
    print(x)
    print(dim(tmp1))
    return(tmp1)
  },simplify=FALSE))
tmp2=sapply(strsplit(fnames,'\\.'),function(x) {return(x[1])})
dtable = data.frame(sample=tmp2,tmp)
rownames(dtable)=tmp2

write.table(dtable,paste(opts$outprefix,'.txt',sep=""),col.names=TRUE,
            row.names=FALSE,quote=FALSE,sep="\t")



#summaryTable = apply(dtable[,-1],2,sum)
#summaryTable = data.frame(KeyName=names(summaryTable),Value=summaryTable)
#write.table(summaryTable,paste(opts$outprefix,'.summary.txt',sep=""),col.names=TRUE,
#            row.names=FALSE,quote=FALSE,sep="\t")

pdf(paste(opts$outprefix,'.pdf',sep=""),width=11,height=8)
par(las=2,mar=c(15,4,2,2),cex=0.5)
for(i in colnames(dtable)[3:ncol(dtable)]){
  barplot(dtable[,i],main=i,names.arg=tmp2)
}
dev.off()
print('done')