R/file.R

In erah: Automated Spectral Deconvolution, Alignment, and Metabolite Identification in GC/MS-Based Untargeted Metabolomics

load.file <- function(filename)
{
  sampleRD <- NULL
  file.extension <- strsplit(filename, split="\\.")[[1]]
  file.type <- file.extension[length(file.extension)]
  
  if(file.type=="cdf") sampleRD <- load.ncdf4(filename)
  if(file.type=="CDF") sampleRD <- load.ncdf4(filename)
  if(file.type=="mzXML" || file.type=="xml") sampleRD <- load.xml(filename)
  if(file.type=="mzML") sampleRD <- load.xml(filename)
  if(file.type=="rdata") sampleRD <- load.erahrd(filename)
  if(is.null(sampleRD)) stop("File extension not recognized. Avalible extensions are: .cdf, .mzXML and .xml")
  
  sampleRD
}

load.erahrd <- function(filename)
{
  sampleRD <- 0
  samp.file <- load(filename)
  sampleObject <- sampleRD
  remove(sampleRD)
  sampleObject
}


load.xml <- function(filename)
{		
  if(requireNamespace("mzR", quietly = TRUE)) {
    xmlO <- mzR::openMSfile(filename)
    metadata <- mzR::runInfo(xmlO)
    scans <- 1:metadata$scanCount
    lowMZ <- metadata$lowMz
    highMZ <- metadata$highMz
    if(lowMZ==0 | highMZ==0 | scans[2]==0)
    {
      peakLst <- mzR::peaks(xmlO)
      mzVct <- unlist(lapply(peakLst, function(x) x[,1]))
      lowMZ <- min(mzVct, na.rm=T)
      highMZ <- max(mzVct, na.rm=T)
      scans <- which(unlist(lapply(peakLst, function(x) nrow(x)))!=0)
    }
    lowMZ <- round(lowMZ + 0.5)
    highMZ <- round(highMZ + 0.5)
    StartTime <- metadata$dStartTime
    ScansPerSecond <- 1/((metadata$dEndTime - metadata$dStartTime)/metadata$scanCount)
    
    log <- utils::capture.output(raw.data <- mzR::get3Dmap(object = xmlO, scans = scans, lowMz = lowMZ, highMz = highMZ, resMz = 1))
    sampleRD <- new("RawDataParameters", data = raw.data, min.mz = lowMZ, max.mz = highMZ, start.time = StartTime, mz.resolution = 1, scans.per.second = ScansPerSecond)
    return(sampleRD)
  }
  else {
    msg <- c("mzR is not installed. eRah and Baitmet can operate withouth mzR, unless you want to process .mzXML files (as in this case). To install the mzR package and be able to use mzXML files, please visit its bioconductor website: http://bioconductor.org/packages/release/bioc/html/mzR.html\nOr, alternatively, execute the following R code:\n\t\t\n\t\t## try http:// if https:// URLs are not supported \n\t\tsource('https://bioconductor.org/biocLite.R')\n\t\tbiocLite('mzR')")
    warning(msg)
  }
}

