R/normalize.R

In MALDIcellassay: Automated MALDI Cell Assays Using Dose-Response Curve Fitting

#' .scalingFactor re implemented from sgibb/MALDIquant
#'
#' @param object AbstractMassObject
#' @param method character
#' @param range double
#'
#' @return
#' double
#' @importFrom MALDIquant trim
#' @noRd
.scalingFactor <- function(object, method=c("TIC", "median"), range) {
  
  method <- match.arg(method)
  
  if (!missing(range)) {
    object <- trim(object, range=range)
  }
  
  switch(method,
         "TIC" = {
           totalIonCurrent(object)
         },
         "median" = {
           median(object@intensity)
         }
  )
}


#' Normalize spectra and peaks
#'
#' @param spec     List of MALDIquant::MassSpectrum
#' @param peaks    List of MALDIquant::MassPeaks
#' @param normMeth Character, normalization method. Options are "TIC", "median" and "mz".
#' @param normMz   Numeric, mz used to normalize.
#' @param normTol  Numeric, tolerance around normMz.
#'
#' @return
#' List of lists of normalized MALDIquant::MassSpectrum, normalized MALDIquant::MassPeaks,
#' normalization factors as well as indicies of spectra containing the `normMz` in case of `normMeth = "mz"`,
#'
#' @export
#' @importFrom MALDIquant totalIonCurrent
#' @examples
#' data(Blank2022spec)
#' data(Blank2022peaks)
#' norm <- normalize(Blank2022spec, Blank2022peaks, normMeth = "mz", normMz = 760.585, normTol = 0.1)
#' 
#' # normalization factors
#' norm$factor 
normalize <- function(spec, peaks, normMeth, normMz, normTol) {
  nm <- names(spec)
  stopifnot(!is.null(nm))
  stopifnot(is.numeric(as.numeric(nm)))

  switch(normMeth,
         "TIC" = {
           tic <- purrr::map_dbl(spec, totalIonCurrent)

           spec <- normalizeByFactor(spec, tic)
           peaks <- normalizeByFactor(peaks, tic)

           norm_fac <- tic
           included_specIdx <- 1:length(spec)
         },
         "median" = {
           median <- vapply(spec,
                            FUN = function(x)
                              .scalingFactor(object = x,
                                                          method = "median"),
                            numeric(1))

           spec <- normalizeByFactor(spec, median)
           peaks <- normalizeByFactor(peaks, median)

           norm_fac <- median
           included_specIdx <- 1:length(spec)
         },
         "mz" = {
           mzNorm <- getNormFactors(
             peaksdf = peaks2df(peaks),
             targetMz = normMz,
             tol = normTol,
             allowNoMatch = TRUE,
             tolppm = FALSE
           )

           norm_fac <- mzNorm$norm_factor
           spec <- normalizeByFactor(spec[mzNorm$specIdx], mzNorm$norm_factor)
           peaks <- normalizeByFactor(peaks[mzNorm$specIdx], mzNorm$norm_factor)
           included_specIdx <- mzNorm$specIdx

           u_nm <- unique(nm)
           u_fil <- unique(nm[included_specIdx])
           if(length(u_nm) != length(u_fil)) {
             # stop if a single condition got filtered completely

             label_removed <- u_nm[which(!(u_nm %in% u_fil))]

             stop("Could not find ", normMz, " in all spectra with label ",
                  paste0(label_removed, collapse = ", "),
                  ".\n Consider increasing tol.\n")
           }
         },
         "none" = {
           norm_fac <- list("norm_factor" = 0)
           included_specIdx <- 1:length(spec)
         }
  )

  return(list(spec = spec,
              peaks = peaks,
              factor = norm_fac,
              idx = included_specIdx))
}