R/gggenes_extensions.R

In djw533/micro.gen.extra: Microbial genomics extensions and functions

#' Create a dataframe for input to gggenes from a list of gff3 files
#'
#' Create a dataframe for input to gggenes from a list of gff3 files
#'
#' @param gff_list Directory containing gff3 files
#'
#' @return Dataframe with position, orientation and description of genomic features from gff files,
#' in a format for plotting gene diagrams with gggenes, or further input into genoplotR dnaseqs formatting.
#' @examples
#' gggenes_df_from_gff_dir("path/to/gff/directory")


gggenes_df_from_gff_list <- function(gff_list) {

  gggenes_df <- data.frame(matrix(nrow = 0 ,ncol = 9))

  colnames(gggenes_df) <- c("number","filename_prefix","start","end","gene","strand","direction","type","contig")

  for (file in gff_list) {
    print(file)

    #set the filename
    filename = tail(unlist(stringr::str_split(file,"/")),1)
    filename_prefix = tail(unlist(stringr::str_split(filename,"[.]gff")),2)[1]


    #replace any hashes in the name with an underscore and write a temporary file:
    changed_filename_prefix <- gsub("#","_",filename_prefix)

    system(glue::glue("sed 's/{filename_prefix}/{changed_filename_prefix}/g' {file} > temp_file.gff"))

    #read in the gff file and coerce into a format gggenes likes
    temp_df <- read.table(file = "temp_file.gff",
                          sep = "\t",
                          comment.char = "",
                          fill = TRUE,
                          quote = "",
                          header = F,
                          col.names = c("contig","method","type", "start","end","strand","direction","score", "details")) %>%
      #rename(contig = V1,method = V2,type = V3, start = V4,end = V5,strand = V6 ,direction = V7 ,score = V8, details  = V9) %>%
      filter(details != "")  %>% # remove lines where there isn't a 9th column
      mutate(filename_prefix = filename_prefix) %>%
      #filter(type == "CDS") %>% # only take CDSs
      # now - if there is ONE  "ID=" string within the details column, create variable "gene" and set as the string immediately after "ID=", but before the next ";". Otherwise set as NA.
      mutate(gene = as.character(purrr::map(details, ~ ifelse(stringr::str_detect(.x, "ID=") & length(unlist(stringr::str_split(.x,"ID="))) == 2,  unlist(stringr::str_split(unlist(stringr::str_split(.x,"ID="))[2],";"))[1] , NA  )))) %>%
      # put in number for gggenes:
      tibble::rownames_to_column(var = "number") %>%
      mutate(number = as.numeric(number)) %>%
      #set the direction and strands:
      mutate(direction = ifelse(direction == "+", 1, -1)) %>%
      mutate(strand = ifelse(direction == 1, "forward","reverse")) %>%
      # now select columns to rbind in:
      select(number,filename_prefix,start,end,gene,strand,direction,type,contig)


    gggenes_df <- rbind(gggenes_df,temp_df)

  }

  file.remove("temp_file.gff")


  return(gggenes_df)


}












#' Create a dataframe for input to gggenes from a directory of gff3 files
#'
#' Create a dataframe for input to gggenes from a directory of gff3 files
#'
#' @param gff_dir Directory containing gff3 files
#'
#' @return Dataframe with position, orientation and description of genomic features from gff files,
#' in a format for plotting gene diagrams with gggenes, or further input into genoplotR dnaseqs formatting.
#' @examples
#' gggenes_df_from_gff_dir("path/to/gff/directory")


gggenes_df_from_gff_dir <- function(gff_dir) {

  gff_files <- list.files(path = gff_dir, full.names = T, pattern = ".gff")

  gggenes_df <- data.frame(matrix(nrow = 0 ,ncol = 9))

  colnames(gggenes_df) <- c("number","filename_prefix","start","end","gene","strand","direction","type","contig")

  for (file in gff_files) {
    print(file)

    #set the filename
    filename = tail(unlist(stringr::str_split(file,"/")),1)
    filename_prefix = tail(unlist(stringr::str_split(filename,"[.]gff")),2)[1]


