R/epiMap.R

In TheJacksonLaboratory/epihet: Determining Epigenetic Heterogeneity from Bisulfite Sequencing Data

#' @title Make Pheatmap from Comparison Matrix
#' @description
#' Creates a pheatmap for the top 'loci.percent' of values
#' of max standard deviation from the comparison matrix
#' generated by compMatrix(). The rows represent the loci
#' of the epiallele and the columns represent the sample
#' names. The columns can be annotated by adding annotation
#' information as a parameter.
#' @param compare.matrix The comparison matrix generated from
#' the compMatrix() function
#' @param value The value to be graphed in the pheatmap.
#' Possible values are 'read', 'pdr', 'meth', 'epipoly',
#' and 'shannon'.
#' @param annotate A dataframe containing the annotation information
#' for the columns of the pheatmap. The row names must be the names
#' of the samples. The columns (any number) are the annotations. E.g. a column
#' called 'TET2' with factors 'Pos' and 'Neg' for each sample that
#' is positive or negative for the TET2 gene
#' @param clustering_distance_rows Distance measure used in clustering rows.
#' @param clustering_distance_cols Distance measure used in clustering columns.
#' @param clustering_method clustering method used.
#' @param annotate.colors A list containing the colors for the
#' annotation information. Each element in the list is a vector
#' of colors with names that correspond to the columns of 'annotate'.
#' @param color a vector of colors used in heatmap.
#' @param loci.percent The top percentage of loci, as a decimal,
#' to be displayed on the pheatmap based on standard deviation,
#' e.g. a value of 0.20 is equivalent to the top 20\% of loci
#' (default: 0.10)
#' @param show.rows A boolean stating if the row names should be
#' displayed on the pheatmap (default: FALSE)
#' @param show.columns A boolean stating if the column names should
#' be displayed on the pheatmap (default: FALSE)
#' @param font.size An integer representing the font size to be used
#' for the pheatmap labels (default: 6)
#' @param pdf.height An integer representing the height (in inches)
#' of the pdf file for the pheatmap (default: 10)
#' @param pdf.width An integer representing the width (in inches)
#' of the pdf file for the pheatmap (default: 10)
#' @param sve A boolean to save the plot (default: FALSE)
#' @param ... any arguments in the function pheatmap()
#' @return A pheatmap object that contains the tree data for
#' both rows and columns and the final pheatmap plot
#' @examples
#' comp.Matrix<-data.frame(
#' p1=c(0.6,0.3,0.5,0.5,0.5,0.6,0.45,0.57,0.45,0.63,0.58,0.67,0.5,0.42,0.67),
#' p2=c(0.62,0.63,0.55,0.75,0.84,0.58,1,0.33,1,0.97,0.57,0.68,0.73,0.72,0.82),
#' p3=c(0.72,0.53,0.62,0.69,0.37,0.85,1,0.63,0.87,0.87,0.82,0.81,0.79,
#' 0.62,0.68),
#' N1=c(0.15,0.24,0.15,0.26,0.34,0.32,0.23,0.14,0.26,0.32,0.12,0.16,0.31,
#' 0.24,0.32),
#' N2=c(0.32,0.26,0.16,0.36,0.25,0.37,0.12,0.16,0.41,0.47,0.13,0.52,0.42,
#' 0.41,0.23),
#' N3=c(0.21,0.16,0.32,0.16,0.36,0.27,0.24,0.26,0.11,0.27,0.39,0.5,0.4,
#' 0.31,0.33),
#' type=rep(c("pdr","epipoly","shannon"),c(5,5,5)),
#' location=rep(c("chr22-327:350:361:364","chr22-755:761:771:773",
#' "chr22-761:771:773:781","chr22-821:837:844:849","chr22-838:845:850:858"),
#' 3),stringsAsFactors =FALSE )
#'
#' subtype <- data.frame(Type= c(rep('CEBPA_sil', 3), rep('Normal', 3)),
#' row.names = colnames(comp.Matrix)[1:6],stringsAsFactors = FALSE)
#'
#' pmap <- epiMap(compare.matrix = comp.Matrix,
#' value = 'epipoly',annotate = subtype,
#' clustering_distance_rows = "euclidean",
#' clustering_distance_cols = "euclidean",
#' clustering_method = "complete",annotate.colors = NA,
#' color= colorRampPalette(c("blue","white","red"))(1000),
#' loci.percent = 1, show.rows = FALSE,
#' show.columns = TRUE, font.size = 15,
#' pdf.height = 10, pdf.width = 10, sve = TRUE)
#' @export
epiMap <- function(compare.matrix, value, annotate,
    clustering_distance_rows = "euclidean",
    clustering_distance_cols = "euclidean",
    clustering_method = "complete", annotate.colors = NA,
    color = colorRampPalette(c("blue", "white", "red"))(1000),
    loci.percent = 0.1, show.rows = FALSE, show.columns = FALSE,
    font.size = 6, pdf.height = 10, pdf.width = 10,
    sve = FALSE, ...) {
    values = c("read", "pdr", "meth", "epipoly", "shannon", "myValues")
    if (!(value %in% values)) {
        stop("Invalid value '", value, "': Possible values are 'read',
           'pdr', 'meth', 'epipoly', 'myValues' or 'shannon'")
    }
    value.matrix <- compare.matrix[compare.matrix$type == value,
        -(length(compare.matrix) - 1)]
    rownames(value.matrix) <- value.matrix$location
    value.matrix <- value.matrix[, -length(value.matrix)]
    value.matrix$sd = apply(value.matrix, 1, sd, na.rm = TRUE)
    value.matrix <- value.matrix[order(-value.matrix$sd), ]
    value.matrix <- value.matrix[, -length(value.matrix)]
    loci.number <- floor(loci.percent * nrow(value.matrix))
    title <- paste0("Top ", loci.percent * 100, "% of ", value, " values")
    matrix <- value.matrix[seq_len(loci.number), ]
    if (sve) {
        pmap <- pheatmap::pheatmap(matrix, show_rownames = show.rows,
            show_colnames = show.columns, color = color,
            clustering_distance_rows = clustering_distance_rows,
            clustering_distance_cols = clustering_distance_cols,
            clustering_method = clustering_method,
            annotation_col = annotate, fontsize = font.size,
            annotation_colors = annotate.colors, height = pdf.height,
            width = pdf.width, main = title,
                filename = paste0(value, "_pheatmap.pdf"), ...)
    } else {
        pmap <- pheatmap::pheatmap(matrix, show_rownames = show.rows,
            show_colnames = show.columns, color = color,
            clustering_distance_rows = clustering_distance_rows,
            clustering_distance_cols = clustering_distance_cols,
            clustering_method = clustering_method,
            annotation_col = annotate, fontsize = font.size,
            annotation_colors = annotate.colors, main = title, ...)
    }
    pmap
}