view hairpinTool.R @ 18:11777fbeac32 default tip

Requirements temporarily removed
author shian_su <registertonysu@gmail.com>
date Wed, 17 Feb 2016 11:15:12 +1100
parents 525bf4364896
children
line wrap: on
line source

# ARGS: 1.inputType         -String specifying format of input (fastq or table)
#    IF inputType is "fastq" or "pairedFastq:
#       2*.fastqPath        -One or more strings specifying path to fastq files
#       2.annoPath          -String specifying path to hairpin annotation table
#       3.samplePath        -String specifying path to sample annotation table
#       4.barStart          -Integer specifying starting position of barcode
#       5.barEnd            -Integer specifying ending position of barcode
#    ###   
#    IF inputType is "pairedFastq":
#       6.barStartRev       -Integer specifying starting position of barcode
#                            on reverse end
#       7.barEndRev         -Integer specifying ending position of barcode
#                            on reverse end
#    ### 
#       8.hpStart           -Integer specifying startins position of hairpin
#                            unique region
#       9.hpEnd             -Integer specifying ending position of hairpin
#                            unique region
#    IF inputType is "counts":
#       2.countPath         -String specifying path to count table
#       3.annoPath          -String specifying path to hairpin annotation table
#       4.samplePath        -String specifying path to sample annotation table
#    ###
#       10.secFactName      -String specifying name of secondary factor
#       11.cpmReq           -Float specifying cpm requirement
#       12.sampleReq        -Integer specifying cpm requirement
#       13.readReq          -Integer specifying read requirement
#       14.fdrThresh        -Float specifying the FDR requirement
#       15.lfcThresh        -Float specifying the log-fold-change requirement
#       16.workMode         -String specifying exact test or GLM usage
#       17.htmlPath         -String specifying path to HTML file
#       18.folderPath       -String specifying path to folder for output
#    IF workMode is "classic" (exact test)
#       19.pairData[2]      -String specifying first group for exact test
#       20.pairData[1]      -String specifying second group for exact test
#    ###
#    IF workMode is "glm"
#       19.contrastData     -String specifying contrasts to be made
#       20.roastOpt         -String specifying usage of gene-wise tests
#       21.hairpinReq       -String specifying hairpin requirement for gene-
#                            wise test
#       22.selectOpt        -String specifying type of selection for barcode
#                            plots
#       23.selectVals       -String specifying members selected for barcode
#                            plots
#    ###
#
# OUT:  Bar Plot of Counts Per Index
#       Bar Plot of Counts Per Hairpin
#       MDS Plot
#       BCV Plot
#       Smear Plot
#       Barcode Plots (If Genewise testing was selected)
#       Top Expression Table
#       Feature Counts Table
#       HTML file linking to the ouputs
#
# Author: Shian Su - registertonysu@gmail.com - Jan 2014

# Record starting time
timeStart <- as.character(Sys.time())

# Loading and checking required packages
library(methods, quietly=TRUE, warn.conflicts=FALSE)
library(statmod, quietly=TRUE, warn.conflicts=FALSE)
library(splines, quietly=TRUE, warn.conflicts=FALSE)
library(edgeR, quietly=TRUE, warn.conflicts=FALSE)
library(limma, quietly=TRUE, warn.conflicts=FALSE)

if (packageVersion("edgeR") < "3.7.17") {
  stop("Please update 'edgeR' to version >= 3.7.17 to run this tool")
}

if (packageVersion("limma")<"3.21.16") {
  message("Update 'limma' to version >= 3.21.16 to see updated barcode graphs")
}

################################################################################
### Function declarations
################################################################################

# Function to load libaries without messages
silentLibrary <- function(...) {
  list <- c(...)
  for (package in list){
    suppressPackageStartupMessages(library(package, character.only=TRUE))
  }
}

# Function to sanitise contrast equations so there are no whitespaces
# surrounding the arithmetic operators, leading or trailing whitespace
sanitiseEquation <- function(equation) {
  equation <- gsub(" *[+] *", "+", equation)
  equation <- gsub(" *[-] *", "-", equation)
  equation <- gsub(" *[/] *", "/", equation)
  equation <- gsub(" *[*] *", "*", equation)
  equation <- gsub("^\\s+|\\s+$", "", equation)
  return(equation)
}

# Function to sanitise group information
sanitiseGroups <- function(string) {
  string <- gsub(" *[,] *", ",", string)
  string <- gsub("^\\s+|\\s+$", "", string)
  return(string)
}

# Function to change periods to whitespace in a string
unmake.names <- function(string) {
  string <- gsub(".", " ", string, fixed=TRUE)
  return(string)
}

# Function has string input and generates an output path string
makeOut <- function(filename) {
  return(paste0(folderPath, "/", filename))
}

# Function has string input and generates both a pdf and png output strings
imgOut <- function(filename) {
  assign(paste0(filename, "Png"), makeOut(paste0(filename,".png")), 
         envir=.GlobalEnv)
  assign(paste0(filename, "Pdf"), makeOut(paste0(filename,".pdf")),
         envir=.GlobalEnv)
}