    #replace any hashes in the name with an underscore and write a temporary file:
    changed_filename_prefix <- gsub("#","_",filename_prefix)

    system(glue::glue("sed 's/{filename_prefix}/{changed_filename_prefix}/g' {file} > temp_file.gff"))

    #read in the gff file and coerce into a format gggenes likes
    temp_df <- read.table(file = "temp_file.gff",
                          sep = "\t",
                          comment.char = "",
                          fill = TRUE,
                          quote = "",
                          header = F,
                          col.names = c("contig","method","type", "start","end","strand","direction","score", "details")) %>%
      filter(details != "")  %>% # remove lines where there isn't a 9th column
      mutate(filename_prefix = filename_prefix) %>%
      #filter(type == "CDS") %>% # only take CDSs
      # now - if there is ONE  "ID=" string within the details column, create variable "gene" and set as the string immediately after "ID=", but before the next ";". Otherwise set as NA.
      mutate(gene = as.character(purrr::map(details, ~ ifelse(stringr::str_detect(.x, "ID=") & length(unlist(stringr::str_split(.x,"ID="))) == 2,  unlist(stringr::str_split(unlist(stringr::str_split(.x,"ID="))[2],";"))[1] , NA  )))) %>%
      # put in number for gggenes:
      tibble::rownames_to_column(var = "number") %>%
      mutate(number = as.numeric(number)) %>%
      #set the direction and strands:
      mutate(direction = ifelse(direction == "+", 1, -1)) %>%
      mutate(strand = ifelse(direction == 1, "forward","reverse")) %>%
      # now select columns to rbind in:
      select(number,filename_prefix,start,end,gene,strand,direction,type,contig)


    gggenes_df <- rbind(gggenes_df,temp_df)

  }

  file.remove("temp_file.gff")


  return(gggenes_df)


}



#' Create a dataframe for input to gggenes from a single gff3 file
#'
#' Create a dataframe for input to gggenes from a single gff3 file
#'
#' @param gff_file Gff file input
#'
#' @return Dataframe with position, orientation and description of genomic features from a single gff file,
#' in a format for plotting gene diagrams with gggenes, or further input into genoplotR dnaseqs formatting.
#' @examples
#' gggenes_df_from_gff_dir("path/to/gff/file")


gggenes_df_from_gff_file <- function(file) {


  #set the filename
  filename = tail(unlist(stringr::str_split(file,"/")),1)
  filename_prefix = tail(unlist(stringr::str_split(filename,"[.]gff")),2)[1]


  #replace any hashes in the name with an underscore and write a temporary file:
  changed_filename_prefix <- gsub("#","_",filename_prefix)

  system(glue::glue("sed 's/{filename_prefix}/{changed_filename_prefix}/g' {file} > temp_file.gff"))

  #read in the gff file and coerce into a format gggenes likes
  temp_df <- read.table(file = "temp_file.gff",
                        sep = "\t",
                        comment.char = "",
                        fill = TRUE,
                        quote = "",
                        header = F,
                        col.names = c("contig","method","type", "start","end","strand","direction","score", "details")) %>%
    filter(details != "")  %>% # remove lines where there isn't a 9th column
    mutate(filename_prefix = filename_prefix) %>%
    #filter(type == "CDS") %>% # only take CDSs
    # now - if there is ONE  "ID=" string within the details column, create variable "gene" and set as the string immediately after "ID=", but before the next ";". Otherwise set as NA.
    mutate(gene = as.character(purrr::map(details, ~ ifelse(stringr::str_detect(.x, "ID=") & length(unlist(stringr::str_split(.x,"ID="))) == 2,  unlist(stringr::str_split(unlist(stringr::str_split(.x,"ID="))[2],";"))[1] , NA  )))) %>%
    # put in number for gggenes:
    tibble::rownames_to_column(var = "number") %>%
    mutate(number = as.numeric(number)) %>%
    #set the direction and strands:
    mutate(direction = ifelse(direction == "+", 1, -1)) %>%
    mutate(strand = ifelse(direction == 1, "forward","reverse")) %>%
    # now select columns to rbind in:
    select(number,filename_prefix,start,end,gene,strand,direction,type,contig)


  file.remove("temp_file.gff")


  return(temp_df)


}