R/eliminate_redundant.R

In mappoly: Genetic Linkage Maps in Autopolyploids

#' Eliminate redundant markers
#'
#' Eliminate markers with identical dosage information for all individuals.
#'
#' @param input.seq an object of class \code{mappoly.sequence}
#'
#' @param data name of the dataset that contains sequence markers (optional, default = NULL)
#'
#' @return An object of class \code{mappoly.unique.seq} which
#'     is a list  containing the following components:
#'     \item{unique.seq}{an object of class \code{mappoly.sequence}
#'           with the redundant markers removed}
#'     \item{kept}{a vector containing the name of the informative markers}
#'     \item{eliminated}{a vector containing the name of the non-informative (eliminated) markers}
#'
#' @examples
#'     all.mrk <- make_seq_mappoly(hexafake, 'all')
#'     red.mrk <- elim_redundant(all.mrk)
#'     plot(red.mrk)
#'     unique.mrks <- make_seq_mappoly(red.mrk)
#'    
#' @author Marcelo Mollinari, \email{mmollin@ncsu.edu}, with minor modifications by Gabriel Gesteira, \email{gdesiqu@ncsu.edu}
#'
#' @references
#'     Mollinari, M., and Garcia, A.  A. F. (2019) Linkage
#'     analysis and haplotype phasing in experimental autopolyploid
#'     populations with high ploidy level using hidden Markov
#'     models, _G3: Genes, Genomes, Genetics_. 
#'     \doi{10.1534/g3.119.400378} 
#'
#' @export elim_redundant
elim_redundant <- function(input.seq, data = NULL)
{
  if (is.null(data)) x <- get(input.seq$data.name, pos = 1)$geno.dose[input.seq$seq.num, ]
  else x = data$geno.dose[input.seq$seq.num, ]
  dat_temp <- unique(x, dimnames = TRUE)
  if (is.null(data)) output.seq <- make_seq_mappoly(get(input.seq$data.name, pos = 1), rownames(dat_temp), data.name = input.seq$data.name)
  else output.seq <- make_seq_mappoly(data, rownames(dat_temp), data.name = data)
  output.seq$chisq.pval.thres  <- input.seq$chisq.pval.thres
  output.seq$chisq.pval  <- input.seq$chisq.pval
  if (is.null(data)) mrk.names <- get(input.seq$data.name, pos = 1)$mrk.names
  else mrk.names <- data$mrk.names
  elim <- setdiff(input.seq$seq.num,output.seq$seq.num)
  n1 <- apply(dat_temp, 1, paste, collapse = "")
  n2 <- apply(x[mrk.names[elim], , drop = FALSE], 1, paste, collapse = "")
  elim.out <- data.frame(kept = rownames(dat_temp)[match(n2,n1)], elim = mrk.names[elim])
  structure(list(unique.seq = output.seq, kept = mrk.names[output.seq$seq.num],
                 elim.correspondence = elim.out),
            class = "mappoly.unique.seq")
}

#' @export
plot.mappoly.unique.seq <- function(x, ...)
{
  slc <- c(nrow(x$elim.correspondence), length(x$kept))
  lbls <- c("eliminated", "kept")
  pct <- round(slc/sum(slc)*100)
  pct <- paste0(slc, " (", pct, "%)")
  lbls <- paste(lbls, pct, sep = "\n")
  pie(slc, labels = lbls)
}

#' @export
print.mappoly.unique.seq <- function(x, ...)
{
  print(x$unique.seq)
  cat("------------\n")
  cat("Eliminated markers: ", nrow(x$elim.correspondence), "\n")
}