R/extract_gene_locations.R

In khughitt/EuPathDB: Provides access to pathogen annotation resources available on EuPathDB databases

#' Clean up the gene location field from eupathdb derived gene location data.
#'
#' The eupathdb encodes its location data for genes in a somewhat peculiar
#' format: chromosome:start..end(strand), but I would prefer to have these
#' snippets of information as separate columns so that I can do things like
#' trivially perform rpkm().
#'
#' @param annot_df Data frame resulting from load_orgdb_annotations()
#' @param location_column Name of the column to extract the start/end/length/etc from.
#' @return Somewhat nicer data frame.
#' @author atb
#' @export
extract_gene_locations <- function(annot_df, location_column = "annot_gene_location_text") {
  newdf <- annot_df %>%
    tidyr::separate(location_column,
                    c("chromosome", "location"), ":") %>%
    tidyr::separate("location", c("start", "end"), "\\.\\.") %>%
    tidyr::separate("end", c("end", "strand"), "\\(")

  newdf[["start"]] <- as.numeric(gsub(pattern = "\\,", replacement = "", x = newdf[["start"]]))
  newdf[["end"]] <- as.numeric(gsub(pattern = "\\,", replacement = "", x = newdf[["end"]]))
  newdf[["strand"]] <- as.factor(gsub(pattern = "\\)", replacement = "", x = newdf[["strand"]]))
  newdf[["length"]] <- abs(newdf[["start"]] - newdf[["end"]])
  return(newdf)
}