An object of class heatmapr includes all the needed information for producing a heatmap. The goal is to separate the pre-processing of the heatmap elements from the graphical rendering of the object, which could be done

(Please submit an issue on github if you have a feature that you wish to have added)

heatmaply_na is a wrapper for `heatmaply` which comes with defaults that are better for exploring missing value (NA) patterns. Specifically, the grid_gap is set to 1, and the colors include two shades of grey. It also calculates the is.na10 automatically.

heatmaply_cor is a wrapper for `heatmaply` which comes with defaults that are better for correlation matrixes. Specifically, the limits are set from -1 to 1, and the color palette is RdBu.

heatmaply(x, ...)

heatmaply_na(x, grid_gap = 1, colors = c("grey80", "grey20"), ...)

heatmaply_cor(x, limits = c(-1, 1), colors = cool_warm, ...)

# S3 method for default
heatmaply(
  x,
  colors = viridis(n = 256, alpha = 1, begin = 0, end = 1, option = "viridis"),
  limits = NULL,
  na.value = "grey50",
  row_text_angle = 0,
  column_text_angle = 45,
  subplot_margin = 0,
  cellnote = NULL,
  draw_cellnote = !is.null(cellnote),
  cellnote_color = "auto",
  cellnote_textposition = "middle right",
  cellnote_size = 12,
  Rowv = NULL,
  Colv = NULL,
  distfun = stats::dist,
  hclustfun = stats::hclust,
  dist_method = NULL,
  hclust_method = NULL,
  distfun_row = distfun,
  hclustfun_row = hclustfun,
  distfun_col = distfun,
  hclustfun_col = hclustfun,
  dendrogram = c("both", "row", "column", "none"),
  show_dendrogram = c(TRUE, TRUE),
  reorderfun = function(d, w) reorder(d, w),
  k_row = 1,
  k_col = 1,
  symm = FALSE,
  revC = symm || (is.dendrogram(Colv) & is.dendrogram(Rowv) & identical(Rowv, rev(Colv))),
  scale = c("none", "row", "column"),
  na.rm = TRUE,
  row_dend_left = FALSE,
  margins = c(NA, NA, NA, NA),
  ...,
  scale_fill_gradient_fun = NULL,
  grid_color = NA,
  grid_gap = 0,
  srtRow = NULL,
  srtCol = NULL,
  xlab = "",
  ylab = "",
  main = "",
  titleX = TRUE,
  titleY = TRUE,
  hide_colorbar = FALSE,
  key.title = NULL,
  return_ppxpy = FALSE,
  row_side_colors = NULL,
  row_side_palette = NULL,
  col_side_colors = NULL,
  col_side_palette = NULL,
  ColSideColors = NULL,
  RowSideColors = NULL,
  seriate = c("OLO", "mean", "none", "GW"),
  heatmap_layers = NULL,
  side_color_layers = NULL,
  dendrogram_layers = NULL,
  branches_lwd = 0.6,
  file = NULL,
  width = NULL,
  height = NULL,
  long_data = NULL,
  plot_method = c("ggplot", "plotly"),
  label_names = c("row", "column", "value"),
  fontsize_row = 10,
  fontsize_col = 10,
  cexRow = NULL,
  cexCol = NULL,
  subplot_widths = NULL,
  subplot_heights = NULL,
  colorbar_len = 0.3,
  colorbar_thickness = 30,
  colorbar_xanchor = if (row_dend_left) "right" else "left",
  colorbar_yanchor = "bottom",
  colorbar_xpos = if (row_dend_left) -0.1 else 1.1,
  colorbar_ypos = 0,
  showticklabels = c(TRUE, TRUE),
  dynamicTicks = FALSE,
  grid_size = 0.1,
  node_type = "heatmap",
  point_size_mat = NULL,
  point_size_name = "Point size",
  label_format_fun = function(...) format(..., digits = 4),
  labRow = NULL,
  labCol = NULL,
  custom_hovertext = NULL,
  col = NULL,
  dend_hoverinfo = TRUE,
  side_color_colorbar_len = 0.3
)

