R/GetPaths.R

In jackbibby1/SCPA: Single Cell Pathway Analysis

#' Convert gene sets to filtered pathway list
#'
#' This function takes one of three inputs:
#' 1) csv file where pathway title is in the first column
#' and subsequent columns contain genes of that pathway
#' 2) A gmt file
#' 3) A tidy formatted list where pathway name is contained in the first column
#' and the genes of that pathway are in the second column
#'
#' @param pathway_filepath filepath to csv or gmt file
#'
#' @examples \dontrun{
#' pathways <- get_paths(
#'      "Documents/gene_list.csv"
#' )
#' }
#'
#' @return list of pathways with corresponding genes
#' @export


get_paths <- function(pathway_filepath) {

  filepath_test <- as.character(pathway_filepath)
  number_of_files <- length(pathway_filepath)

    message("Generating gene sets from: \n", paste(filepath_test, collapse = "\n"), "\n")

    if (all(stringr::str_ends(filepath_test, "gmt"))) {

      pathways <- lapply(pathway_filepath, function(x) {

        clustermole::read_gmt(x, geneset_label = "Pathway", gene_label = "Genes") %>%
          dplyr::group_split(Pathway) %>%
          lapply(function(x) x[x$Genes != "", ])

      })

      pathways <- purrr::list_c(pathways)

    } else if (all(stringr::str_ends(filepath_test, "csv"))) {

      pathways <- utils::read.csv(pathway_filepath, row.names = 1, header = F)
      pathways <- as.data.frame(t(pathways))
      pathways <- tidyr::pivot_longer(pathways, cols = 1:length(pathways), names_to = "Pathway", values_to = "Genes")
      pathways <- dplyr::group_split(pathways, Pathway)
      pathways <- lapply(pathways, function(x) x[x$Genes != "",])

    } else {

      stop("Gene set file extension needs to be .csv or .gmt")

    }

}