Is there a way to write a genbank file in R?

Question

[Jack]

Hi All,

I'm trying to edit some genbank file fields in R (removing whitespace from gene feature tags) and write the updated output to a new file. Whilst I can read .gbk files (using gggenomes) and perform the edits I want, I cannot for the life of me find any R package that allows me to write a genbank file.

Does anyone know of a way to do this? I noticed BioPython has a number of options but I'm not as familiar with Python so would struggle more with performing the edits I need to make, though if no R option is available I suppose I'll make the switch.

Thank you!

Answer 1

[yura.grabovska]

From the gggenomes github:

#' Read genbank files
#'
#' Genbank flat files (.gb/.gbk/.gbff) and their ENA and DDBJ equivalents have a
#' particularly gruesome format. That's why [read_gbk()] is just a wrapper
#' around a Perl-based `gb2gff` converter and [read_gff3()].
#'
#' @importFrom readr read_tsv
#' @inheritParams read_gff3
#' @export
#' @return tibble
read_gbk <- function(file, sources = NULL, types = NULL, infer_cds_parents = TRUE) {
  gb2gff <- base::system.file("exec/gb2gff", package = "gggenomes")

  if (file_is_zip(file) && file_ext(file, ignore_zip = FALSE) != "gz") {
    abort(str_glue("Decompressing for genbank only works with gzipped files, not `{suf}`"))
  }

  if (file_is_url(file) && file_is_zip(file)) {
    file <- pipe(str_glue("curl {file} | gzip -cd | {gb2gff} -S"))
  } else if (file_is_url(file)) {
    file <- pipe(str_glue("curl {file} | {gb2gff} -S"))
  } else if (file_is_zip(file)) {
    suf <- file_ext(file, ignore_zip = FALSE)
    if (suf != "gz") {
      abort(str_glue("Decompressing for genbank only works with gzipped files, not `{suf}`"))
    }
    file <- pipe(str_glue("gzip -cd {file} | {gb2gff} -S"))
  } else {
    file <- pipe(str_glue("{gb2gff} -S {file}"))
  }

  read_gff3(file,
    sources = sources, types = types,
    infer_cds_parents = infer_cds_parents
  )
}

All the R command here does is run gb2gff and then reads in the gff file using read_gff3()internally.

So while it looks like this package found a workaround by going via gff3. It has a converse write_gff3() function and then perhaps you can use other tools outside of this, such as https://bioinformatics.stackexchange.com/questions/11115/existing-tool-for-converting-gff3-to-genbank-gbk