R/heatmap_dge.R

In erkutilaslan/biovizR: A Package to Visualize Biological Data Designed for Wet Lab Scientists.

#' Heatmap for differential gene expression data.
#'
#' This function visualizes differential gene expresion data
#' as heatmap.
#'
#' @param heatmap_data Path to the input file.
#' @param gene_list Path to the list of genes file which will be visualized.
#' @param header Title for the heatmap.
#' @return A heatmap of heatmap_data.
#' @import ComplexHeatmap
#' @export

heatmap_dge <- function(heatmap_data,
		       	gene_list = "",
			header = "") {

#data import
if (is.character(heatmap_data) == TRUE) {
 
  if (grepl(".csv", as.character(heatmap_data)) == TRUE) {
    
    heatmap_data <- read.csv(heatmap_data, header = TRUE) 

  } else {

    heatmap_data <- readxl::read_excel(heatmap_data, sheet = 1)

  }

}

#NA omit otherwise we can't move HGNC to rownames
heatmap_data  <- na.omit(heatmap_data)

#removing row names
heatmap_data <- tibble::remove_rownames(heatmap_data)

#moving HGNC to rownames
colnames(heatmap_data)[1] <- "hgnc_symbol"
heatmap_data <- tibble::column_to_rownames(heatmap_data, var = "hgnc_symbol")

#z-score calculation
heatmap_data <- t(scale(t(heatmap_data)))

#selecing of genes to visualize e.g. infertility vectors
if (gene_list != "") {

  gene_list <- read.table(gene_list, header = FALSE, sep = " ")
  gene_list <- gene_list$V1


  #need to take the columns with the genes i want to visualize
  heatmap_data <- heatmap_data[as.character(gene_list), ]

  }

#removing na values once a mfing time
heatmap_data  <- na.omit(heatmap_data)

# Visualization
ComplexHeatmap::Heatmap(heatmap_data,
                        heatmap_legend_param = list(title = "z-score",
                                                    legend_height = grid::unit(5, "cm"),
                                                    grid_width = grid::unit(0.75, "cm"),
                                                    labels_gp = grid::gpar(fontsize = 12),
                                                    title_gp = grid::gpar(fontsize = 14)),
                        column_title_gp = grid::gpar(fontsize = 24),
                        column_title = header,
                        row_names_side = "left",
                        row_names_gp = grid::gpar(fontsize = 14),
                        column_names_gp = grid::gpar(fontsize = 16),
                        cluster_columns = FALSE,
                        show_row_dend = FALSE,
                        clustering_method_rows = "ward.D2",
                        clustering_distance_rows = "pearson")

}