# S3 method for heatmapr
heatmaply(
  x,
  colors = viridis(n = 256, alpha = 1, begin = 0, end = 1, option = "viridis"),
  limits = NULL,
  na.value = "grey50",
  row_text_angle = 0,
  column_text_angle = 45,
  subplot_margin = 0,
  row_dend_left = FALSE,
  margins = c(NA, NA, NA, NA),
  ...,
  scale_fill_gradient_fun = scale_fill_gradientn(colors = if (is.function(colors))
    colors(256) else colors, na.value = na.value, limits = limits),
  grid_color = NA,
  grid_gap = 0,
  srtRow = NULL,
  srtCol = NULL,
  xlab = "",
  ylab = "",
  main = "",
  titleX = TRUE,
  titleY = TRUE,
  hide_colorbar = FALSE,
  key.title = NULL,
  return_ppxpy = FALSE,
  draw_cellnote = FALSE,
  cellnote_color = "auto",
  cellnote_textposition = "middle right",
  cellnote_size = 12,
  row_side_colors = x[["row_side_colors"]],
  row_side_palette = NULL,
  col_side_colors = x[["col_side_colors"]],
  col_side_palette = NULL,
  plot_method = c("ggplot", "plotly"),
  ColSideColors = NULL,
  RowSideColors = NULL,
  heatmap_layers = NULL,
  side_color_layers = NULL,
  dendrogram_layers = NULL,
  branches_lwd = 0.6,
  label_names = c("row", "column", "value"),
  fontsize_row = 10,
  fontsize_col = 10,
  subplot_widths = NULL,
  subplot_heights = NULL,
  colorbar_xanchor = if (row_dend_left) "right" else "left",
  colorbar_yanchor = "bottom",
  colorbar_xpos = if (row_dend_left) -0.1 else 1.1,
  colorbar_ypos = 0,
  colorbar_len = 0.3,
  colorbar_thickness = 30,
  showticklabels = c(TRUE, TRUE),
  dynamicTicks = FALSE,
  node_type = c("scatter", "heatmap"),
  grid_size = 0.1,
  point_size_mat = x[["matrix"]][["point_size_mat"]],
  point_size_name = "Point size",
  label_format_fun = function(...) format(..., digits = 4),
  custom_hovertext = x[["matrix"]][["custom_hovertext"]],
  dend_hoverinfo = TRUE,
  side_color_colorbar_len = 0.3,
  height = NULL,
  width = NULL
)

Arguments

x

can either be a heatmapr object, or a numeric matrix Defaults to TRUE unless x contains any NAs.

...

other parameters passed to heatmapr (currently, various parameters may be ignored.

grid_gap

this is a fast alternative to grid_color. The default is 0, but if a larger value is used (for example, 1), then the resulting heatmap will have a white grid which can help identify different cells. This is implemented using style (with xgap and ygap).

colors, col

a vector of colors to use for heatmap color. The default uses viridis(n=256, alpha = 1, begin = 0, end = 1, option = "viridis") It is passed to scale_fill_gradientn. If colors is a color function (with the first argument being `n` = the number of colors), it will be used to create 256 colors from that function. (col is there to stay compatible with heatmap.2)

limits

a two dimensional numeric vector specifying the data range for the scale.

na.value

color to use for missing values (default is "grey50").

row_text_angle

numeric (Default is 0), the angle of the text of the rows. (this is called srtRow in heatmap.2)

column_text_angle

numeric (Default is 45), the angle of the text of the columns. (this is called srtCol in heatmap.2)

subplot_margin

Currently not well implemented. It is passed to subplot. Default is 0. Either a single value or four values (all between 0 and 1). If four values are provided, the first is used as the left margin, the second is used as the right margin, the third is used as the top margin, and the fourth is used as the bottom margin. If a single value is provided, it will be used as all four margins.

cellnote

Values to be shown as annotations atop the heatmap cells.

draw_cellnote

Should the cellnote annotations be drawn? Defaults is FALSE, if cellnote is not supplied, TRUE if cellnote is supplied. If TRUE and cellnote is not supplied, x will be used for cellnote.

cellnote_color

The color of the cellnote text to be used.

cellnote_textposition

The text positioning/centering of the cellnote. Default is "middle right". Options are "top left", "top center", "top right", "middle left", "middle center", "middle right", "bottom left", "bottom center", "bottom right"

cellnote_size

The font size (HTML/CSS) of the cellnote. Default is 12.