# Create cat function default path set, default seperator empty and appending
# true by default (Ripped straight from the cat function with altered argument
# defaults)
cata <- function(..., file=htmlPath, sep="", fill=FALSE, labels=NULL, 
                 append=TRUE) {
  if (is.character(file)) 
    if (file == "") 
      file <- stdout()
  else if (substring(file, 1L, 1L) == "|") {
    file <- pipe(substring(file, 2L), "w")
    on.exit(close(file))
  }
  else {
    file <- file(file, ifelse(append, "a", "w"))
    on.exit(close(file))
  }
  .Internal(cat(list(...), file, sep, fill, labels, append))
}

# Function to write code for html head and title
HtmlHead <- function(title) {
  cata("<head>\n")
  cata("<title>", title, "</title>\n")
  cata("</head>\n")
}

# Function to write code for html links
HtmlLink <- function(address, label=address) {
  cata("<a href=\"", address, "\" target=\"_blank\">", label, "</a><br />\n")
}

# Function to write code for html images
HtmlImage <- function(source, label=source, height=600, width=600) {
  cata("<img src=\"", source, "\" alt=\"", label, "\" height=\"", height)
  cata("\" width=\"", width, "\"/>\n")
}

# Function to write code for html list items
ListItem <- function(...) {
  cata("<li>", ..., "</li>\n")
}

TableItem <- function(...) {
  cata("<td>", ..., "</td>\n")
}

TableHeadItem <- function(...) {
  cata("<th>", ..., "</th>\n")
}
################################################################################
### Input Processing
################################################################################

# Grabbing arguments from command line
argv <- commandArgs(TRUE)

inputType <- as.character(argv[1])
if (inputType == "fastq") {

  fastqPath <- as.character(gsub("fastq::", "", argv[grepl("fastq::", argv)], 
                                 fixed=TRUE))

  # Remove fastq paths
  argv <- argv[!grepl("fastq::", argv, fixed=TRUE)] 

  fastqPathRev <- NULL
  annoPath <- as.character(argv[2])
  samplePath <- as.character(argv[3])
  barStart <- as.numeric(argv[4])
  barEnd <- as.numeric(argv[5])
  barStartRev <- NULL
  barStartRev <- NULL
  hpStart <- as.numeric(argv[8])
  hpEnd <- as.numeric(argv[9])
} else if (inputType=="pairedFastq") {

  fastqPath <- as.character(gsub("fastq::", "", argv[grepl("fastq::", argv)], 
                                 fixed=TRUE))
  
  fastqPathRev <- as.character(gsub("fastqRev::", "", 
                               argv[grepl("fastqRev::", argv)], fixed=TRUE))

  # Remove fastq paths
  argv <- argv[!grepl("fastq::", argv, fixed=TRUE)]
  argv <- argv[!grepl("fastqRev::", argv, fixed=TRUE)] 

  annoPath <- as.character(argv[2])
  samplePath <- as.character(argv[3])
  barStart <- as.numeric(argv[4])
  barEnd <- as.numeric(argv[5])
  barStartRev <- as.numeric(argv[6])
  barEndRev <- as.numeric(argv[7])
  hpStart <- as.numeric(argv[8])
  hpEnd <- as.numeric(argv[9])
} else if (inputType == "counts") {
  countPath <- as.character(argv[2])
  annoPath <- as.character(argv[3])
  samplePath <- as.character(argv[4])
}

secFactName <- as.character(argv[10])
cpmReq <- as.numeric(argv[11])
sampleReq <- as.numeric(argv[12])
readReq <- as.numeric(argv[13])
fdrThresh <- as.numeric(argv[14])
lfcThresh <- as.numeric(argv[15])
selectDirection <- as.character(argv[16])
workMode <- as.character(argv[17])
htmlPath <- as.character(argv[18])
folderPath <- as.character(argv[19])

if (workMode == "classic") {
  pairData <- character()
  pairData[2] <- as.character(argv[20])
  pairData[1] <- as.character(argv[21])
} else if (workMode == "glm") {
  contrastData <- as.character(argv[20])
  roastOpt <- as.character(argv[21])
  hairpinReq <- as.numeric(argv[22])
  selectOpt <- as.character(argv[23])
  selectVals <- as.character(argv[24])
}

# Read in inputs

samples <- read.table(samplePath, header=TRUE, sep="\t")

anno <- read.table(annoPath, header=TRUE, sep="\t")

if (inputType == "counts") {
  counts <- read.table(countPath, header=TRUE, sep="\t")
}

