Interactive Heatmaps for Delta-Scale Pairwise Correlations

delta_heatmap() is designed to take an \(n \times p \times t\) array and display a heatmap to visualize the observed delta-scalecorrelation matrix. This delta-scale correlation matrix is used in the prolong model as the adjacency matrix for the graph whose laplacian is used in the (group) lasso + laplacian penalty used.

Usage

delta_heatmap(
  x,
  timediff = "2-1",
  object = NULL,
  selected = NULL,
  interactive = TRUE,
  grayscale = FALSE
)

Arguments

x: Input covariate array, with n rows, p columns, and t slices
timediff: Pair of time points to use in heatmap. Should be in format 't2-t1'
object: A prolong model object. Can be left NULL if selected is provided or if all variables are to be used
selected: A character list of variable names or a numeric index of variables of interest to be included in the heatmap Can be left NULL if object is provided or if all variables are to be used
interactive: If TRUE, a shiny app will open in browser that will display an interactive version of the heatmap where sub-heatmaps can be selected for display
grayscale: If TRUE the viridis colors will be desaturated to grayscale

Value

A heatmap, possibly interactive in a shiny app

References

Gu Z (2022). “Complex Heatmap Visualization.” iMeta. doi:10.1002/imt2.43 . Gu Z, Huebschmann D (2021). “Make Interactive Complex Heatmaps in R.” Bioinformatics. doi:10.1093/bioinformatics/btab806 .

Examples

if (FALSE) {
delta_heatmap(Xarray)
delta_heatmap(Xarray, timediff = "4-2", interactive = F, grayscale = T)
}