Mercurial > repos > marpiech > rnaseq_pro_workflow_tools
diff tools/script.R @ 0:c5a812cdf478 draft
planemo upload
author | marpiech |
---|---|
date | Fri, 09 Dec 2016 10:52:35 -0500 |
parents | |
children | db5d126bf8d0 |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tools/script.R Fri Dec 09 10:52:35 2016 -0500 @@ -0,0 +1,68 @@ +#!/usr/bin/env Rscript + +args = commandArgs(TRUE) +countDataPath = args[1] +statsDataPath = args[2] +logFC = args[3] +logCPM = args[4] +pValue = args[5] +fdr = args[6] + +require(preprocessCore) +require(gplots) + +#prepare counts data -------------------------------------------------------- +countData = read.table(countDataPath, comment = "", + sep = "\t") + +groups = sapply(as.character(countData[1, -1]), strsplit, ":") +groups = as.vector(t(countData[1, -1])) + +names = as.vector(t(countData[2, -1])) + +countData = countData[-c(1,2), ] +rownames(countData) = countData[, 1] +countData = countData[, -1] +colnames(countData) = names + +countData = countData[, order(groups, names)] + +# prepare stats data ------------------------------------------------------ + +statsData = read.table(statsDataPath, sep = "\t", + header = T) + +colnames(statsData)[-1] = sapply(colnames(statsData)[-1], function(x){ + unlist(strsplit(x, ".", fixed = T))[3] +}) + +wh = which(abs(statsData$logFC) >= logFC & statsData$logCPM >= logCPM & statsData$PValue <= pValue & statsData$FDR <= fdr) + +for(i in 1:ncol(countData)){ + countData[,i] = as.numeric(countData[,i]) +} + +countDataNorm = normalize.quantiles(as.matrix(countData), copy = T) +countDataNormLog = log(countDataNorm + 1, 2) + +colnames(countDataNormLog) = colnames(countData) +rownames(countDataNormLog) = rownames(countData) + +#svg("heatmap.svg", width = 3+length(names), height = 1/2*length(wh)) +pdf("heatmap.pdf") + +heatmap.2( + countDataNormLog[wh, ], + density.info=c("none"), + hclustfun = function(x) hclust(x, method = "average"), + distfun = function(x) as.dist(1-cor(t(x))), + col = bluered(50), + scale = 'row', + trace = "none", #lwid = c(1, length(names)), lhei = c(1,1/3*length(wh)), +# Rowv=NA, + Colv = NA, + margins = c(7, 8) +) + +dev.off() +