###################### Check inputs for correctness ############################
samples$ID <- make.names(samples$ID)

if ( !any(grepl("group", names(samples))) ) {
  stop("'group' column not specified in sample annotation file")
} # Check if grouping variable has been specified

if (secFactName != "none") {
  if ( !any(grepl(secFactName, names(samples))) ) {
  tempStr <- paste0("Second factor specified as \"", secFactName, "\" but ",
                    "no such column name found in sample annotation file")
  stop(tempStr)
  } # Check if specified secondary factor is present 
}


if ( any(table(samples$ID) > 1) ){
  tab <- table(samples$ID)
  offenders <- paste(names(tab[tab > 1]), collapse=", ")
  offenders <- unmake.names(offenders)
  stop("'ID' column of sample annotation must have unique values, values ",
       offenders, " are repeated")
} # Check that IDs in sample annotation are unique

if (inputType == "fastq" || inputType == "pairedFastq") {
  
  if ( any(table(anno$ID) > 1) ){
    tab <- table(anno$ID)
    offenders <- paste(names(tab[tab>1]), collapse=", ")
    stop("'ID' column of hairpin annotation must have unique values, values ",
    offenders, " are repeated")
  } # Check that IDs in hairpin annotation are unique
  
} else if (inputType == "counts") {
  # The first element of the colnames will be 'ID' and should not match
  idFromSample <- samples$ID
  idFromTable <- colnames(counts)[-1]
  if (any(is.na(match(idFromTable, idFromSample)))) {
    stop("not all samples have groups specified")
  } # Check that a group has be specifed for each sample
  
  if ( any(table(counts$ID) > 1) ){
    tab <- table(counts$ID)
    offenders <- paste(names(tab[tab>1]), collapse=", ")
    stop("'ID' column of count table must have unique values, values ",
    offenders, " are repeated")
  } # Check that IDs in count table are unique
}
if (workMode == "glm") {
  if (roastOpt == "yes") {
    if (is.na(match("Gene", colnames(anno)))) {
      tempStr <- paste("Gene-wise tests selected but'Gene' column not",
                       "specified in hairpin annotation file")
      stop(tempStr)
    }
  }
}

if (secFactName != "none") {
  if (workMode != "glm") {
    tempStr <- paste("only glm analysis type possible when secondary factor",
                     "used, please change appropriate option.")
  }
}

################################################################################

# Process arguments
if (workMode == "glm") {
  if (roastOpt == "yes") {
    wantRoast <- TRUE
  } else {
    wantRoast <- FALSE
  }
}

# Split up contrasts seperated by comma into a vector and replace spaces with
# periods
if (exists("contrastData")) {
  contrastData <- unlist(strsplit(contrastData, split=","))
  contrastData <- sanitiseEquation(contrastData)
  contrastData <- gsub(" ", ".", contrastData, fixed=TRUE)
}

# Replace spaces with periods in pair data
if (exists("pairData")) {
  pairData <- make.names(pairData)
}

# Generate output folder and paths
dir.create(folderPath, showWarnings=FALSE)

# Generate links for outputs
imgOut("barHairpin")
imgOut("barIndex")
imgOut("mds")
imgOut("bcv")
if (workMode == "classic") {
  smearPng <- makeOut(paste0("smear(", pairData[2], "-", pairData[1],").png"))
  smearPdf <- makeOut(paste0("smear(", pairData[2], "-", pairData[1],").pdf"))
  topOut <- makeOut(paste0("toptag(", pairData[2], "-", pairData[1],").tsv"))
} else if (workMode == "glm") {
  smearPng <- character()
  smearPdf <- character()
  topOut <- character()
  roastOut <- character()
  barcodePng <- character()
  barcodePdf <- character()
  for (i in 1:length(contrastData)) {
    smearPng[i] <- makeOut(paste0("smear(", contrastData[i], ").png"))
    smearPdf[i] <- makeOut(paste0("smear(", contrastData[i], ").pdf"))
    topOut[i] <- makeOut(paste0("toptag(", contrastData[i], ").tsv"))
    roastOut[i] <- makeOut(paste0("gene_level(", contrastData[i], ").tsv"))
    barcodePng[i] <- makeOut(paste0("barcode(", contrastData[i], ").png"))
    barcodePdf[i] <- makeOut(paste0("barcode(", contrastData[i], ").pdf"))
  }
}
countsOut <- makeOut("counts.tsv")
sessionOut <- makeOut("session_info.txt")

# Initialise data for html links and images, table with the link label and
# link address
linkData <- data.frame(Label=character(), Link=character(),
                       stringsAsFactors=FALSE)
imageData <- data.frame(Label=character(), Link=character(),
                        stringsAsFactors=FALSE)
                        
# Initialise vectors for storage of up/down/neutral regulated counts
upCount <- numeric()
downCount <- numeric()
flatCount <- numeric()

################################################################################
### Data Processing
################################################################################

# Transform gene selection from string into index values for mroast
if (workMode == "glm") {
  if (selectOpt == "rank") {
    selectVals <- gsub(" ", "", selectVals, fixed=TRUE)
    selectVals <- unlist(strsplit(selectVals, ","))
    
    for (i in 1:length(selectVals)) {
      if (grepl(":", selectVals[i], fixed=TRUE)) {
        temp <- unlist(strsplit(selectVals[i], ":"))
        selectVals <- selectVals[-i]
        a <- as.numeric(temp[1])
        b <- as.numeric(temp[2])
        selectVals <- c(selectVals, a:b)         
      }
    }
    selectVals <- as.numeric(unique(selectVals))
  } else {
    selectVals <- gsub(" ", "", selectVals, fixed=TRUE)
    selectVals <- unlist(strsplit(selectVals, ","))
  }                                                           
}
                                                  
if (inputType == "fastq" || inputType == "pairedFastq") {
  # Use EdgeR hairpin process and capture outputs

  hpReadout <- capture.output(
  data <- processAmplicons(readfile=fastqPath, readfile2=fastqPathRev,
                            barcodefile=samplePath, 
                            hairpinfile=annoPath,
                            barcodeStart=barStart, barcodeEnd=barEnd,
                            barcodeStartRev=barStartRev, 
                            barcodeEndRev=barEndRev,
                            hairpinStart=hpStart, hairpinEnd=hpEnd, 
                            verbose=TRUE)
  )