load.ncdf4 <- function(filename)
{	
  if(!requireNamespace("ncdf4", quietly = TRUE)){
    msg <- c("ncdf4 is not installed. eRah and Baitmet can operate withouth ncdf4, unless you want to process .CDF files (as in this case). To install the ncdf4 package use: install.packages('ncdf4')")
    stop(msg)
  }  
  isExact <- FALSE
  measurement = ncdf4::nc_open(filename)
  mass_values <- ncdf4::ncvar_get(measurement, "mass_values")
  rndSmplColl <- sample(mass_values, 500)
  if (any(rndSmplColl != (rndSmplColl^2/trunc(rndSmplColl)))) isExact <- TRUE
  mass_intensities <- ncdf4::ncvar_get(measurement, "intensity_values")
  scan_indexes <- ncdf4::ncvar_get(measurement, "scan_index")
  min_mz <- round(min(mass_values)) - 1
  max_mz <- round(max(mass_values)) + 1
  start_time <- as.numeric(ncdf4::ncvar_get(measurement, "scan_acquisition_time", count = 1))
  rndScan <- as.numeric(ncdf4::ncvar_get(measurement, "scan_acquisition_time", count = 10))[10]
  rndScan2 <- as.numeric(ncdf4::ncvar_get(measurement, "scan_acquisition_time", count = 11))[11]
  scans_per_second <- as.numeric((1/(rndScan2 - rndScan)))
  if (isExact) {
    full.matrix <- matrix(0, length(scan_indexes), ((max_mz - min_mz) + 1))
    mass_values <- round(mass_values - (min_mz - 1))
    for (i in 1:(length(scan_indexes) - 1)) {
      MssLoc <- mass_values[(scan_indexes[i] + 1):scan_indexes[i +  1]]
      MssInt <- mass_intensities[(scan_indexes[i] + 1):scan_indexes[i + 1]]
      MssInt <- as.vector(unlist(lapply(split(MssInt, MssLoc), sum)))
      MssLoc <- unique(MssLoc)
      full.matrix[i, MssLoc] <- MssInt
    }
  }
  else {
    full.matrix <- matrix(0, length(scan_indexes), ((max_mz - min_mz) + 1))
    mass_values <- mass_values - (min_mz - 1)
    for (i in 1:(length(scan_indexes) - 1)) {
      full.matrix[i, mass_values[(scan_indexes[i] + 1):scan_indexes[i + 1]]] <- mass_intensities[(scan_indexes[i] + 1):scan_indexes[i + 1]]
    }
  }
  sampleRD <- new("RawDataParameters", data = full.matrix, min.mz = min_mz, max.mz = max_mz, start.time = start_time, mz.resolution = 1, scans.per.second = scans_per_second)
  sampleRD
}

#' @name createdt
#' @aliases createdt
#' @title Creating Experiment Tables
#' @description eRah requires an instrumental and (optionally) phenotype .csv file for starting/creating a new eRah project/experiment. This function automatically creates the Phenoytpe and Instrumental data .csv files.
#' @usage createdt(path)
#' @param path the path where the experiment-folder is (where the experiment samples are stored).
#' @details 
#' The experiment has to been organized as follows: all the samples related to each class have to be stored in the same folder (one folder = one class), and all the class-folders in one folder, which is the experiment folder. 
#'
#' Two things have to be considered at this step: .csv files are different when created by American and European computers, so errors may raise due to that fact. Also, the folder containing the samples, must contain only folders. If the folder contains files (for example, already created .csv files), eRah will prompt an error.
#'
#' See eRah vignette for more details. To open the vignette, execute the following code in R:
#'  vignette("eRahManual", package="erah")
#' @examples \dontrun{
#' # Store all the raw data files in one different folder per class,
#' # and all the class-folders in one folder, which is the experiment
#' # folder. Then execute
#'
#' createdt(path)
#'
#' # where path is the experiment folder path.
#' # The experiment can be now startd by:
#'
#' ex <- newExp(instrumental="path/DEMO_inst.csv", 
#' phenotype="path/DEMO_pheno.csv", info="DEMO Experiment")
#' }
#' @seealso \code{\link{newExp}}
#' @export
#' @importFrom utils write.table

createdt <- function(path)
{
  #path <- "Valli/GenCond"
  #path.dir <- list.files(path)
  path.name <- strsplit(path, "/")[[1]]
  path.name <- path.name[length(path.name)]
  
  #dirs.c <- unlist(apply(as.matrix(path.dir), 1, function(x) rep(x,length(list.files(paste(path, x, sep="/"))))))
  #path.dir.c <- apply(as.matrix(path.dir),1, function(x) paste(path, x, sep="/"))
  #files.c <- list.files(path.dir.c)
  
  #files.name <- apply(as.matrix(1:length(dirs.c)),1, function(x) paste(dirs.c[x],files.c[x], sep="/"))
  #files.name <- list.files(path.dir.c, full.name=T)
  #files.ID <- apply(as.matrix(files.c), 1, function(x) strsplit(x, "\\.")[[1]][1])
  
  files.name <- list.files(path, recursive=T)
  
  if(any(apply(as.matrix(files.name), 1, function(x) length(strsplit(x, "/")[[1]]))==1)) stop("There are files without directory in the selected path. Remove all the files in the path, only folders are allowed")
  
