#install.packages('doSNOW');
#install.packages('ggplot2');
#install.packages('directlabels');
#install.packages('gridExtra');
#install.packages('hexbin');
#install.packages('Hmisc');
#install.packages('gplots');
#install.packages('GMD');
#install.packages('FactoMineR');
#install.packages('reshape');
#install.packages('svDialogs');
#install.packages('stringr');
library(ggplot2);
library(directlabels);
library(gridExtra);
library(hexbin);
library(Hmisc);
library(gplots);
library(RColorBrewer);
library(GMD);
library(FactoMineR);
library(parallel);
library(foreach);
library(doParallel);
library(plyr);
library(reshape);
library(doSNOW);
library(svDialogs);
library(stringr);
#####################################################################################
####                         User-defined Functions                              ####
#####################################################################################

## Filters out psms with "Rejected" in the PSM Ambiguity column
ambfilter <- function(x){
  x[!(x$PSM.Ambiguity == "Rejected"),]
}

## Averages the psm quantification from each TMT label
addAvgSN <- function(inDF, tmts){
  cbind(inDF, AvgSN = rowMeans(inDF[,tmts, drop = FALSE]))
}

## Filters out psms with an average Signal:Noise < 10
Snfilter <- function(x){
  x[!(x$AvgSN < 10),]
}

## Filters out psms with an isolation interferenece > 25
specfilter <- function(x){
  x[!(x$Isolation.Interference.... > 25),]
}

## Extract the columns we need
extract <- function(x, tmts){
  x[,c("Protein", tmts)]
}

## Sum the values for each unique protein ID
sumvals <- function(x){
  ddply(x,.(Protein),summarize,
        X126.1=sum(X126), X127N.1=sum(X127N),X127C.1=sum(X127C),X128N.1=sum(X128N),X128C.1=sum(X128C),
        X129N.1=sum(X129N),X129C.1=sum(X129C),X130N.1=sum(X130N),X130C.1=sum(X130C),X131.1=sum(X131))
}

## Change column names to append file they came from 
namexchange <- function(x,y){
  colnames(x)[names(x) == "Protein"] <- c(paste("ProteinID"))
  colnames(x)[names(x) == "X126.1"] <- c(paste("X126", {y}, sep = "_"))
  colnames(x)[names(x) == "X127N.1"] <- c(paste("X127n", {y}, sep = "_"))
  colnames(x)[names(x) == "X127C.1"] <- c(paste("X127c", {y}, sep = "_"))
  colnames(x)[names(x) == "X128N.1"] <- c(paste("X128n", {y}, sep = "_"))
  colnames(x)[names(x) == "X128C.1"] <- c(paste("X128c", {y}, sep = "_"))
  colnames(x)[names(x) == "X129N.1"] <- c(paste("X129n", {y}, sep = "_"))
  colnames(x)[names(x) == "X129C.1"] <- c(paste("X129c", {y}, sep = "_"))
  colnames(x)[names(x) == "X130N.1"] <- c(paste("X130n", {y}, sep = "_"))
  colnames(x)[names(x) == "X130C.1"] <- c(paste("X130c", {y}, sep = "_"))
  colnames(x)[names(x) == "X131.1"] <- c(paste("X131", {y}, sep = "_"))
  return(x)
}

namexchange2 <- function(x,y){
  colnames(x)[names(x) == "Protein"] <- c(paste("ProteinID"))
  if(exists("X126.1")) colnames(x)[names(x) == "X126.1"] <- c(paste("X126", {y}, sep = "_"))
  if(exists("X127N.1")) colnames(x)[names(x) == "X127N.1"] <- c(paste("X127n", {y}, sep = "_"))
  if(exists("X127C.1")) colnames(x)[names(x) == "X127C.1"] <- c(paste("X127c", {y}, sep = "_"))
  if(exists("X128N.1")) colnames(x)[names(x) == "X128N.1"] <- c(paste("X128n", {y}, sep = "_"))
  if(exists("X128C.1")) colnames(x)[names(x) == "X128C.1"] <- c(paste("X128c", {y}, sep = "_"))
  if(exists("X129N.1")) colnames(x)[names(x) == "X129N.1"] <- c(paste("X129n", {y}, sep = "_"))
  if(exists("X129C.1")) colnames(x)[names(x) == "X129C.1"] <- c(paste("X129c", {y}, sep = "_"))
  if(exists("X130N.1")) colnames(x)[names(x) == "X130N.1"] <- c(paste("X130n", {y}, sep = "_"))
  if(exists("X130C.1")) colnames(x)[names(x) == "X130C.1"] <- c(paste("X130c", {y}, sep = "_"))
  if(exists("X131.1")) colnames(x)[names(x) == "X131.1"] <- c(paste("X131", {y}, sep = "_"))
  return(x)
}
#####################################################################################
####                         Data Processing Code                                ####
#####################################################################################