  # Remove function output entries that show processing data or is empty
  hpReadout <- hpReadout[hpReadout!=""]
  hpReadout <- hpReadout[!grepl("Processing", hpReadout)]
  hpReadout <- hpReadout[!grepl("in file", hpReadout)]
  hpReadout <- gsub(" -- ", "", hpReadout, fixed=TRUE)

  # Make the names of groups syntactically valid (replace spaces with periods)
  data$samples$group <- make.names(data$samples$group)
  if (secFactName != "none") {
    data$samples[[secFactName]] <- make.names(data$samples[[secFactName]])
  }
} else if (inputType == "counts") {
  # Process counts information, set ID column to be row names
  rownames(counts) <- counts$ID
  counts <- counts[ , !(colnames(counts) == "ID")]
  countsRows <- nrow(counts)
  
  # Process group information
  sampleNames <- colnames(counts)
  matchedIndex <- match(sampleNames, samples$ID)
  factors <- samples$group[matchedIndex]

  if (secFactName != "none") {
    secFactors <- samples[[secFactName]][matchedIndex]
  }
  
  annoRows <- nrow(anno)
  anno <- anno[match(rownames(counts), anno$ID), ]
  annoMatched <- sum(!is.na(anno$ID))
  
  if (any(is.na(anno$ID))) {
    warningStr <- paste("count table contained more hairpins than",
                        "specified in hairpin annotation file")
    warning(warningStr)
  }
  
  # Filter out rows with zero counts
  sel <- rowSums(counts)!=0
  counts <- counts[sel, ]
  anno <- anno[sel, ]
  
  # Create DGEList
  data <- DGEList(counts=counts, lib.size=colSums(counts), 
                  norm.factors=rep(1,ncol(counts)), genes=anno, group=factors)
  
  # Make the names of groups syntactically valid (replace spaces with periods)
  data$samples$group <- make.names(data$samples$group)
}

# Filter out any samples with zero counts
if (any(data$samples$lib.size == 0)) {
  sampleSel <- data$samples$lib.size != 0
  filteredSamples <- paste(data$samples$ID[!sampleSel], collapse=", ")
  data$counts <- data$counts[, sampleSel]
  data$samples <- data$samples[sampleSel, ]
}

# Filter hairpins with low counts
preFilterCount <- nrow(data)
selRow <- rowSums(cpm(data$counts) > cpmReq) >= sampleReq
selCol <- colSums(data$counts) > readReq
data <- data[selRow, selCol]

# Check if any data survived filtering
if (length(data$counts) == 0) {
  stop("no data remains after filtering, consider relaxing filters")
}

# Count number of filtered tags and samples
postFilterCount <- nrow(data)
filteredCount <- preFilterCount - postFilterCount
sampleFilterCount <- sum(!selCol)

if (secFactName == "none") {
  # Estimate dispersions
  data <- estimateDisp(data)
  commonBCV <- round(sqrt(data$common.dispersion), 4)
} else {
  # Construct design
  if (inputType == "counts") {
      
    sampleNames <- colnames(counts)
    matchedIndex <- match(sampleNames, samples$ID)
    factors <- factor(make.names(samples$group[matchedIndex]))

    secFactors <- factor(make.names(samples[[secFactName]][matchedIndex]))

  } else if (inputType == "fastq" || inputType == "pairedFastq") {

    factors <- factor(data$sample$group)
    secFactors <- factor(data$sample[[secFactName]])
  
  }

  design <- model.matrix(~0 + factors + secFactors)
  
  # Estimate dispersions
  data <- estimateDisp(data, design=design)
  commonBCV <- round(sqrt(data$common.dispersion), 4)
}


################################################################################
### Output Processing
################################################################################

# Plot number of hairpins that could be matched per sample
png(barIndexPng, width=600, height=600)
barplot(height<-colSums(data$counts), las=2, main="Counts per index", 
        cex.names=1.0, cex.axis=0.8, ylim=c(0, max(height)*1.2))
imageData[1, ] <- c("Counts per Index", "barIndex.png")
invisible(dev.off())

pdf(barIndexPdf)
barplot(height<-colSums(data$counts), las=2, main="Counts per index", 
        cex.names=1.0, cex.axis=0.8, ylim=c(0, max(height)*1.2))
linkData[1, ] <- c("Counts per Index Barplot (.pdf)", "barIndex.pdf")
invisible(dev.off())