  files.class <- apply(as.matrix(files.name), 1, function(x) strsplit(x, "/")[[1]][1])
  files.ID <- apply(as.matrix(files.name), 1, function(x) {
    out.s <- strsplit(x, "/")[[1]]
    strsplit(out.s[length(out.s)], "\\.")[[1]][1] 
  })
  
  files.path <- apply(as.matrix(files.name),1, function(x) paste(path, x, sep="/"))	
  files.cdate <- apply(as.matrix(files.path), 1, function(x) as.character(file.info(x)$mtime))
  files.date <- apply(as.matrix(files.cdate),1, function(x) strsplit(x, " ")[[1]][1])
  files.time <- apply(as.matrix(files.cdate),1, function(x) strsplit(x, " ")[[1]][2])
  
  inst.table <- matrix(0, ncol=4, nrow=length(files.ID))
  colnames(inst.table) <- c("sampleID", "filename", "date", "time") 
  
  inst.table[,1] <- files.ID
  inst.table[,2] <- files.name
  inst.table[,3] <- files.date
  inst.table[,4] <- files.time
  
  meta.table <- matrix(0, ncol=2, nrow=length(files.ID))
  colnames(meta.table) <- c("sampleID", "class") 
  
  meta.table[,1] <- files.ID
  meta.table[,2] <- files.class
  
  inst.file <- paste(path, "/", path.name, "_inst.csv", sep="")
  meta.file <- paste(path, "/", path.name, "_pheno.csv", sep="")
  
  write.table(inst.table, file=inst.file, sep=";", row.names=FALSE, eol="\n", quote=F)
  write.table(meta.table, file=meta.file, sep=";", row.names=FALSE, eol="\n", quote=F)	
  
}

#' @title Create Instrumental Table
#' @description Create table containing instrumental information such as sample IDs and file names.
#' @param files File paths to experiment samples.
#' @details Creates instrumental information table based on experiment sample file paths. Columns containing further information can also be added to this. 
#' @examples \dontrun{
#' library(gcspikelite)
#' 
#' files <- list.files(system.file('data',package = 'gcspikelite'),full.names = TRUE)
#' files <- files[sapply(files,grepl,pattern = 'CDF')]
#' 
#' instrumental <- createInstrumentalTable(files)
#' }
#' @seealso \code{\link{newExp}} \code{\link{createPhenoTable}}
#' @importFrom tibble tibble
#' @export

createInstrumentalTable <- function(files){
  
  files.ID <- apply(as.matrix(files), 1, function(x) {
    out.s <- strsplit(x, "/")[[1]]
    strsplit(out.s[length(out.s)], "\\.")[[1]][1] 
  })
  
  files.cdate <- apply(as.matrix(files), 1, function(x) as.character(file.info(x)$mtime))
  files.date <- apply(as.matrix(files.cdate),1, function(x) strsplit(x, " ")[[1]][1])
  files.time <- apply(as.matrix(files.cdate),1, function(x) strsplit(x, " ")[[1]][2])
  
  inst.table <- tibble(
    sampleID = files.ID,
    filename = files,
    date = files.date,
    time = files.time
  )
 
  return(inst.table)
}

#' @title Create Phenotype Table
#' @description Create table containing sample meta information such as as sample ID and class.
#' @param files File paths to experiment samples.
#' @param cls Character vector containing sample classes.
#' @details Creates phenotype information table based on experiment sample file paths and sample classes. Columns containing further information can also be added to this. 
#' @examples \dontrun{
#' library(gcspikelite)
#' data(targets)
#' 
#' files <- list.files(system.file('data',package = 'gcspikelite'),full.names = TRUE)
#' files <- files[sapply(files,grepl,pattern = 'CDF')]
#' 
#' phenotype <- createPhenoTable(files,as.character(targets$Group[order(targets$FileName)]))
#' }
#' @seealso \code{\link{newExp}} \code{\link{createInstrumentalTable}}
#' @importFrom tibble tibble
#' @export

createPhenoTable <- function(files,cls){
  files.ID <- apply(as.matrix(files), 1, function(x) {
    out.s <- strsplit(x, "/")[[1]]
    strsplit(out.s[length(out.s)], "\\.")[[1]][1] 
  })
  
  meta.table <- tibble(
    sampleID = files.ID,
    class = cls
  )
  
  return(meta.table)
}