Rowv

determines if and how the row dendrogram should be reordered. By default, it is TRUE, which implies dendrogram is computed and reordered based on row means. If NULL or FALSE, then no dendrogram is computed and no reordering is done. If a dendrogram (or hclust), then it is used "as-is", ie without any reordering. If a vector of integers, then dendrogram is computed and reordered based on the order of the vector.

Colv

determines if and how the column dendrogram should be reordered. Has the options as the Rowv argument above and additionally when x is a square matrix, Colv = "Rowv" means that columns should be treated identically to the rows.

distfun

function used to compute the distance (dissimilarity) between both rows and columns. Defaults to dist. The options "pearson", "spearman" and "kendall" can be used to use correlation-based clustering, which uses as.dist(1 - cor(t(x))) as the distance metric (using the specified correlation method).

hclustfun

function used to compute the hierarchical clustering when Rowv or Colv are not dendrograms. Defaults to hclust.

dist_method

default is NULL (which results in "euclidean" to be used). Can accept alternative character strings indicating the method to be passed to distfun. By default distfun. is dist hence this can be one of "euclidean", "maximum", "manhattan", "canberra", "binary" or "minkowski".

hclust_method

default is NULL (which results in "complete" to be used). Can accept alternative character strings indicating the method to be passed to hclustfun By default hclustfun is hclust hence this can be one of "ward.D", "ward.D2", "single", "complete", "average" (= UPGMA), "mcquitty" (= WPGMA), "median" (= WPGMC) or "centroid" (= UPGMC). Specifying hclust_method=NA causes heatmaply to use find_dend to find the "optimal" dendrogram for the data.

distfun_row

distfun for row dendrogram only.

hclustfun_row

hclustfun for col dendrogram only.

distfun_col

distfun for row dendrogram only.

hclustfun_col

hclustfun for col dendrogram only.

dendrogram

character string indicating whether to compute 'none', 'row', 'column' or 'both' dendrograms. Defaults to 'both'. However, if Rowv (or Colv) is FALSE or NULL and dendrogram is 'both', then a warning is issued and Rowv (or Colv) arguments are honoured. It also accepts TRUE/FALSE as synonyms for "both"/"none".

show_dendrogram

Logical vector of length two, controlling whether the row and/or column dendrograms are displayed. If a logical scalar is provided, it is repeated to become a logical vector of length two.

reorderfun

function(d, w) of dendrogram and weights for reordering the row and column dendrograms. The default uses statsreorder.dendrogram

k_row

an integer scalar with the desired number of groups by which to color the dendrogram's branches in the rows (uses color_branches) If NA then find_k is used to deduce the optimal number of clusters.

k_col

an integer scalar with the desired number of groups by which to color the dendrogram's branches in the columns (uses color_branches) If NA then find_k is used to deduce the optimal number of clusters.

symm

logical indicating if x should be treated symmetrically; can only be true when x is a square matrix.

revC

logical indicating if the column order should be reversed for plotting. Default (when missing) - is FALSE, unless symm is TRUE. This is useful for cor matrix.

scale

character indicating if the values should be centered and scaled in either the row direction or the column direction, or none. The default is "none".

na.rm

logical (default is TRUE) indicating whether NA's should be removed when scaling (i.e.: when using rowMeans/colMeans). Generally it should always be kept as TRUE, and is included here mainly to stay backward compatible with gplots::heatmap.2. This argument does not effect the presence of NA values in the matrix itself. For removing rows/columns with NAs you should pre-process your matrix using na.omit (or some form of imputation).

row_dend_left

logical (default is FALSE). Should the row dendrogram be plotted on the left side of the heatmap. If false then it will be plotted on the right side.

