R/get_inflection_points.R

In ytakemon/GINIR: Genetic inteRaction and EssenTiality neTwork mApper

#' @title Calculate inflection points of co-essential data frame
#' 
#' @description Calculates the inflection point of the positive and negative curves to determine threshold for 
#' co-essential genes. See `?coessential_map` for details.
#' 
#' @param input_coessential_df data frame, A data frame output from `coessential_map()`, Default: NULL.
#' @param test logical. Used for examples. Default: FALSE.
#'
#' @return A data frame containing rank at which the threshold should be drawn for positive and negative co-essential genes.
#' 
#' @details Description of output data frame
#' * `Inflection_point_pos_byRank` - Rank threshold for positive curve.
#' * `Inflection_point_neg_byRank` - Rank threshold for negative curve.
#' @md
#' 
#' @examples 
#' gretta_data_dir <- './GRETTA_example/'
#' gretta_output_dir <- './GRETTA_example_output/'
#' 
#' if(!dir.exists(gretta_data_dir)){
#'   download_example_data(".")
#' }
#' 
#' inflection_points <- get_inflection_points(input_coessential_df = coess_df, test = TRUE)
#' 
#' @rdname get_inflection_points
#' @export 
#' @importFrom dplyr mutate filter pull rename arrange case_when
#' @importFrom tibble tibble
#' @importFrom RootsExtremaInflections inflexi

get_inflection_points <- function(input_coessential_df = NULL, test = FALSE) {
  if (test == FALSE) {
    # Checkpoint
    if (is.null(input_coessential_df)) {
      stop("No coessential dataframe found!")
    }
    if (!is.data.frame(input_coessential_df)) {
      stop("Input is not a dataframe, please check the input")
    }
    
    All_res <- NULL
    gene_list <- unique(input_coessential_df$GeneNameID_A)
    for(g in seq_len(length(gene_list))){
      # g <- 1
      gene <- gene_list[g]
      
      # Calculate inflection point
      message(gene, " selected. This may take a few mins...")
      inflection_points <- NULL
      
      # Positive curve
      subset_df_pos <- input_coessential_df %>%
        dplyr::filter(.data$GeneNameID_A %in% gene) %>%
        dplyr::arrange(.data$estimate, .data$Padj_BH) %>%
        dplyr::filter(.data$estimate > 0)
      message("Calculating inflection point of positive curve.\n")
      x_pos <- subset_df_pos$Rank
      y_pos <- subset_df_pos$estimate
      fit_pos <- RootsExtremaInflections::inflexi(x_pos,
                                                  y_pos, 1, length(x_pos), 5, 5,
                                                  plots = FALSE,
                                                  doparallel = FALSE
      )
      fit_pos$an
      fit_pos$finfl
      inflection_point_pos <- fit_pos$finfl[2]
      
      # Negative curve
      subset_df_neg <- input_coessential_df %>%
        dplyr::filter(.data$GeneNameID_A %in% gene) %>%
        dplyr::arrange(.data$estimate, .data$Padj_BH) %>%
        dplyr::filter(.data$estimate < 0)
      message("Calculating inflection point of negative curve.\n")
      x_neg <- subset_df_neg$Rank
      y_neg <- subset_df_neg$estimate
      fit_neg <- RootsExtremaInflections::inflexi(x_neg,
                                                  y_neg, 1, length(x_neg), 5, 5,
                                                  plots = FALSE,
                                                  doparallel = FALSE
      )
      fit_neg$an
      fit_neg$finfl
      inflection_point_neg <- fit_neg$finfl[2]
      
      res <- tibble::tibble(
        GeneNameID_A = gene,
        Inflection_point_pos_byRank = inflection_point_pos,
        Inflection_point_neg_byRank = inflection_point_neg
      )
      All_res <- dplyr::bind_rows(All_res, res)
    }
    
    return(All_res)
  } else {
    message("This tool requires correlation coefficients to be calculated.")
  }
}