server-scripts/qhtcp-workflow/apps/r/addShiftVals.R

#!/usr/bin/env Rscript
# This script will add the shift data to the finalTable.csv file
#
# May want to reorder columns in excel before making heatmaps - otherwise all the shift data will be plotted next to each other.

library(plyr)
library(dplyr)
library(sos)

args <- commandArgs(TRUE)

if (length(args) >= 1) {
  finalTable <- file.path(args[1])
} else {
  finalTable <- "REMcRdy_lm_only.csv-finalTable.csv"  # for legacy workflow
}

if (length(args) >= 2) {
  shiftFile <- file.path(args[2])
} else {
  shiftFile <- "Shift_only.csv"  # for legacy workflow
}

if (length(args) >= 3) {
  studyInfo <- file.path(args[3])
} else {
  studyInfo <- "../Code/StudyInfo.csv" # for legacy workflow
}

if (length(args) >= 4) {
  output <- file.path(args[4])
} else {
  output <- "REMcHeatmaps/REMcWithShift.csv" # for legacy workflow
}

# Read in the REMc finalTable data
X <- data.frame(read.csv(file = finalTable, header = TRUE, stringsAsFactors = FALSE))
# Read in the shift data From ../JoinInteractions
Y <- data.frame(read.csv(file = shiftFile, header = TRUE, stringsAsFactors = FALSE))
Labels <- read.delim(studyInfo, skip = 0, as.is = TRUE, row.names = 1, strip.white = TRUE)

# Determine the number of cols - needed to create the correct number of new cols
Xcolnum <- length(X[1, ])
ADDnum <- Xcolnum + length(Y[1, ]) - 2

# Create new columns filled with NAs to be filled with data
Xtemp <- X
Xtemp[, (Xcolnum + 1):ADDnum] <- NA

# Match the orf names in each row to a orf name in the shift data file and then add the shift data to the finalTable file
shiftTbl < - as.data.frame(matrix(nrow = 1, ncol = length(Y) - 2)) #the df shiftTbl must be initialized before for loop

for (i in 1:length(X[, 1])) {
  Shiftrownum <- match(X[i, 2], Y[, 1])
  shiftTbl[i, ] <- Y[Shiftrownum, 3:length(Y[1, ])]
  Xtemp[i, (Xcolnum + 1):ADDnum] <- Y[Shiftrownum, 3:length(Y[1, ])]
}
headerX <- colnames(Xtemp)
headerY <- colnames(Y)
shfHdr <- headerY[3:length(headerY)]
combTbl <- X[, 1:3]
lmTbl <- select(Xtemp,  contains("Z_lm"))  #X[,(4:Xcolnum-2)]
shiftTbl <- select(Xtemp,  contains("V"))
clustTbl <- select(Xtemp, contains("cluster."))

# Give the new column names the same names as in the shift file
Xcols <- colnames(X)
Ycols <- colnames(Y)[3:length(Y[1, ])]
newCols <- c(Xcols[1:Xcolnum], Ycols)

# Reorder columns for generating heatmaps
combI <- combTbl  #Starting Template orf, Genename columns
headersRemc <- newCols  #colnames(X)
newHeaders <- newCols[1:3]
lmHdr <- colnames(lmTbl)   #newCols[4:(length(Xcols)-2)]
clstHdr <- colnames(clustTbl)  #select(newCols, contains('cluster.'))  #newCols[3+length(lmHdr):2]

intLvHdr <- vector()
#Reorder columns to produce an interleaved set of Z_lm and Shift data for all the cpps.
for (i in 1:(length(shiftTbl[1, ]))) {
  combI <- cbind.data.frame(combI, shiftTbl[i])
  combI <- cbind.data.frame(combI, lmTbl[i])
  intLvHdrx <- c(shfHdr[i], lmHdr[i])
  intLvHdr <- c(intLvHdr, intLvHdrx)
}

combIHdr <- c(colnames(combTbl), intLvHdr, clstHdr)
combI <- cbind.data.frame(combI, clustTbl)
colnames(combI) <- combIHdr
write.csv(combI, file = output, row.names = FALSE)