margins

numeric vector of length 4 (default is c(50,50,NA,0)) containing the margins (see layout) for column, row and main title names, respectively. The top margin is NA by default. If main=="" then the top margin will be set to 0, otherwise it will get 30. For a multiline title a larger default for the 3rd element should be set. The right margin is NA by default, meaning it will be zero if row_dend_left is FALSE, or 100 if row_dend_left is TRUE.

scale_fill_gradient_fun

A function that creates a smooth gradient for the heatmap. The default uses scale_fill_gradientn with the values of colors, limits, and na.value that are supplied by the user. The user can input a customized function, such as scale_color_gradient() in order to get other results (although the virids default is quite recommended)

grid_color

control the color of the heatmap grid. Default is NA. Value passed to geom_tile. Do not use this parameter on larger matrix sizes, as it can dramatically prolong the build time of the heatmap. (another parameter, grid_color, will be added in the future - once it is implemented in plotly) In the meantime it is MUCH better to use the grid_gap argument.

srtRow

if supplied, this overrides row_text_angle (this is to stay compatible with heatmap.2)

srtCol

if supplied, this overrides column_text_angle (this is to stay compatible with heatmap.2)

xlab

A character title for the x axis.

ylab

A character title for the y axis.

main

A character title for the heatmap.

titleX

logical (TRUE). should x-axis titles be retained? (passed to subplot).

titleY

logical (TRUE). should y-axis titles be retained? (passed to subplot).

hide_colorbar

logical (FALSE). If TRUE, then the color bar (i.e.: the legend) is hidden.

key.title

(character) main title of the color key. If set to NULL (default) no title will be plotted.

return_ppxpy

logical (FALSE). If TRUE, then no plotting is done and the p, px and py objects are returned (before turning into plotly objects). This is a temporary option which might be removed in the future just to make it easy to create a ggplot heatmaps.

row_side_colors, col_side_colors

data.frame of factors to produce row/column side colors in the style of heatmap.2/heatmap.3. When a data.frame is provided, the column names are used as the label names for each of the newly added row_side_colors. When a vector is provided it is coerced into a data.frame and the name of the side color will be just row_side_colors.

row_side_palette, col_side_palette

Color palette functions to be used for row_side_colors and col_side_colors respectively.

ColSideColors, RowSideColors

passed to row_side_colors,col_side_colors in order to keep compatibility with heatmap.2

seriate

character indicating the method of matrix sorting (default: "OLO"). Implemented options include: "OLO" (Optimal leaf ordering, optimizes the Hamiltonian path length that is restricted by the dendrogram structure - works in O(n^4) ) "mean" (sorts the matrix based on the reorderfun using marginal means of the matrix. This is the default used by heatmap.2), "none" (the default order produced by the dendrogram), "GW" (Gruvaeus and Wainer heuristic to optimize the Hamiltonian path length that is restricted by the dendrogram structure)

heatmap_layers

ggplot object(s) (eg, list(theme_bw())) to be added to the heatmap before conversion to a plotly object.

side_color_layers

ggplot2 objects to be added to side color plots, similar to heatmap_layers.

dendrogram_layers

ggplot2 objects to be added to dendrograms, similar to heatmap_layers and side_color_layers.

branches_lwd

numeric (default is 0.6). The width of the dendrograms' branches. If NULL then it is ignored. If the "lwd" is already defined in Rowv/Colv then this parameter is ignored (it is checked using has_edgePar("lwd")).

file

name of the file(s) into which to save the heatmaply output. Should be a character vector of strings ending with ".html" for a dynamic output, or ".png", ".jpeg", ".pdf" for a static output.

For example: heatmaply(x, file = "heatmaply_plot.html") or dir.create("folder"); heatmaply(x, file = "folder/heatmaply_plot.html")

This is based on saveWidget, and webshot for the static files. For more refined control over the static file output, you should save the heatmaply object using export and pass the arguments you want based on the ones in webshot.

