Make an interactive heatmap with heatmaply — interactive

This is a generic function which may be useful outside of this package. It produces a heatmap based on an expression matrix and accompanying experiment data in the form of a frame, using heatmaply::heatmaply().

Usage

interactive_heatmap(
  plotmatrix,
  displaymatrix,
  sample_annotation,
  cluster_rows = TRUE,
  cluster_cols = FALSE,
  scale = "row",
  row_labels,
  colors = viridisLite::viridis(100),
  cexCol = 0.9,
  cexRow = 0.9,
  display_numbers = FALSE,
  hide_colorbar = FALSE,
  grid_gap = 1,
  cor_method = "spearman",
  cluster_method = "ward.D2",
  plot_height = 600,
  ...
)

Arguments

plotmatrix: Expression/ other data matrix
displaymatrix: A matrix of values shown on hover, in addition to the (possibly scaled/ transformed) values in plotmatrix
sample_annotation: A data frame with sample metadata, used to draw column side colors and an accompanying legend
cluster_rows: Cluster rows?
cluster_cols: Cluster columns?
scale: 'row', 'column' or none
row_labels: Vector labels to use for rows
colors: A vector of colors for the heatmap
cexCol: Character expansion factor passed to heatmaply()
cexRow: Character expansion factor passed to heatmaply()
display_numbers: Boolean, should the (possibly scaled/ transformed) values in plotmatrix be displayed on the heatmap cells?
hide_colorbar: Boolean, should the color scale legend be hidden?
grid_gap: Pixel gap drawn between cells, passed through to heatmaply(). The default of 1 reads fine for small heatmaps, but on a heatmap with hundreds of rows the gaps can visually dominate and obscure cross-column colour patterns; pass 0 for contiguous cells in that case.
cor_method: Correlation method passed to calculate_dendrogram when clustering rows and/or columns (default: 'spearman'). Match this to the method used by another clustering of the same data (e.g. a ComplexHeatmap::Heatmap() call's clustering_distance_columns) to reproduce its dendrogram topology here.
cluster_method: Clustering method passed to calculate_dendrogram (default: 'ward.D2'). Match this to another clustering's clustering_method_columns/-_rows for the same reason as cor_method.
plot_height: The total rendered height of the plot in pixels. Passed through to heatmaply() as its height argument, and also used to convert the fixed-pixel annotation row height into the fraction heatmaply() expects. When displayed inside a plotlyOutput(), the latter's own height argument should match this value so the container and the widget agree.
...: Additional arguments passed to heatmaply(). For example, heatmaply() reorders dendrogram leaves for visual clarity (seriate = "OLO") by default; pass seriate = "none" here to keep the plain hclust leaf order instead, matching ComplexHeatmap::Heatmap()'s default display.

Value

output A plotly htmlwidget as produced by heatmaply()

Examples

mat <- matrix(rnorm(24), nrow = 6, ncol = 4,
  dimnames = list(paste0("gene", 1:6), paste0("s", 1:4)))
sample_annotation <- data.frame(
  condition = rep(c("treated", "control"), each = 2),
  row.names = colnames(mat)
)
interactive_heatmap(mat, mat, sample_annotation, row_labels = rownames(mat))