R/createRegulonList.R

In saezlab/CARNIVAL-Bioconductor-Dev: A CAusal Reasoning tool for Network Identification (from gene expression data) using Integer VALue programming

#'\code{createRegulonList}
#'
#' Creating regulon list similar to vipper from an interaction data-frame
#' 
#' @param regulon_table the regulon matrix with columns the source, sign and
#' target of the TF interaction.
#'
#'Enio Gjerga, 2020

createRegulonList <- function(regulon_table = regulon_table){
  
  
  if((!is.character(regulon_table[, 1])) || 
     !is.character(regulon_table[, 3])){
    
    stop("Source and Target columns should be of class character")
    
  }
  
  if(!((is.character(regulon_table[, 2])) || 
     (is.numeric(regulon_table[, 2])))){
    
    stop("Interaction sign column should either be of class character/numeric")
    
  }
  
  sgn <- as.numeric(as.character(regulon_table[, 2]))
  if(!(all(sgn%in%c(1, -1)))){
    stop("Please check your interaction sign column. 
         They should be either 1/-1.")
  }
  
  regulon <- list()
  sourceList <- unique(regulon_table[, 1])
  
  for(ii in 1:length(sourceList)){
    
    currRegulon <- list()
    vec <- as.numeric(
      as.character(regulon_table[which(regulon_table[, 1]==sourceList[ii]), 2]))
    names(vec) <- regulon_table[which(regulon_table[, 1]==sourceList[ii]), 3]
    
    currRegulon[[length(currRegulon)+1]] <- vec
    currRegulon[[length(currRegulon)+1]] <- rep(1, length(vec))
    
    names(currRegulon) <- c("tfmode", "likelihood")
    
    regulon[[length(regulon)+1]] <- currRegulon
    
  }
  
  names(regulon) <- sourceList
  
  return(regulon)
  
}