R/comp_import_openms.R

In YasinEl/mzRAPP: Benchmark Dataset Generation and Non-Targeted Data Pre-Processing Assessment

#' import_ungrouped_openms
#'
#' @param file_list file_list
#' @param options_dt options_dt
#'
#' @importFrom  data.table as.data.table
#'
#' @keywords internal
import_ungrouped_openms <- function(file_list, options_dt){

  message('start OpenMS unaligned import')

  if(is.null(file_list)){
    stop('No ungrouped files selected')
  }

  if(length(file_list) < 2){
    stop('There should be multiple files for the unaligned OpenMS output!')
  }



  doFuture::registerDoFuture()
  future::plan("sequential")

  t <- Sys.time()
  `%dopar%` <- foreach::`%dopar%`
  Output <- list()
  Output <-
    foreach::foreach(i = 1:length(file_list),
            .packages = c("mzRAPP")) %dopar% {



                  file_path <- file_list[i]
                  file_name <- tools::file_path_sans_ext(basename(file_path))

              #Check if ug_table exists, if not: create

                    cols <- max(utils::count.fields(file_path, sep = ","))

                    suppressWarnings(
                    ug_table <- utils::read.table(file_path,
                                 col.names = paste("V", seq(cols)),
                                 fill = TRUE,
                                 sep = ",",
                                 comment.char = "",
                                 row.names = NULL,
                                 colClasses = "character")
                   )

                    ug_table <- as.data.table(ug_table)
                    colnames(ug_table)[1:2] <- c("V.1", "rpl")
                    if(!("#FEATURE" %in% unlist(ug_table[,"V.1"]))) {stop(paste("There is no information on FEATURES in " , file_path))}
                    ug_table <- ug_table[V.1 == "#FEATURE" | V.1 == "FEATURE"]
                    ug_table <- ug_table[,1:length(as.character(ug_table[1,])[as.character(ug_table[1,]) != ""])]
                    colnames(ug_table) <- as.character(unlist(unname(ug_table[V.1 == "#FEATURE"])))
                    ug_table <- ug_table[!1 ,!1]
                    ug_table <- ug_table[, sample_name := file_name]
                    return(ug_table)
            }

  ug_table <- data.table::rbindlist(Output, fill = TRUE, use.names = TRUE)


  #Check if all columns defined in optionsframe are present
  ug_req_cols <- stats::na.omit(options_dt$ug_columns)
  if(!all(ug_req_cols %in% colnames(ug_table))){
    cols_not_found <- setdiff(ug_req_cols, colnames(ug_table))
    stop('Columns defined in options but not present in OpenMS output: ', paste0(cols_not_found, sep = " - "))
  }

  #rename all columns for internal use according to option frame
  ug_table <- rename_columns_from_options(ug_table, options_dt, 'ug_columns', 'internal_columns')

  #Add a sample_id based on the sample_names in options_dt
  ug_table <- ug_table[options_dt, ':=' (sample_id = i.sample_id), on=c(sample_name = 'ug_samples')]

  #convert factors to numeric
  ug_table <- ug_table[, 'peak_area' := as.numeric(peak_area)]
  ug_table <- ug_table[, 'mz' := as.numeric(mz)]
  ug_table <- ug_table[, 'rt' := as.numeric(rt)]
  ug_table <- ug_table[, 'rt_start' := as.numeric(rt_start)]
  ug_table <- ug_table[, 'rt_end' := as.numeric(rt_end)]

  #Remove peaks where height and area are below 0
  ug_table <- ug_table[peak_area > 0] #height does not exist

  #Generate comp_id for each peak
  ug_table$comp_id <- seq.int(nrow(ug_table))

  #Add "_ug" as suffix to each column name
  colnames(ug_table) <- paste(colnames(ug_table), 'ug', sep = '_')

  message(paste0('Successful OpenMS unaligned import. No. of peaks imported: ', nrow(ug_table)))

  return(ug_table)
}





#' import_grouped_openms
#'
#' @param file_path file_path
#' @param options_table options_table
#'
#' @importFrom data.table as.data.table
#'
#' @keywords internal
import_grouped_openms <- function(file_path, options_table){

  message('Starting OpenMS aligned import')

  if(is.null(file_path)){
    return(NULL)
    stop('No grouped file selected')
  }

