R/priming_graph.R

In selcenari/ceRNAnetsim: Regulation Simulator of Interaction between miRNA and Competing RNAs (ceRNA)

#' @title normalize
#'
#' @description normalizes the values according to maximum values
#' inside a group. The helper function of priming_graph.
#'
#' @param x The variable name that is normalized.
#'
#' @return normalized values
#' @keywords internal
#'


normalize <- function(x) {
  abs(x)/max(abs(x))
}

#' Converts the given dataframe using first variable as
#' competing and the second as miRNA.
#' The function converts the given dataframe using first
#' variable as competing and the second as miRNA. If user
#' defines interaction factors as affinity or degradation,
#' the factors are taken into account.
#'
#' @importFrom purrr reduce
#' @importFrom igraph edges edge
#' @importFrom stats na.omit
#' @importFrom rlang enquo enquos
#' @return the graph object.
#'
#' @param df A data frame that includes the miRNA and competing targets.
#' @param competing_count The counts (or expression) of competing elements of the dataset.
#' @param miRNA_count The counts (or expression) of repressive element (miRNA) of the dataset.
#' @param aff_factor The parameter/s of binding between miRNA and targets.
#' @param deg_factor The parameter/s for degradation of bound miRNA:target complex.
#'
#' @details priming_graph provides grouping of competing targets
#' and evaluation of targets within the groups taking into account
#' miRNA:target, target:total target, interaction and degradation
#' parameters. The target groups are determined according to miRNAs.
#' If the factors that are important in target interactions are
#' specified as arguments, the factors also are evaluated separately
#' within each group. priming_graph also calculates the miRNA efficiency
#' in steady-state conditions. It is assumed that quantity of competing
#' targets and miRNAs are shown in the steady-state system after the
#' miRNAs exhibit repressive efficiency. Note that the data must not
#' include missing values such as NA or '-'.
#'
#'
#' @examples
#'
#' data('minsamp')
#'
#' priming_graph(minsamp, Competing_expression, miRNA_expression)
#'
#' priming_graph(minsamp, Competing_expression, miRNA_expression,
#'     aff_factor = c(seed_type,energy), deg_factor = region)
#'
#'
#' @export

priming_graph <- function(df, competing_count, miRNA_count, aff_factor = dummy,
                          deg_factor = dummy) {

  if (any(is.na(df))) {

    warning(paste("Dataframe includes", sum(is.na(df)), "NA values. Dataframe will be processed after NA removing. "))
    df <- na.omit(df)

  }

  competing_exp <- enquo(competing_count)
  mirna_exp <- enquo(miRNA_count)
  affinity <- enquos(aff_factor)
  degradation <- enquos(deg_factor)

  df <- df %>% dplyr::mutate(competing = .[[1]], miRNA = .[[2]], Competing_name = .[[1]],
                             miRNA_name = .[[2]], dummy = 1) %>% dplyr::select(competing, miRNA,
                                                                               Competing_name, miRNA_name, !!competing_exp, !!mirna_exp, !!!affinity,
                                                                               !!!degradation, dummy)


  input_graph <- df %>% dplyr::group_by(miRNA) %>% dplyr::mutate_at(dplyr::vars(!!!affinity),
                                                                    list(anorm = ~normalize(.))) %>% 
                                                   dplyr::mutate_at(dplyr::vars(!!!degradation),
                                                                    list(dnorm = ~normalize(.))) %>% 
                                                   dplyr::ungroup() %>% 
    dplyr::mutate(afff_factor = dplyr::select(., dplyr::ends_with("anorm")) %>% 
                    reduce(`*`, .init = 1), degg_factor = dplyr::select(., dplyr::ends_with("dnorm")) %>% 
                    reduce(`*`, .init = 1)) %>% 
    tidygraph::as_tbl_graph() %>%
    tidygraph::activate(nodes) %>% tidygraph::mutate(type = ifelse(tidygraph::centrality_degree(mode = "in") >
                                                                     0, "miRNA", "Competing")) %>% tidygraph::mutate(node_id = dplyr::row_number()) %>%
    tidygraph::activate(edges) %>% tidygraph::mutate(comp_count_list = as.list(!!competing_exp),
                                                     comp_count_pre = !!competing_exp, comp_count_current = !!competing_exp,
                                                     mirna_count_list = as.list(!!mirna_exp), mirna_count_pre = !!mirna_exp,
                                                     mirna_count_current = !!mirna_exp) %>% tidygraph::group_by(to) %>%
    tidygraph::mutate(mirna_count_per_dep = mirna_count_current * comp_count_current *afff_factor/sum(comp_count_current * afff_factor), 
                      mirna_count_per_dep = ifelse(is.na(mirna_count_per_dep),0, mirna_count_per_dep)) %>% 
    tidygraph::ungroup() %>% tidygraph::mutate(effect_current = mirna_count_per_dep * degg_factor, 
                                               effect_pre = effect_current, 
                                               effect_list = as.list(effect_current)) %>%
    tidygraph::select(-dplyr::ends_with("norm"), dummy)

  warning("First column is processed as competing and the second as miRNA.
")


  return(input_graph %>% update_nodes(once = TRUE))

}