#A) Set your working directory to the location of your input file
setwd("/Users/johnlapek/documents")

ldf <- list() # creates a list for the data files
nonrejected <-list() #creates a list for the non-rejected data
na.as.one <-list() #creates a list when replacing the NA values of the quantifications with 1's for proper averaging
signal <-list() #creates a list for data summing sn for all channels
Sn <-list() #creates a list for data removing below SN threshold
spec <-list()#creates a list for data removing below specificity threshold
spec2 <- list()
NAgone <-list()#creates a list for data removing rows with NAs
extractA <-list()#creates a list for dataframes with extracted columns
extractB <-list()#creates a list for dataframes with extracted columns
values <-list()#creates a list for final summed values
final <-list()#creates a list for final values with named columns
filePD <- list()#creates a list for the file from Proteome Discoverer
listcsv <- list()#creates a list for separated files from the PD output
tmtlablist <- list("X126", "X127N", "X127C", "X128N","X128C", "X129N","X129C", "X130N","X130C", "X131") #List of potential TMT labels
labelsperfile <- list()
 # list("X126.1", "X127n.1", "X127c.1", "X128n.1", "X128c.1", "X129n.1", "X129c.1", "X130n.1", "X130c.1", "X131.1")
temp <- list()

#B) Read in .csv Proteome Discoverer file
filePD[[1]] = read.csv("Peptides_Gencode.csv")

#addinfo <- c(as.character(filePD[[1]]$Protein), filePD[[1]]$Coverage, as.character(filePD[[1]]$Description))
#addinfo  <- matrix(addinfo ,ncol=3,byrow=FALSE)
#colnames(addinfo) <- c("ProteinID", "%_Coverage", "Description")
#addinfo  <- unique(addinfo)

#C) Get user input on number of files put into PD
n <- dlgInput(message = "How many MS runs did you perform? ", default = "")$res
tempPos <- gregexpr(':',n)[[1]][2]
n <- as.numeric(substr(n, tempPos+1, nchar(n)))

#Create empty vector for user defined file IDs
fileIDs <- c()

#Read in user defined file IDs
for(i in 1:n){
  tempID <- dlgInput(message = "Enter a unique string from each MS run: ", default = "")$res
  tempPos <- gregexpr(':',tempID)[[1]][2]
  tempID <- substr(tempID, tempPos+1, nchar(tempID))
  fileIDs <- c(fileIDs, tempID)
}

#Allow user to select TMT labels used for each file
for(i in 1:n){
  tmtlab <- dlgList(tmtlablist, multiple = TRUE)$res
  labelsperfile[i] <- list(tmtlab)
}

dir.create("CSVs/")
setwd("CSVs/")
#Parse PD output into .csvs based on user-specified file number and IDs
for(i in 1:n){
  t1 <- filePD[[1]][,"Spectrum.File"][grepl(fileIDs[i], filePD[[1]][,"Spectrum.File"])]
  t2 <- filePD[[1]][filePD[[1]][,"Spectrum.File"] %in% t1,]
  write.csv(t2, file=paste(fileIDs[i], ".csv", sep=""))
}