Another example: heatmaply(x, file = c("heatmaply_plot.html", "heatmaply_plot.png"))

width, height

The width and height of the output htmlwidget, or the output file if exporting to png/pdf/etc. Presumed to be in pixels, but if a plotly internal function decides it's in other units you may end up with a huge file! Default is 800x500 when exporting to a file, and 100 as a htmlwidget.

long_data

Data in long format. Replaces x, so both should not be used. Colnames must be c("name", "variable", "value"). If you do not have a names column you can simply use a sequence of numbers from 1 to the number of "rows" in the data.

plot_method

Use "ggplot" or "plotly" to choose which library produces heatmap and dendrogram plots

label_names

Names for labels of x, y and value/fill mouseover.

fontsize_row, fontsize_col, cexRow, cexCol

Font size for row and column labels.

subplot_widths, subplot_heights

The relative widths and heights of each subplot. The length of these vectors will vary depending on the number of plots involved.

colorbar_len

The length of the colorbar/color key relative to the total plot height. Only used if plot_method = "plotly"

colorbar_thickness

The thickness (width) of the colorbar/color key in pixels. Only used if plot_method = "plotly".

colorbar_xanchor, colorbar_yanchor

The x and y anchoring points of the colorbar/color legend. Can be "left", "middle" or "right" for colorbar_xanchor, and "top", "middle" or "bottom" for colorbar_yanchor. See colorbar for more details.

colorbar_xpos, colorbar_ypos

The x and y co-ordinates (in proportion of the plot window) of the colorbar/color legend. See colorbar for more details.

showticklabels

A logical vector of length two (default is TRUE). If FALSE, then the ticks are removed from the sides of the plot. The first location refers to the x axis and the second to the y axis. If only one value is supplied (TRUE/FALSE) then it is replicated to get to length 2. When using this parameter, it might be worth also adjusting margins. This option should be used when working with medium to large matrix size as it makes the heatmap much faster (and the hover still works).

dynamicTicks

(default: FALSE). passed to ggplotly: should plotly.js dynamically generate axis tick labels? Dynamic ticks are useful for updating ticks in response to zoom/pan interactions; however, they can not always reproduce labels as they would appear in the static ggplot2 image.

grid_size

When node_type is "scatter", this controls point size. When node_type is "heatmap", this controls the size of the grid between heatmap cells.

node_type

For plot_method = "ggplot", should the heatmap be rendered as a x-y scatter plot (node_type = "scatter") or a heatmap (node_type = "heatmap"). Default is node_type = "heatmap".

point_size_mat

Matrix to map to point size

point_size_name

Name of point size mapping (for hovertext/legend)

label_format_fun

Function to format hovertext (eg, function(...) round(..., digits=3) or function(...) format(..., digits=3)

labRow, labCol

character vectors with row and column labels to use; these default to rownames(x) or colnames(x), respectively. if set to NA, they change the value in showticklabels to be FALSE. This is mainly to keep backward compatibility with gplots::heatmap.2.

custom_hovertext

Custom hovertext matrix (the same dimensions as the input). If plot_method is "plotly" then just this text is displayed; if plot_method if "ggplot" then it is appended to the existing text.

dend_hoverinfo

Boolean value which controls whether mouseover text is shown for the row and column dendrograms.

side_color_colorbar_len

As with colorbar_len, this controls the length of the colorbar/color key relative to the total plot height. This argument controls the colorbar_len of the side colour plots. Only used if plot_method = "plotly".

Examples

if (FALSE) {

# mtcars
# x <- heatmapr(mtcars)
library(heatmaply)
heatmaply(iris[, -5], k_row = 3, k_col = 2)
heatmaply(cor(iris[, -5]))
heatmaply(cor(iris[, -5]), limits = c(-1, 1))
heatmaply(mtcars, k_row = 3, k_col = 2)
# heatmaply(mtcars, k_row = 3, k_col = 2, grid_color = "white")
heatmaply(mtcars, k_row = 3, k_col = 2, grid_gap = 1)

# make sure there is enough room for the labels:
heatmaply(mtcars, margins = c(40, 130))
# this is the same as using:
heatmaply(mtcars) %>% layout(margin = list(l = 130, b = 40))

# control text angle
heatmaply(mtcars, column_text_angle = 90, margins = c(40, 130))
# the same as using srtCol:
# heatmaply(mtcars, srtCol = 90) %>% layout(margin = list(l = 130, b = 40))



x <- mtcars
# different colors
heatmaply(x, colors = heat.colors(200))
# using special scale_fill_gradient_fun colors
heatmaply(x, scale_fill_gradient_fun = scale_color_gradient())


