R/make_unlabeled.R

In danielbraas/MetabFUN: This is a collection of functions that I use for metabolomics data analysis

#' This function produces a data frame with fractional contribution data.
#' @author Daniel Braas
#' @param DF The input data. This should be a data frame that contains the MID envelope. The data frame should have the following
#' columns: Used_ID, Iso and each sample separated into individual columns.
#' @param max.iso The maximal number of isotopologues in the data (default is 50).
#' @return A data frame with fractional contribution data.
#' @export

make_unlabeled <- function(DF, max.iso=50){

  if (exists('Title')==F) stop('Title not specified')


  flat <- function(dat){
    if (nrow(dat)==1) return(dat)
    else {
      for (i in 1:nrow(dat)){
        if (is.na(dat$Value[i])) {
          dat$Value[i:nrow(dat)] <- NA
          return(dat)
        }
      }
    }
  }

  DF$Iso <- factor(DF$Iso, levels=paste0('M', 0:max.iso))
  DF$Value[DF$Value==0] <- NA

  if (sum(grepl('Name', names(DF))) > 0) {
    Order <- levels(DF$Condition)
    DF <- rename(DF, Used_ID=Name) %>%
      unite(Sample, Condition, Exp, sep='_') %>%
      split(.[c('Used_ID','Sample')]) %>%
      map(~ flat(.)) %>%
      do.call(rbind,.) %>%
      separate(Sample, c('Condition','Exp'), sep='_') %>%
      mutate(Condition = factor(Condition, levels = Order)) %>%
      rename(Name=Used_ID)
    rownames(DF) <- NULL
  }

  else {
    DF <- DF %>%
      gather(Sample, Value,-Used_ID, -Iso) %>%
      arrange(Used_ID, Sample, Iso) %>%
      split(.[c('Used_ID','Sample')]) %>%
      map(~ flat(.)) %>%
      do.call(rbind,.) %>%
      spread(Sample, Value)
  }

  write.csv(DF, paste(Title, '-Uncorrected raw data.csv'), row.names = F)
  return(DF)
}