# Plot per hairpin totals across all samples
png(barHairpinPng, width=600, height=600)
if (nrow(data$counts)<50) {
  barplot(height<-rowSums(data$counts), las=2, main="Counts per hairpin",
          cex.names=0.8, cex.axis=0.8, ylim=c(0, max(height)*1.2))
} else {
  barplot(height<-rowSums(data$counts), las=2, main="Counts per hairpin",
          cex.names=0.8, cex.axis=0.8, ylim=c(0, max(height)*1.2),
          names.arg=FALSE)
}
imageData <- rbind(imageData, c("Counts per Hairpin", "barHairpin.png"))
invisible(dev.off())

pdf(barHairpinPdf)
if (nrow(data$counts)<50) {
  barplot(height<-rowSums(data$counts), las=2, main="Counts per hairpin",
          cex.names=0.8, cex.axis=0.8, ylim=c(0, max(height)*1.2))
} else {
  barplot(height<-rowSums(data$counts), las=2, main="Counts per hairpin",
          cex.names=0.8, cex.axis=0.8, ylim=c(0, max(height)*1.2),
          names.arg=FALSE)
}
newEntry <- c("Counts per Hairpin Barplot (.pdf)", "barHairpin.pdf")
linkData <- rbind(linkData, newEntry)
invisible(dev.off())

# Make an MDS plot to visualise relationships between replicate samples
png(mdsPng, width=600, height=600)
plotMDS(data, labels=data$samples$group, col=as.numeric(data$samples$group),
        main="MDS Plot")
imageData <- rbind(imageData, c("MDS Plot", "mds.png"))
invisible(dev.off())

pdf(mdsPdf)
plotMDS(data, labels=data$samples$group, col=as.numeric(data$samples$group),
        main="MDS Plot")
newEntry <- c("MDS Plot (.pdf)", "mds.pdf")
linkData <- rbind(linkData, newEntry)
invisible(dev.off())

# BCV Plot
png(bcvPng, width=600, height=600)
plotBCV(data, main="BCV Plot")
imageData <- rbind(imageData, c("BCV Plot", "bcv.png"))
invisible(dev.off())

pdf(bcvPdf)
plotBCV(data, main="BCV Plot")
newEntry <- c("BCV Plot (.pdf)", "bcv.pdf")
linkData <- rbind(linkData, newEntry)
invisible(dev.off())

if (workMode == "classic") {
  # Assess differential representation using classic exact testing methodology 
  # in edgeR
  testData <- exactTest(data, pair=pairData)
  
  top <- topTags(testData, n=Inf)

  if (selectDirection == "all") {
    topIDs <- top$table[(top$table$FDR < fdrThresh) &
                      (abs(top$table$logFC) > lfcThresh), 1]
  } else if (selectDirection == "up") {
    topIDs <- top$table[(top$table$FDR < fdrThresh) &
                      (top$table$logFC > lfcThresh), 1]
  } else if (selectDirection == "down") {
  topIDs <- top$table[(top$table$FDR < fdrThresh) &
                      (top$table$logFC < -lfcThresh), 1]
}
                      
  write.table(top, file=topOut, row.names=FALSE, sep="\t")
  
  linkName <- paste0("Top Tags Table(", pairData[2], "-", pairData[1], 
                     ") (.tsv)")
  linkAddr <- paste0("toptag(", pairData[2], "-", pairData[1], ").tsv")
  linkData <- rbind(linkData, c(linkName, linkAddr))
  
  upCount[1] <- sum(top$table$FDR < fdrThresh & top$table$logFC > lfcThresh)

  downCount[1] <- sum(top$table$FDR < fdrThresh & 
                      top$table$logFC < -lfcThresh)

  flatCount[1] <- sum(top$table$FDR > fdrThresh |
                      abs(top$table$logFC) < lfcThresh)
  
  
  
  # Select hairpins with FDR < 0.05 to highlight on plot
  png(smearPng, width=600, height=600)
  plotTitle <- gsub(".", " ", 
                    paste0("Smear Plot: ", pairData[2], "-", pairData[1]),
                    fixed=TRUE)
  plotSmear(testData, de.tags=topIDs, 
            pch=20, cex=1.0, main=plotTitle)
  abline(h=c(-1, 0, 1), col=c("dodgerblue", "yellow", "dodgerblue"), lty=2)
  imgName <- paste0("Smear Plot(", pairData[2], "-", pairData[1], ")")
  imgAddr <- paste0("smear(", pairData[2], "-", pairData[1],").png")
  imageData <- rbind(imageData, c(imgName, imgAddr))
  invisible(dev.off())
  
  pdf(smearPdf)
  plotTitle <- gsub(".", " ", 
                    paste0("Smear Plot: ", pairData[2], "-", pairData[1]),
                    fixed=TRUE)
  plotSmear(testData, de.tags=topIDs, 
            pch=20, cex=1.0, main=plotTitle)
  abline(h=c(-1, 0, 1), col=c("dodgerblue", "yellow", "dodgerblue"), lty=2)
  imgName <- paste0("Smear Plot(", pairData[2], "-", pairData[1], ") (.pdf)")
  imgAddr <- paste0("smear(", pairData[2], "-", pairData[1], ").pdf")
  linkData <- rbind(linkData, c(imgName, imgAddr))
  invisible(dev.off())
  
