Plots a ggplot heatmap

ggheatmap() is the main function of the ggheatmap package. It constructs a tile-able plot using ggplot and patchwork, that can be added to and used in constructing panels.

ggheatmap(
  table,
  colv = NULL,
  rowv = NULL,
  hm_colors = "RdYlBu",
  hm_color_values = NULL,
  hm_color_breaks = waiver(),
  hm_color_limits = NULL,
  scale = FALSE,
  center = FALSE,
  cluster_rows = TRUE,
  cluster_cols = TRUE,
  dist_method = "euclidean",
  clustering_method = "complete",
  raster = FALSE,
  rows_title = "",
  column_title = "",
  colors_title = "value",
  fontsize = 11,
  show_rownames = TRUE,
  show_colnames = TRUE,
  show_dend_row = TRUE,
  show_dend_col = TRUE,
  dend_lwd = 0.3,
  dend_prop_row = 0.1,
  dend_prop_col = 0.1,
  group_track = TRUE,
  group_label = FALSE,
  group_track_topslack = 0,
  group_label_angle = 0,
  group_label_size = 3.5,
  group_label_position = "center",
  group_prop = 0.1,
  group_colors = NULL,
  group_lines = FALSE,
  group_line_color = "black",
  group_lty = "solid",
  group_lwd = 0.3,
  group_leg_ncol = 3,
  row_facetting_space = 3,
  colorbar_dir = "vertical"
)

Arguments

table

A table that can follow four formats:

  • If table is a data.frame or tbl and colv and rowv are provided, colv will be plotted into the columns (observations), while the variables provided as rowv will be plotted into the rows (variables).

  • If table is a grouped_tbl and cluster_cols = TRUE, a semi-supervised clustering of the data will be performed instead only clustering within groups.

  • If table is a data.frame or tbl and only colv, all other columns will be used as rowv.

  • If table is a matrix or data.frame and colv and rowv are NULL, ggheatmap will plot it as-is (columns will be plotted as columns, rows as rows of the heatmap). It is convenient, but less powerful.

colv

Either NULL or the name of the column that contains the observation ID variable for the data.

rowv

Either NULL, a vector with names of columns to be plotted into the rows of the heatmap, or a named list of vectors for a faceted plot (show_dend_row will then be ignored).

hm_colors

Can be either:

hm_color_values

Either NULL or a vector of values used for interpolation between colors. Passed to values from ggplot2::scale_fill_gradientn()

hm_color_breaks

Breaks for heatmap colors. Passed to breaks from ggplot2::scale_fill_gradientn()

hm_color_limits

A vector with length = 2 for limits of the heatmap color scale. Passed to limits from ggplot2::scale_fill_gradientn().

scale

If TRUE, data will be scaled (z-score)

center

If TRUE, data will be centered

cluster_rows

If TRUE, rows will be clustered with stats::hclust()

cluster_cols

If TRUE, columns will be clustered with stats::hclust()

dist_method

Distance method passed to stats::dist(). Also supports correlation distance (1-cor()), with dist_method passed to method from stats::cor

clustering_method

Clustering method is passed to method from stats::hclust()

raster

If TRUE, ggplot2::geom_raster() will be used for the heatmap tiles instead of ggplot2::geom_tile(). Will be recommended by the function for large tables.

rows_title

A title for row variables.

column_title

A title for the column variable.

colors_title

A title for the color legend.

fontsize

Base fontsize for plot, which will be used by the theme. Ultimately passed to ggplot2::theme_minimal() as base_size.

show_rownames

If TRUE, row names will be shown in the heatmap.

show_colnames

If TRUE, column names will be shown in the heatmap.

show_dend_row

If TRUE, the clustering dendrogram for the row variables will be shown to the left of the heatmap

show_dend_col

If TRUE, the clustering dendrogram for the column variable will be shown on top of the heatmap.

dend_lwd

Linewidth for the dendrogram drawing. Passed to ggtree::ggtree as size.

dend_prop_row

A value between 0 and 1. The proportion of the width occupied by the dendrogram.

dend_prop_col

A value between 0 and 1. The proportion of the height occupied by the dendrogram.

group_track

If table is a grouped_tbl and group_track = TRUE, a track will be plot between the dendrogram and heatmap.

group_label

a logical indicating whether to label the groups directly. If TRUE, show_dend_col will be set to FALSE.

group_track_topslack

a numeric, indicating space to add at the top of the group track. Useful for fitting names

group_label_angle

an angle for the group labels

group_label_size

a numeric value to set the size of the group labels

group_label_position

one of: "left", "right" or "center", indicating where to write the group label if `group_label = TRUE

group_prop

The proportion of the height of the heatmap that will be used for the group track.

group_colors

A named vector with colors for each level in the grouping variable. If NULL, automatic colors will be used.

group_lines

If TRUE, vertical lines will separate supervised clustering groups

group_line_color

Color of the vertical line

group_lty

Group line type. See graphics::par

group_lwd

Group line width. See graphics::par

group_leg_ncol

Number of columns in the groups legend. Passed to ggplot2::guide_legend().

row_facetting_space

space in pts between row facets, if they exist

colorbar_dir

Places the colorbar either horizontal or vertically. See: ggplot2::guide_colorbar

Value

A ggplot object with class ggheatmap.