R/get_Formula.R

In lishensuo/M2D: find Molecule(s) to Disease by reverse transcriptome

#' get two types of Formula for disease
#'
#' @param score_RRS result of score_RRS()
#' @param enrich_double result of gsea_disease_set()
#'
#' @return two types of Formula for disease
#' @export
#'
#' @examples
#' library(M2D)
#' data("score_RRS_result")
#' data("gsea_disease_set_result")
#' formula_list <- get_Formula(
#'   score_RRS = score_RRS_result,
#'   enrich_double = gsea_disease_set_result$gsea_sig
#' )
#' formula_1 <- formula_list$formula1
#' head(formula_1)
#' formula_2 <- formula_list$formula2
get_Formula <- function(score_RRS,
                        enrich_double) {
  # idea1: best single molecule
  ## 根据method1的逆向调控总分，计算小分子排名；
  ## 再取method2 significant molecule 交集
  formula1 <- score_RRS %>%
    dplyr::select(molecule, RRS_sum) %>%
    dplyr::distinct(molecule, .keep_all = T) %>%
    dplyr::arrange(dplyr::desc(RRS_sum)) %>%
    dplyr::mutate(Rank = 1:length(unique(score_RRS$molecule))) %>%
    dplyr::inner_join(enrich_double[, c("molecule", "Type")]) %>%
    dplyr::distinct(molecule, .keep_all = T)

  # idea2: formula molecules for as many as possible pathways
  ## step1: 挑选出每个通路的候选小分子(score>0)
  ## step2: 针对每个通路的候选小分子，根据总调控分数，选取最佳小分子
  formula2 <- score_RRS %>%
    dplyr::select(Term, molecule, RRS, RRS_sum) %>%
    dplyr::filter(
      RRS > 0,
      molecule %in% enrich_double$molecule
    ) %>%
    dplyr::arrange(Term, RRS_sum) %>%
    dplyr::distinct(Term, .keep_all = T)
  return(list(
    formula1 = formula1,
    formula2 = formula2
  ))
}