} else if (workMode == "glm") {
  # Generating design information
  if (secFactName == "none") {

    factors <- factor(data$sample$group)
    design <- model.matrix(~0 + factors)
    
    colnames(design) <- gsub("factors", "", colnames(design), fixed=TRUE)

  } else {

    factors <- factor(data$sample$group)

    if (inputType == "counts") {
      
      sampleNames <- colnames(counts)
      matchedIndex <- match(sampleNames, samples$ID)
      factors <- factor(samples$group[matchedIndex])

      secFactors <- factor(samples[[secFactName]][matchedIndex])

    } else if (inputType == "fastq" || inputType == "pairedFastq") {

      secFactors <- factor(data$sample[[secFactName]])
    
    }

    design <- model.matrix(~0 + factors + secFactors)
    
    colnames(design) <- gsub("factors", "", colnames(design), fixed=TRUE)
    colnames(design) <- gsub("secFactors", secFactName, colnames(design), 
                              fixed=TRUE)
  }
  
  
  # Split up contrasts seperated by comma into a vector
  contrastData <- unlist(strsplit(contrastData, split=","))
  
  for (i in 1:length(contrastData)) {
    # Generate contrasts information
    contrasts <- makeContrasts(contrasts=contrastData[i], levels=design)
    
    # Fit negative bionomial GLM
    fit <- glmFit(data, design)
    # Carry out Likelihood ratio test
    testData <- glmLRT(fit, contrast=contrasts)
    
    # Select hairpins with FDR < 0.05 to highlight on plot
    top <- topTags(testData, n=Inf)

    if (selectDirection == "all") {
      topIDs <- top$table[(top$table$FDR < fdrThresh) &
                        (abs(top$table$logFC) > lfcThresh), 1]
    } else if (selectDirection == "up") {
      topIDs <- top$table[(top$table$FDR < fdrThresh) &
                        (top$table$logFC > lfcThresh), 1]
    } else if (selectDirection == "down") {
      topIDs <- top$table[(top$table$FDR < fdrThresh) &
                        (top$table$logFC < -lfcThresh), 1]
    }

    write.table(top, file=topOut[i], row.names=FALSE, sep="\t")
    
    linkName <- paste0("Top Tags Table(", contrastData[i], ") (.tsv)")
    linkAddr <- paste0("toptag(", contrastData[i], ").tsv")
    linkData <- rbind(linkData, c(linkName, linkAddr))
    
    # Collect counts for differential representation
    upCount[i] <- sum(top$table$FDR < fdrThresh & top$table$logFC > lfcThresh)
    downCount[i] <- sum(top$table$FDR < fdrThresh & 
                        top$table$logFC < -lfcThresh)
    flatCount[i] <- sum(top$table$FDR > fdrThresh |
                        abs(top$table$logFC) < lfcThresh)
    
    # Make a plot of logFC versus logCPM
    png(smearPng[i], height=600, width=600)
    plotTitle <- paste("Smear Plot:", gsub(".", " ", contrastData[i], 
                       fixed=TRUE))
    plotSmear(testData, de.tags=topIDs, pch=20, cex=0.8, main=plotTitle)
    abline(h=c(-1, 0, 1), col=c("dodgerblue", "yellow", "dodgerblue"), lty=2)
    
    imgName <- paste0("Smear Plot(", contrastData[i], ")")
    imgAddr <- paste0("smear(", contrastData[i], ").png")
    imageData <- rbind(imageData, c(imgName, imgAddr))
    invisible(dev.off())
    
    pdf(smearPdf[i])
    plotTitle <- paste("Smear Plot:", gsub(".", " ", contrastData[i], 
                       fixed=TRUE))
    plotSmear(testData, de.tags=topIDs, pch=20, cex=0.8, main=plotTitle)
    abline(h=c(-1, 0, 1), col=c("dodgerblue", "yellow", "dodgerblue"), lty=2)
    
    linkName <- paste0("Smear Plot(", contrastData[i], ") (.pdf)")
    linkAddr <- paste0("smear(", contrastData[i], ").pdf")
    linkData <- rbind(linkData, c(linkName, linkAddr))
    invisible(dev.off())
    
    genes <- as.character(data$genes$Gene)
    unq <- unique(genes)
    unq <- unq[!is.na(unq)]
    geneList <- list()
    for (gene in unq) {
      if (length(which(genes == gene)) >= hairpinReq) {
        geneList[[gene]] <- which(genes == gene)
      }
    }
    
    if (wantRoast) {
      # Input preparaton for roast
      nrot <- 9999
      set.seed(602214129)
      roastData <- mroast(data, index=geneList, design=design,
                         contrast=contrasts, nrot=nrot)
      roastData <- cbind(GeneID=rownames(roastData), roastData)
      write.table(roastData, file=roastOut[i], row.names=FALSE, sep="\t")
      linkName <- paste0("Gene Level Analysis Table(", contrastData[i], 
                         ") (.tsv)")
      linkAddr <- paste0("gene_level(", contrastData[i], ").tsv")
      linkData <- rbind(linkData, c(linkName, linkAddr))
      if (selectOpt == "rank") {
        selectedGenes <- rownames(roastData)[selectVals]
      } else {
        selectedGenes <- selectVals
      }
      
