96-well Growth Curve Analysis

number_of_plates <- 5
number_of_reps <- 3


raw_dataset <- vector("list", length = number_of_reps)

for(repl in 1:number_of_reps){
	rawData <- c()
	for(plate in 1:number_of_plates){
		fileName=paste0("plate",plate,"-Rep",repl,".txt")
		cat(paste0("Opening input connection: `", fileName,"`..."), "\n")
		con=file(fileName,open="r")
		is.first = TRUE
		nline <- 0
		while ( length(line <- readLines(con, n = 1)) > 0 ) {
			nline <- nline + 1
			# if line is blank is skipped
			if (line == ""){
				cat(paste0("   blank line:", nline," skipped"),"\n")
				if(is.first == FALSE){
					rawData <- cbind(rawData, tmp)
				}

			}else{
			# if line contains comment beginning with: "[Table" is skipped
				if( substr(line, 1, 6) == "[Table" ){
					cat(paste0("   comment line: ",nline," skipped"),"\n")
				}else{
					if( substr(line, 1, 4) == "Time" ){
						# reading header and discard first "Time" column
						myColNames <- strsplit(line, split = ";")[[1]][-1]
						# add plate info to column names:
						myColNames <- paste0("plate_", plate, "_", myColNames)
						# initialize empty array
						tmp <- array(data = NA, dim = c(0, length(myColNames)))
						colnames(tmp) <- myColNames
					}else{
						# reading data
						tmp <- rbind(tmp, 
								as.numeric(strsplit(line, split = ";")[[1]][-1])
							)
						is.first = FALSE
					}
		
				}
			}

		}
		rawData <- cbind(rawData, tmp)
		cat(paste0("Closing input connection: `", fileName,"`."), "\n\n")
		close(con)
	}
	raw_dataset[[repl]] <- rawData
}

# the raw_dataset contains all data, with dimensions: 3 replicates x 289 time-points x 480 time-series

# clean raw_dataset by filtering out blanks and weirdos

averagedData <- array(data = NA, dim = dim(raw_dataset[[1]]))
n <- dim(raw_dataset[[1]])[2]
nTime <- dim(raw_dataset[[1]])[1]
for(i in 1:n){
	repl <- 1
	averagedData[,i] <- raw_dataset[[repl]][,i]
	for(repl in 2:number_of_reps){
		averagedData[,i] <- averagedData[,i] + raw_dataset[[repl]][,i]
	}
	
}
averagedData <- averagedData/number_of_reps
discardWells <- which(averagedData[nTime, ] < 0.015)	# this threshold excludes all blank wells, it should be adjusted accordingly to other applications. Here, a blank well is defined as one where the average growth level at the last time-point (t = 289) is less than 0.015.
discardWells <- c(discardWells, c(36, 75))	# these two cases are also excluded because they don't behave nice

myDataList <- lapply(raw_dataset, function(y)y[,-discardWells])	# final dataset with N = 411 time-series.
lapply(myDataList, function(y)dim(y))