R/NR.R

In emosca-cnr/dmfind002: Network diffusion-based analysis of omics data

#' Local Monte Carlo p values
#' @param edge_list edge list
#' @param ranked_vector scores vector; it must have the same names and size of the vertices of G
#' @param k number of permutations
#' @param mc.cores number of cores
#' @export
#' @import parallel


NR <- function(edge_list, ranked_vector, k=99, mc.cores=1) {

  #real values
  ranked_vector_norm <- ranked_vector / max(ranked_vector)

  omega_vect <- omega(edge_list, ranked_vector_norm)

  #permutations of ranked_vector_norm
  #idx.perm <- lapply(1:k, function(x) array(sample(1:length(dS), length(dS)), dimnames = list(names(dS))))
  idx.perm <- lapply(1:k, function(x) array(sample(1:length(ranked_vector_norm)), dimnames = list(names(ranked_vector_norm))))

  if(mc.cores > 1){
    res <- parallel::mclapply(idx.perm, omega_perm, edge_list, ranked_vector_norm, mc.cores = mc.cores)
  }else{
    res <- lapply(idx.perm, omega_perm, edge_list, ranked_vector_norm)
  }

  #n-by-k matrix
  res <- t(do.call(rbind, res))

  #LMC p values
  out <- apply(res, 2, function(x) sign(x >= omega_vect))
  out <- rowSums(out)
  out <- (out + 1) / (k+1)

  return(list(NR_summary=data.frame(id=names(out), rank=1:length(ranked_vector_norm), ranking_score=ranked_vector_norm, omega=omega_vect, p=out, stringsAsFactors = FALSE), omega_perm=res))

}