      if (packageVersion("limma")<"3.19.19") {
        png(barcodePng[i], width=600, height=length(selectedGenes)*150)
      } else {
        png(barcodePng[i], width=600, height=length(selectedGenes)*300)
      }
      par(mfrow=c(length(selectedGenes), 1))
      for (gene in selectedGenes) {
        barcodeplot(testData$table$logFC, index=geneList[[gene]],
                    main=paste("Barcode Plot for", gene, "(logFCs)", 
                               gsub(".", " ", contrastData[i])),
                    labels=c("Positive logFC", "Negative logFC"))
      }
      imgName <- paste0("Barcode Plot(", contrastData[i], ")")
      imgAddr <- paste0("barcode(", contrastData[i], ").png")
      imageData <- rbind(imageData, c(imgName, imgAddr))
      dev.off()
      if (packageVersion("limma")<"3.19.19") {
        pdf(barcodePdf[i], width=8, height=2)
      } else {
        pdf(barcodePdf[i], width=8, height=4)
      }
      for (gene in selectedGenes) {
        barcodeplot(testData$table$logFC, index=geneList[[gene]],
                    main=paste("Barcode Plot for", gene, "(logFCs)", 
                               gsub(".", " ", contrastData[i])),
                    labels=c("Positive logFC", "Negative logFC"))
      }
      linkName <- paste0("Barcode Plot(", contrastData[i], ") (.pdf)")
      linkAddr <- paste0("barcode(", contrastData[i], ").pdf")
      linkData <- rbind(linkData, c(linkName, linkAddr))
      dev.off()
    }
  }
}

# Generate data frame of the significant differences
sigDiff <- data.frame(Up=upCount, Flat=flatCount, Down=downCount)
if (workMode == "glm") {

  row.names(sigDiff) <- contrastData

} else if (workMode == "classic") {

  row.names(sigDiff) <- paste0(pairData[2], "-", pairData[1])

}

# Output table of summarised counts
ID <- rownames(data$counts)
outputCounts <- cbind(ID, data$counts)
write.table(outputCounts, file=countsOut, row.names=FALSE, sep="\t",
            quote=FALSE)
linkName <- "Counts table (.tsv)"
linkAddr <- "counts.tsv"
linkData <- rbind(linkData, c(linkName, linkAddr))

# Record session info
writeLines(capture.output(sessionInfo()), sessionOut)
linkData <- rbind(linkData, c("Session Info", "session_info.txt"))

# Record ending time and calculate total run time
timeEnd <- as.character(Sys.time())
timeTaken <- capture.output(round(difftime(timeEnd,timeStart), digits=3))
timeTaken <- gsub("Time difference of ", "", timeTaken, fixed=TRUE)
################################################################################
### HTML Generation
################################################################################
# Clear file
cat("", file=htmlPath)

cata("<html>\n")
HtmlHead("EdgeR Output")

cata("<body>\n")
cata("<h3>EdgeR Analysis Output:</h3>\n")
cata("<h4>Input Summary:</h4>\n")
if (inputType == "fastq" || inputType == "pairedFastq") {

  cata("<ul>\n")
  ListItem(hpReadout[1])
  ListItem(hpReadout[2])
  cata("</ul>\n")
  cata(hpReadout[3], "<br />\n")
  cata("<ul>\n")
  ListItem(hpReadout[4])
  ListItem(hpReadout[7])
  cata("</ul>\n")
  cata(hpReadout[8:11], sep="<br />\n")
  cata("<br />\n")
  cata("<b>Please check that read percentages are consistent with ")
  cata("expectations.</b><br >\n")

} else if (inputType == "counts") {

  cata("<ul>\n")
  ListItem("Number of Samples: ", ncol(data$counts))
  ListItem("Number of Hairpins: ", countsRows)
  ListItem("Number of annotations provided: ", annoRows)
  ListItem("Number of annotations matched to hairpin: ", annoMatched)
  cata("</ul>\n")

}

cata("The estimated common biological coefficient of variation (BCV) is: ", 
     commonBCV, "<br />\n")

if (secFactName == "none") {

  cata("No secondary factor specified.<br />\n")

} else {

  cata("Secondary factor specified as: ", secFactName, "<br />\n")

}

cata("<h4>Output:</h4>\n")
cata("PDF copies of JPEGS available in 'Plots' section.<br />\n")
for (i in 1:nrow(imageData)) {
  if (grepl("barcode", imageData$Link[i])) {

    if (packageVersion("limma")<"3.19.19") {

      HtmlImage(imageData$Link[i], imageData$Label[i], 
                height=length(selectedGenes)*150)

    } else {

      HtmlImage(imageData$Link[i], imageData$Label[i], 
                height=length(selectedGenes)*300)

    }
  } else {

    HtmlImage(imageData$Link[i], imageData$Label[i])