#D) Read in parsed data
listcsv <- dir(pattern = "*.csv")
temp <- NULL
temp2 <- NULL
#peptides <- NULL
for (k in 1:length(listcsv)){
  temp <- labelsperfile[[k]]
  temp2 <- paste(temp, ".1", sep="")
  ldf[[k]] = read.csv(listcsv[k])

  nonrejected[[k]] = data.frame(lapply(ldf[k], ambfilter))
  nonrejected[[k]][,39:ncol(nonrejected[[k]])][is.na(nonrejected[[k]][,39:ncol(nonrejected[[k]])])] <- 1
  
  signal[[k]] = data.frame(lapply(nonrejected[k], addAvgSN, temp))
  Sn[[k]] = data.frame(lapply(signal[k], Snfilter))
  spec[[k]] = data.frame(lapply(Sn[k], specfilter))
  #peptides[[k]] = data.frame(table(unlist(spec[[k]][2])))
  
  extractA[[k]] = data.frame(lapply(spec[k], extract, unlist( tmtlablist)))
  #colnames(peptides[[k]]) <- c(as.character("ProteinID"), paste("Peptides_",as.character(fileIDs[k]),sep=""))
  values[[k]] = data.frame(lapply(extractA[k], sumvals))
  extractB[[k]] = data.frame(lapply(values[k], extract, temp2))
  
  print(k)
  
  #Change column names to append file they came from
  final[[k]] = data.frame(lapply(extractB[k], namexchange, listcsv[k]))
}

#Merge all data frames in list by the column ProteinID
finalvalues <- data.frame(Reduce(function(x,y) merge(x,y,all=TRUE), final))
#pepNos <- data.frame(Reduce(function(x,y) merge(x,y,all=TRUE), peptides))

#finalvalues <- finalvalues[2:nrow(finalvalues),]
#rm(NAgone, Sn, extractA, extractB, final, ldf, signal, spec, values)
write.table(finalvalues, 'nonnormalizeddataall.txt', sep = "\t", row.names = FALSE, col.names = TRUE)

#Populate average matrix
avgMat <- NULL
avgVec <- NULL
temp <- NULL
for (i in 1:length(listcsv)){
  if(i==1){
    x1 <- 2
    x2 <- length(labelsperfile[[i]]) + 1
  }else{
    x1 <- x1 + length(labelsperfile[[i-1]])
    x2 <- x2 + length(labelsperfile[[i]])
  }
  
  tempRow <- finalvalues[,x1:x2]
  tempMean <- rowMeans(tempRow)
  avgVec <- c(avgVec, tempMean)
}
avgMat <- matrix(avgVec,ncol=length(listcsv),byrow=FALSE)
allMedian <- median(avgMat, na.rm = TRUE)

avgnorm = c();
for (i in 1:length(listcsv)){
  if(i==1){
    x1 <- 2
    x2 <- length(labelsperfile[[i]]) + 1
  }else{
    x1 <- x1 + length(labelsperfile[[i-1]])
    x2 <- x2 + length(labelsperfile[[i]])
  }

  subset <- finalvalues[,x1:x2]
  subset2 <- subset/(avgMat[,i]/allMedian)
  
  if(i==1){
    avgnorm <- subset2
  }else{
    avgnorm <- cbind(avgnorm, subset2)
  }
}

datamedian <- median(unlist(avgnorm[,1:length(avgnorm)]), na.rm = TRUE)

finalnorm = data.frame()
colmeds <- list()
for (j in 1:dim(avgnorm)[2]){
  colmeds[[j]] = median(unlist(avgnorm[,j]), na.rm = TRUE)
  if(j == 1)
  {
    finalnorm = data.frame(avgnorm[,j]/(as.numeric(colmeds[j])/datamedian));
  } else
  {
    finalnorm = data.frame(cbind(finalnorm, avgnorm[,j]/(as.numeric(colmeds[j])/datamedian)));
  }
}

names(finalnorm) <- names(avgnorm)
finalnorm <- cbind(finalvalues["ProteinID"], finalnorm)

#finalnorm <- merge(finalnorm, pepNos, by = "ProteinID")
#finalnorm <- merge(finalnorm, addinfo, by = "ProteinID")

write.table(finalnorm, 'NormalizedDataAll.txt', sep = "\t", col.names = TRUE, row.names = FALSE)

commonreps <- na.omit(finalnorm)

write.table(commonreps, 'NormalizedCommonReps.txt', sep = "\t", col.names = TRUE, row.names = FALSE)