  #Check if filetype is csv
  if(tools::file_ext(file_path) != 'csv'){
    stop('grouped dataset is not a valid csv file')
  }

  if(length(file_path) != 1){
    stop('There should only be one file for the aligned MS-DIAL output!')
  }

  #Import csv file
  ######################g_table <- data.table::fread(file_path)#########################

  cols <- max(utils::count.fields(file_path, sep = ","))

  openms_exp <- utils::read.table(file_path,
                           col.names = paste("V", seq(cols)),
                           fill = TRUE,
                           sep = ",",
                           comment.char = "",
                           row.names = NULL,
                           colClasses = "character")

  openms_exp <- as.data.table(openms_exp)

  if(nrow(openms_exp[V.1 == "#CONSENSUS"]) != 1){
    stop('There is no #CONSENSUS row in aligned OpenMS export!')
  }


  #generate g_table
  g_table <- openms_exp[V.1 == "#CONSENSUS" | V.1 == "CONSENSUS"]
  g_table <- g_table[, 1:sum(g_table[1,] != "")]
  colnames(g_table) <- as.character(unname(unlist(g_table[1,])))
  g_table <- g_table[!1,!1]

  #agenerate sample id table
  sampid <- openms_exp[V.1 == "#MAP" | V.1 == "MAP"]
  sampid <- sampid[, 1:sum(sampid[1,] != "")]
  colnames(sampid) <- as.character(unname(unlist(sampid[1,])))
  sampid <- sampid[!1,!1]
  sampid[, vars := paste0("intensity_", id)]
  sampid <- sampid[, sample_name := tools::file_path_sans_ext(basename(as.character(filename)))]

  #Check if all columns defined in optionsframe are present
  g_req_cols <- stats::na.omit(options_table$g_columns)
  if(!all(g_req_cols %in% colnames(g_table))){
    cols_not_found <- setdiff(g_req_cols, colnames(g_table))
    stop('Columns defined in options but not present in aligned OpenMS export: ', paste0(cols_not_found, sep = " - "))
  }


  #Add feature_id for each row
  g_table[, feature_id := seq.int(nrow(g_table))]

  #transforming table from wide to long format, creating 1 peak-per-row format
  id_vars <- append(stats::na.omit(options_table[['g_columns']]), 'feature_id')
  measure_vars <- sampid$vars

  g_table <- data.table::melt(g_table,
                  id.vars = id_vars,
                  measure.vars = measure_vars,
                  variable.name = 'vars',
                  value.name = 'peak_area'
                  )

  #convert factors to numeric
  g_table <- g_table[, 'peak_area' := as.numeric(as.character(peak_area))]
  g_table <- sampid[, c("sample_name", "vars")][g_table, on=.(vars)][, !"vars"]

  #rename all columns for internal use according to options frame
  g_table <- rename_columns_from_options(g_table, options_table, 'g_columns', 'internal_columns')

  #Add a sample_id column based on the sample_names in options_dt
  g_table <- g_table[options_table, ':=' (sample_id = i.sample_id), on=c(sample_name = 'g_samples')]

  #convert factors to numeric
  g_table <- g_table[, 'mz' := as.numeric(mz)]
  g_table <- g_table[, 'rt' := as.numeric(rt)]

  #Remove peaks where area is below or equal 0
  g_table <- g_table[peak_area > 0]

  #Add comp_id for each peak
  g_table$comp_id <- seq.int(nrow(g_table))

  #Add "_g" as suffix to each column name
  colnames(g_table) <- paste(colnames(g_table), 'g', sep = '_')

  message(paste0('Successful OpenMS algined import. No. of peaks imported: ', nrow(g_table)))

  return(g_table)
}