R/filter.R

In gsrc: Genome Structure Rearrangement Calling in Genomes with High Synteny

#' Filter Samples
#'
#' Failed or negative control samples (e.g. H2O) should be filtered out before the normalization.
#'
#' @param dat raw_data object.
#' @param samples Character vector with sample names, prefixes of samples to
#' filter or sample indices.
#' If prefixes fit many samples, indices are highly recommended.
#' @return raw_data object
#' @examples
#' if(require(brassicaData)){
#' data("raw_napus", package = "brassicaData", envir = environment())
#' raw_napus <- filt_samp(raw_napus, check_raw(raw = raw_napus, plot = FALSE, thresh = 28000))
#' }
#' @export
filt_samp <- function(dat, samples) {
  if (is.character(samples)) {
    cnames <- dat$samples
    filt <- c()
    for (i in samples) {
      filt <- c(filt, grep(i, cnames))
    }
  }else{
    filt <- samples
  }
  if (length(filt) > 0) {
    dat$samples <- dat$samples[-filt]
    if (is.matrix(dat$raw))
      dat$raw <- dat$raw[,-filt]
    if (is.matrix(dat$sd) &
        dim(dat$sd)[2] == length(samples))
      dat$sd <- dat$sd[,-filt]
    if (is.matrix(dat$beads) &
        dim(dat$beads)[2] == length(samples))
      dat$beads <- dat$beads[,-filt]
    if (is.matrix(dat$intensity))
      dat$intensity <- dat$intensity[,-filt]
    if (is.matrix(dat$theta))
      dat$theta <- dat$theta[,-filt]
    if (is.matrix(dat$baf))
      dat$baf <- dat$baf[,-filt]
    if (is.matrix(dat$rratio))
      dat$rratio <- dat$rratio[,-filt]
    if (is.matrix(dat$geno))
      dat$geno <- dat$geno[,-filt]
    if (is.matrix(dat$cnv))
      dat$cnv <- dat$cnv[,-filt]
    if (is.matrix(dat$tl))
      dat$tl <- dat$tl[,-filt]
    if (is.data.frame(dat$cna))
      dat$cna <- dat$cna[,dat$cna$ID %in% dat$samples]
  }
  dat
}

#' Filter SNPs
#'
#' Some SNPs do not work as well as others and might be filtered out.
#'
#' @param dat Matrix with raw data or list with intensity, theta, position and chromosome objects.
#' @param filt Character or numberic vector with SNP names or rownumbers to filter out.
#'
#' @return Filtered matrix.
#' @examples
#' if(require(brassicaData)){
#' data("raw_napus", package = "brassicaData", envir = environment())
#' \dontshow{
#' raw_napus <- filt_samp(raw_napus, raw_napus$samples[-(1:100)])
#' raw_napus <- filt_snps(raw_napus, raw_napus$snps[-(1:10)])
#' }
#' dat <- intens_theta(raw_napus)
#' dat <- remove_suffix(dat, "_Grn")
#' dat <- geno_baf_rratio(dat, delthresh = 11)
#' dat <- filt_snps(dat, dat$snps[is.na(rowMeans(dat$baf, na.rm = TRUE))])
#' }
#' @export filt_snps
#' @rdname filt_snps
filt_snps <- function(dat, filt) {
  UseMethod("filt_snps")
}

#' @rdname filt_snps
#' @export
filt_snps.raw_data <- function(dat, filt) {
  if (is.character(filt)) {
    filt <- which(dat$snps %in% filt)
  }
  if (length(filt) > 0) {
    if (is.matrix(dat$raw))
      dat$raw <- dat$raw[-filt,]
    if (is.vector(dat$snps))
      dat$snps <- dat$snps[-filt]
    if (is.matrix(dat$sd))
      dat$sd <- dat$sd[-filt,]
    if (is.matrix(dat$beads))
      dat$beads <- dat$beads[-filt,]
    if (is.vector(dat$pos))
      dat$pos <- dat$pos[-filt]
    if (is.vector(dat$chr))
      dat$chr <- dat$chr[-filt]
  }
  dat
}


#' @rdname filt_snps
#' @export
filt_snps.norm_data <- function(dat, filt) {
  if (is.character(filt)) {
    filt <- which(dat$snps %in% filt)
  }
  if (length(filt) > 0) {
    if (is.matrix(dat$intensity))
      dat$intensity <- dat$intensity[-filt, ]
    if (is.matrix(dat$theta))
      dat$theta <- dat$theta[-filt, ]
    if (is.matrix(dat$baf))
      dat$baf <- dat$baf[-filt, ]
    if (is.matrix(dat$rratio))
      dat$rratio <- dat$rratio[-filt, ]
    if (is.matrix(dat$geno))
      dat$geno <- dat$geno[-filt,]
    if (is.matrix(dat$cnv))
      dat$cnv <- dat$cnv[-filt,]
    if (is.vector(dat$pos))
      dat$pos <- dat$pos[-filt]
    if (is.vector(dat$chr))
      dat$chr <- dat$chr[-filt]
    if (is.vector(dat$snps))
      dat$snps <- dat$snps[-filt]
    if (is.matrix(dat$cnv))
      dat$cnv <- dat$cnv[-filt, ]
    if (is.matrix(dat$beads))
      dat$beads <- dat$beads[-filt, ]
    if (is.matrix(dat$sd))
      dat$sd <- dat$sd[-filt, ]
  }
  dat
}

#' Filter Copy number variations
#'
#' Short CNV stretches can be filtered out.
#'
#' @param dat norm_data object.
#' @param thresh Integer, lower threshold for filtering CNVs.
#' @return norm_data object
#' @examples
#' \dontrun{
#' if(require(brassicaData)){
#' data("raw_napus", package = "brassicaData", envir = environment())
#' dat <- intens_theta(raw_napus)
#' dat <- remove_suffix(dat, "_Grn")
#' dat <- geno_baf_rratio(dat, delthresh = 11)
#' dat <- segm(dat)
#' dat <- cnv(dat, dup = 0.03, del = -0.06)
#' dat <- filter_cnv(dat)
#' }
#' }
#' @export
filter_cnv <- function(dat, thresh = 5) {
  stub_zero <- function(x, y) {
    run <- rle(x)
    chpos <- which(run$lengths < y & run$values == 1)
    chneg <- which(run$lengths < y & run$values == -1)
    if (length(chpos) > 0) {
      if (chpos[1] == 1)
        x[1:run$lengths[1]] <- 0
      for (i in chpos) {
        start <- sum(run$lengths[1:(i - 1)])
        x[(start + 1):(start + run$lengths[i])] <- 0
      }
    }
    if (length(chneg) > 0) {
      if (chneg[1] == 1)
        x[1:run$lengths[1]] <- 0
      for (i in chneg) {
        start <- sum(run$lengths[1:(i - 1)])
        x[(start + 1):(start + run$lengths[i])] <- 0
      }
    }
    x
  }
  dat$cnv <- apply(dat$cnv, 2, stub_zero, y = thresh)
  dat
}