R/morphology_corr_knignenburg.R

In bjreisman/DebarcodeR: Automated Demultiplexing for Fluorescent Cell Barcoding

#' Corrects morphology based on scatter or uptake control.
#'
#' @param fcb the barcoded dataframe, post compensation and preprocessing
#' @param uptake Optional: a dataframe consisting of all cells barcoded with a single level of the barcoding dye
#' @param channel The name (string) of the channel to be corrected, ie. the column name in 'fcb_df'
#' @param subsample Integer, number of cells to sample (with replacement) for the morphology correction, defaults to 10,000.
#' @param updateProgress used in reactive context (shiny) to return progress information to GUI#'
#'
#' @return a tibble/data.frame with the selected channel corrected for fsc and ssc
#'
#' @seealso \code{\link{selectDenseScatterArea}} \code{\link{doRegressConstrained}}
#' @export

morphology_corr.knijnenburg <- function(fcb,
                                        uptake,
                                        channel,
                                        fsc_ssc = c(fsc = 'FSC-A', ssc = 'SSC-A'),
                                        subsample = 10e3,
                                        updateProgress = NULL,
                                        ret.model = FALSE) {
  if (is.function(updateProgress)) {
    updateProgress(detail = "Mapping cellular Density")
  }

  area_density <- selectDenseScatterArea(uptake,
                                         subsample = subsample,
                                         fsc_ssc = fsc_ssc)


  if (is.function(updateProgress)) {
    updateProgress(detail = "Performing Morphology Correction")
  }
  regression.output <- doRegressConstrained(
    uptake,
    fcb,
    fsc_ssc = fsc_ssc,
    Loc = area_density$loc,
    weight = area_density$c,
    trans = "none",
    columns = c(channel),
    monodir = c(1, 1)
  )
  cor.data <- regression.output[[1]]
  fcb[, channel] <- cor.data[, channel]

  if (ret.model == FALSE) {
    return(list(values = fcb[, channel]))
  } else{
    return(list(values = fcb[, channel],
                model = regression.output))
  }
}