  }
}
cata("<br />\n")

cata("<h4>Differential Representation Counts:</h4>\n")

cata("<table border=\"1\" cellpadding=\"4\">\n")
cata("<tr>\n")
TableItem()
for (i in colnames(sigDiff)) {
  TableHeadItem(i)
}
cata("</tr>\n")
for (i in 1:nrow(sigDiff)) {
  cata("<tr>\n")
  TableHeadItem(unmake.names(row.names(sigDiff)[i]))
  for (j in 1:ncol(sigDiff)) {
    TableItem(as.character(sigDiff[i, j]))
  }
  cata("</tr>\n")
}
cata("</table>")

cata("<h4>Plots:</h4>\n")
for (i in 1:nrow(linkData)) {
  if (grepl(".pdf", linkData$Link[i])) {
    HtmlLink(linkData$Link[i], linkData$Label[i])
  }
}

cata("<h4>Tables:</h4>\n")
for (i in 1:nrow(linkData)) {
  if (grepl(".tsv", linkData$Link[i])) {
    HtmlLink(linkData$Link[i], linkData$Label[i])
  }
}

cata("<p>Alt-click links to download file.</p>\n")
cata("<p>Click floppy disc icon on associated history item to download ")
cata("all files.</p>\n")
cata("<p>.tsv files can be viewed in Excel or any spreadsheet program.</p>\n")

cata("<h4>Additional Information:</h4>\n")

if (inputType == "fastq") {

  ListItem("Data was gathered from fastq raw read file(s).")

} else if (inputType == "counts") {

  ListItem("Data was gathered from a table of counts.")

}

if (cpmReq != 0 && sampleReq != 0) {
  tempStr <- paste("Target sequences without more than", cpmReq,
                   "CPM in at least", sampleReq, "samples are insignificant",
                   "and filtered out.")
  ListItem(tempStr)

  filterProp <- round(filteredCount/preFilterCount*100, digits=2)
  tempStr <- paste0(filteredCount, " of ", preFilterCount," (", filterProp,
                   "%) target sequences were filtered out for low ",
                   "count-per-million.")
  ListItem(tempStr)
}

if (readReq != 0) {
  tempStr <- paste("Samples that did not produce more than", readReq,
                   "counts were filtered out.")
  ListItem(tempStr)

  tempStr <- paste0(sampleFilterCount, " samples were filtered out for low ",
                    "counts.")
  ListItem(tempStr)
}

if (exists("filteredSamples")) {
  tempStr <- paste("The following samples were filtered out for having zero",
                   "library size: ", filteredSamples)
  ListItem(tempStr)
}

if (workMode == "classic") {
  ListItem("An exact test was performed on each target sequence.")
} else if (workMode == "glm") {
  ListItem("A generalised linear model was fitted to each target sequence.")
}

cit <- character()
link <-character()
link[1] <- paste0("<a href=\"",
                  "http://www.bioconductor.org/packages/release/bioc/",
                  "vignettes/limma/inst/doc/usersguide.pdf",
                  "\">", "limma User's Guide", "</a>.")
link[2] <- paste0("<a href=\"",
                  "http://www.bioconductor.org/packages/release/bioc/",
                  "vignettes/edgeR/inst/doc/edgeRUsersGuide.pdf",
                  "\">", "edgeR User's Guide", "</a>")
                  
cit[1] <- paste("Robinson MD, McCarthy DJ and Smyth GK (2010).",
                "edgeR: a Bioconductor package for differential",
                "expression analysis of digital gene expression",
                "data. Bioinformatics 26, 139-140")
cit[2] <- paste("Robinson MD and Smyth GK (2007). Moderated statistical tests",
                "for assessing differences in tag abundance. Bioinformatics",
                "23, 2881-2887")
cit[3] <- paste("Robinson MD and Smyth GK (2008). Small-sample estimation of",
                "negative binomial dispersion, with applications to SAGE data.",
                "Biostatistics, 9, 321-332")

cit[4] <- paste("McCarthy DJ, Chen Y and Smyth GK (2012). Differential",
                "expression analysis of multifactor RNA-Seq experiments with",
                "respect to biological variation. Nucleic Acids Research 40,",
                "4288-4297")

cata("<h4>Citations</h4>")
cata("<ol>\n")
ListItem(cit[1])
ListItem(cit[2])
ListItem(cit[3])
ListItem(cit[4])
cata("</ol>\n")

cata("<p>Report problems to: su.s@wehi.edu.au</p>\n")

for (i in 1:nrow(linkData)) {
  if (grepl("session_info", linkData$Link[i])) {
    HtmlLink(linkData$Link[i], linkData$Label[i])
  }
}

cata("<table border=\"0\">\n")
cata("<tr>\n")
TableItem("Task started at:"); TableItem(timeStart)
cata("</tr>\n")
cata("<tr>\n")
TableItem("Task ended at:"); TableItem(timeEnd)
cata("</tr>\n")
cata("<tr>\n")
TableItem("Task run time:"); TableItem(timeTaken)
cata("<tr>\n")
cata("</table>\n")

cata("</body>\n")
cata("</html>")