# We can join two heatmaps together:
library(heatmaply)
hm1 <- heatmaply(mtcars, margins = c(40, 130))
hm2 <- heatmaply(mtcars, scale = "col", margins = c(40, 130))
subplot(hm1, hm2, margin = .2)

# If we want to share the Y axis, then it is risky to keep any of the dendrograms:
library(heatmaply)
hm1 <- heatmaply(mtcars, Colv = FALSE, Rowv = FALSE, margins = c(40, 130))
hm2 <- heatmaply(mtcars,
  scale = "col", Colv = FALSE, Rowv = FALSE,
  margins = c(40, 130)
)
subplot(hm1, hm2, margin = .02, shareY = TRUE)


# We can save heatmaply as an HTML file by using:
heatmaply(iris[, -5], file = "heatmaply_iris.html")
# or a png/pdf/jpeg file using:
heatmaply(iris[, -5], file = "heatmaply_iris.png")
# or just doing it in one go:
heatmaply(iris[, -5], file = c("heatmaply_iris.html", "heatmaply_iris.png"))



# If we don't want the HTML to be selfcontained, we can use the following:
library(heatmaply)
library(htmlwidgets)
heatmaply(iris[, -5]) %>%
  saveWidget(file = "heatmaply_iris.html", selfcontained = FALSE)


# Example for using RowSideColors

x <- as.matrix(datasets::mtcars)
rc <- colorspace::rainbow_hcl(nrow(x))

library(gplots)
library(viridis)
heatmap.2(x,
  trace = "none", col = viridis(100),
  RowSideColors = rc
)

heatmaply(x,
  seriate = "mean",
  RowSideColors = rc
)


heatmaply(x[, -c(8, 9)],
  seriate = "mean",
  col_side_colors = c(rep(0, 5), rep(1, 4)),
  row_side_colors = x[, 8:9]
)
heatmaply(x[, -c(8, 9)],
  seriate = "mean",
  col_side_colors = data.frame(a = c(rep(0, 5), rep(1, 4))),
  row_side_colors = x[, 8:9]
)


## Example of using Rowv And Colv for custumized dendrograms.


x <- as.matrix(datasets::mtcars)

# now let's spice up the dendrograms a bit:
library(dendextend)

row_dend <- x %>%
  dist() %>%
  hclust() %>%
  as.dendrogram() %>%
  set("branches_k_color", k = 3) %>%
  set("branches_lwd", 4) %>%
  ladderize()
#    rotate_DendSer(ser_weight = dist(x))
col_dend <- x %>%
  t() %>%
  dist() %>%
  hclust() %>%
  as.dendrogram() %>%
  set("branches_k_color", k = 2) %>%
  set("branches_lwd", 4) %>%
  ladderize()
#    rotate_DendSer(ser_weight = dist(t(x)))

heatmaply(x, Rowv = row_dend, Colv = col_dend)


heatmaply(is.na10(airquality))
heatmaply(is.na10(airquality), grid_gap = 1)

# grid_gap can handle quite large data matrix
heatmaply(matrix(1:10000, 100, 100), k_row = 3, k_col = 3, grid_gap = 1)

# Examples of playing with font size:
heatmaply(mtcars, fontsize_col = 20, fontsize_row = 5, margin = c(100, 90))



# Example for using subplot_width/subplot_height

heatmaply(percentize(mtcars),
  subplot_widths = c(0.6, 0.4),
  subplot_heights = c(0.05, 0.95)
)



# Example of removing labels and thus making the plot faster
heatmaply(iris, showticklabels = c(T, F), margins = c(80, 10))

# this is what allows for a much larger matrix to be printed:
set.seed(2017 - 05 - 18)
large_x <- matrix(rnorm(19), 1000, 100)
heatmaply(large_x, dendrogram = F, showticklabels = F, margins = c(1, 1))
}
if (FALSE) {
heatmaply_na(airquality)
}
if (FALSE) {
heatmaply_cor(cor(mtcars))
}