R/check_tax_uniqueness.R

In vmikk/metagMisc: Miscellaneous functions for metagenomic analysis

#' @title Check taxonomy uniqueness.
#' @description This function checks for the duplicate taxonomy names that can occur in different groups (e.g., the same genus names in different families).
#' @param x Data frame with taxonomy (columns = taxonomy ranks, rows = species)
#' @param col Column name (taxonomy rank) that should be checked for uniqueness
#' @param return_all Logical; default, FALSE - only non-unique values will be returned
#' @param dropNA Logical; default, TRUE - missing values within the selected column ('col') will be removed
#' @return Data frame with counts of taxon occurence (by default, only non-unique values will be shown).
#' @export
#'
#' @examples
#' datt <- data.frame(
#'   Kingdom = rep(LETTERS[1:2], each = 6),
#'   Phylum = rep(LETTERS[3:6], each = 3),
#'   Class = rep(letters[1:6], times = 2),
#'   stringsAsFactors = F
#'   )
#'
#' check_tax_uniqueness(datt, col = "Phylum") # All ranks of Phylum are unique
#' check_tax_uniqueness(datt, col = "Class")  # Classes are duplicated
#'
check_tax_uniqueness <- function(x, col = "k", return_all = F, dropNA = T){
  # x = data frame
  #NB! Columns should be ordered, missing values are coded with NA

  # require(plyr)

  x <- as.data.frame(x)

  ## Where is the selected rank?
  COLID <- which(colnames(x) == col)

  ## Remove NAs
  if(dropNA == TRUE){
    nn <- is.na(x[,COLID])
    if(any(nn)){ x <- x[!nn, ] }
  }

  ## Remove lower ranks and count number of unique values
  res <- plyr::ddply(.data = x[, 1:COLID], .variables = col, .fun = function(z){
    rez <- z[!duplicated(z),]    # remove duplicates
    rezz <- data.frame(UniqueCombs = nrow(rez))
    return(rezz)
    })

  ## Remove unique
  if(return_all == FALSE){
    res <- subset(res, UniqueCombs > 1)
    if(nrow(res) == 0){ cat("All ranks of ", col, " are unique.\n") }
  }
  return(res)
}