R/tr_anno_cleaner.R

In travis-m-blimkie/tRavis: tRavis: an R package by Travis Blimkie

#' Clean annotations from pseudomonas.com
#'
#' @param input_file Path to the input TSV or CSV file.
#' @param extra_cols Logical to determine if start, end, and strand columns
#'   should be included. Defaults to `FALSE`.
#' @param fill_names Logical to determine if blank/NA genes names should be
#'   filled in with corresponding locus tag. Defaults to `FALSE`.
#'
#' @return A data frame (tibble) of the cleaned input file
#' @export
#'
#' @import dplyr
#' @importFrom janitor clean_names
#' @importFrom readr cols read_delim
#'
#' @description Given an input CSV or TSV annotation file from
#'   <https://pseudomonas.com>, separates and cleans the data, returning a tidy
#'   tibble with the following columns: "locus_tag", "gene_name", and
#'   "product_name". Setting the `extra_cols` argument to `TRUE` will add the
#'   columns "start", "end", and "strand". Enabling `fill_names` will populate
#'   missing gene names with the corresponding locus tag.
#'
#' @references Download annotation files from <https://pseudomonas.com>
#' @seealso <https://www.github.com/travis-m-blimkie/tRavis>
#'
#' @examples
#' tr_anno_cleaner(
#'   input_file = paste0(
#'   "https://pseudomonas.com/downloads/pseudomonas/pgd_r_22_1/",
#'   "Pseudomonas_aeruginosa_PAO1_107/Pseudomonas_aeruginosa_PAO1_107.csv.gz"
#'   )
#' )
#'
tr_anno_cleaner <- function(
    input_file,
    extra_cols = FALSE,
    fill_names = FALSE
) {

  file_type <- grep(x = input_file, pattern = "\\.(c|t)sv", value = TRUE)

  stopifnot("'input_file' must be a '.csv' or '.tsv' file" = !is.na(file_type))

  step1 <- read_delim(
    input_file,
    delim = switch(file_type, csv = ",", tsv = "\t"),
    col_types = cols()
  ) %>%
    clean_names()

  if (!extra_cols) {
    final_cols <- c("locus_tag", "gene_name", "product_name")
  } else {
    final_cols <-
      c("locus_tag", "gene_name", "product_name", "start", "end", "strand")
  }

  step2 <- step1 %>%
    select(all_of(final_cols)) %>%
    distinct(locus_tag, .keep_all = TRUE)

  # Conditionally fill blank gene names with corresponding locus tags
  step3 <- step2

  if (fill_names) {
    step3 <- mutate(
      step3,
      gene_name = if_else(is.na(gene_name), locus_tag, gene_name)
    )
  }

  return(step3)
}