R/module.overlap.R

In DEComplexDisease: A tool for differential expression analysis and DEGs based investigation to complex diseases by bi-clustering analysis

#' Plot the overlap among predicted DEG modules
#'
#' Plot the overlap among predicted DEG modules
#'
#' @docType methods
#' @import ComplexHeatmap
#'
#' @name module.overlap
#' @param res.module a 'seed.module' or 'cluster.module' object returned by \code{\link{seed.module}} or \code{\link{cluster.module}}
#' @param show.mods the modules to display
#' @param type the module type to display
#' @param max.n the maximum number of modules to plot. If 'show.mods' is set, this option will be ignored.
#' @param show.overlap boolen, display the overlap number
#' @param cex the font cex to display the overlap number, default 10.
#'
#'
#' @author Guofeng Meng
#'
#'
#'
#' @details The DEG modules may have partial overlaps for either genes or patients. This function plots the overlaps for genes and patients from DEG modules in two heatmaps. It can be used to find DEG patterns.
#'
#' @return The heatmap plot for gene and patient overlaps.
#'
#' @examples
#' data(seed.mod)
#' module.overlap(seed.mod, max.n=5,type='model')
#' @export

module.overlap <- function(res.module, show.mods = NULL, type = c("model", "max.patients", 
    "max.genes")[1], max.n = 30, show.overlap = TRUE, cex = 10) {
    if (!is(res.module, "seed.module") & !is(res.module, "cluster.module")) 
        stop("Error: reg.module: must be the output of 'seed.module' or 'cluster.module'!")
    if (!any(c("model", "max.patients", "max.genes") == type)) 
        stop("Error: type: should one of model, max.patients and max.genes!")
    mods = names(res.module)
    mods = mods[mods != "decd.specific" & mods != "decd.input" & mods != "decd.clustering"]
    if (is.null(res.module[[1]][[type]])) 
        stop("Error: res.module has no selected type")
    if (!is.null(show.mods)) {
        show.mods = show.mods[show.mods %in% mods]
    } else {
        show.mods = .select.mod(res.module, max.n, type = type)
    }
    n = length(show.mods)
    olp1 = matrix(ncol = n, nrow = n)
    olp2 = matrix(ncol = n, nrow = n)
    olp3 = matrix(ncol = n, nrow = n)
    olp4 = matrix(ncol = n, nrow = n)
    for (i in seq_len(n)) {
        pas1 = res.module[[show.mods[i]]][[type]][["patients"]]
        ges1 = res.module[[show.mods[i]]][[type]][["genes"]]
        for (j in seq_len(n)) {
            pas2 = res.module[[show.mods[j]]][[type]][["patients"]]
            ges2 = res.module[[show.mods[j]]][[type]][["genes"]]
            
            olp1[i, j] = olp1[j, i] = length(intersect(pas1, pas2))
            olp2[i, j] = olp2[j, i] = length(intersect(ges1, ges2))
            
        }
    }
    for (i in seq_len(n)) {
        for (j in seq_len(n)) {
            olp3[i, j] = olp1[i, j]/olp1[j, j]
            olp4[i, j] = olp2[i, j]/olp2[j, j]
        }
    }
    row.names(olp1) <- show.mods
    row.names(olp2) <- show.mods
    colnames(olp1) <- show.mods
    colnames(olp2) <- show.mods
    row.names(olp3) <- show.mods
    row.names(olp4) <- show.mods
    colnames(olp3) <- show.mods
    colnames(olp4) <- show.mods
    if (length(unique(as.vector(olp1))) > 1) {
        if (show.overlap) {
            Heatmap(olp3, cluster_rows = TRUE, cluster_columns = TRUE, name = "Patients", 
                column_title = "Patients", cell_fun = function(j, i, x, y, w, h, 
                  col) {
                  grid.text(olp1[i, j], x, y, gp = gpar(fontsize = cex))
                }) + Heatmap(olp4, cluster_rows = TRUE, cluster_columns = TRUE, name = "Genes", 
                column_title = "genes", cell_fun = function(j, i, x, y, w, h, col) {
                  grid.text(olp2[i, j], x, y, gp = gpar(fontsize = cex))
                })
        } else {
            Heatmap(olp3, cluster_rows = TRUE, cluster_columns = TRUE, name = "Patients", 
                column_title = "Patients") + Heatmap(olp4, cluster_rows = TRUE, cluster_columns = TRUE, 
                name = "Genes", column_title = "genes")
        }
    } else {
        print("Warning: all the modules have the same patients")
        if (show.overlap) {
            Heatmap(olp4, cluster_rows = TRUE, cluster_columns = TRUE, name = "Genes", 
                column_title = "genes", cell_fun = function(j, i, x, y, w, h, col) {
                  grid.text(olp2[i, j], x, y, gp = gpar(fontsize = cex))
                })
        } else {
            Heatmap(olp4, cluster_rows = TRUE, cluster_columns = TRUE, name = "Genes", 
                column_title = "genes")
        }
    }
}