R/load_bed.R

In andygxzeng/ECSI: Espresso - Somatic Mutation Calling using Contextual Error Models

#' Load Bed File as GRanges
#'
#' Loads in sequence ranges from bed file and presents it as a GRanges object. Can handle headers and different Chr naming conventions.
#'
#' @param file Location of bed file
#' @param genome Reference genome to use, default is hg19
#' @export
#' @examples
#' \dontrun{
#' regions_of_interest <- load_bed(file = "AML_exons.bed", genome = "hg19")
#' }
#' @return This function returns a \code{GRanges} object including:
#' \itemize{
#'	\item seqnames
#' 	\item ranges
#' }

load_bed <-
function(file, genome = c("hg19", "hg38")) {

  # Check that genome was specified by the user
  if(length(genome) > 1){ stop("Need to specify genome, either hg19 or hg38.") }

  # load as datatable (automatic handling of headers)
  bed_dt <- data.table::fread(file)

  # convert to data.frame, fill in colnames, standardize chr names
  bed_df <- as.data.frame(bed_dt)
  colnames(bed_df)[1:3] <- c("chr", "start", "stop")
  bed_df$chr <- ifelse(stringr::str_detect(bed_df$chr, "[Cc][Hh][Rr]"), stringr::str_replace(bed_df$chr, "[Cc][Hh][Rr]", ""), bed_df$chr)

  # convert to GRanges
  bed_gr <- GenomicRanges::GRanges(
    seqnames = paste0("chr", bed_df$chr),
    ranges = IRanges::IRanges(bed_df$start, bed_df$stop))

  # specify genome
  GenomeInfoDb::genome(bed_gr) = genome

  return(bed_gr)
}