ComplexHeatmapPlot-class: The ComplexHeatmapPlot panel
In iSEE: Interactive SummarizedExperiment Explorer

Description ComplexHeatmapPlot slot overview Constructor Supported methods Author(s) See Also Examples

The ComplexHeatmapPlot is a panel class for creating a Panel that displays an assay of a SummarizedExperiment object as a Heatmap with features as rows and samples and columns, respectively. It provides slots and methods for specifying the features of interest, which assay to display in the main heatmap, any transformations to perform on the data, and which metadata variables to display as row and column heatmap annotations.

The following slots control the rows that are used:

CustomRows, a logical scalar indicating whether the custom list of rows should be used. If FALSE, the incoming selection is used instead. Defaults to TRUE.
CustomRowsText, string containing newline-separated row names. This specifies which rows of the SummarizedExperiment object are to be shown in the heatmap. If NA, defaults to the first row name of the SummarizedExperiment.

The following slots control the metadata variables that are used:

ColumnData, a character vector specifying columns of the colData to show as columnAnnotation. Defaults to character(0).
RowData, a character vector specifying columns of the rowData to show as columnAnnotation. Defaults to character(0).

The following slots control the choice of assay values:

Assay, string specifying the name of the assay to use for obtaining expression values. Defaults to the first valid assay name (see ?".refineParameters,ComplexHeatmapPlot-method" for details).

The following slots control the clustering of rows:

ClusterRows, a logical scalar indicating whether rows should be clustered by their assay values. Defaults to FALSE.
ClusterRowsDistance, string specifying a distance measure to use. This can be any one of "euclidean", "maximum", "manhattan", "canberra", "binary", "minkowski", "pearson", "spearman", or "kendall". Defaults to "spearman".
ClusterRowsMethod, string specifying a linkage method to use. This can be any one of "ward.D", "ward.D2", "single", "complete", "average", "mcquitty", "median", or "centroid". Defaults to "ward.D2".

The following control transformations applied to rows:

AssayCenterRows is a logical scalar indicating whether assay values should be centered for each row.
AssayScaleRows is a logical scalar indicating whether assay values should be scaled for each row. This transformation is only applicable if AssayCenterRows is TRUE.

The following slots control the color scale:

CustomBounds is logical scale indicating whether the color scale should be constrained by an upper and/or a lower bounds.
LowerBound is a numerical value setting the lower bound of the color scale; or NA to disable the lower bound when CustomBounds is TRUE.
UpperBound is a numerical value setting the lower bound of the color scale; or NA to disable the upper bound when CustomBounds is TRUE.
DivergentColormap is a character scalar indicating a 3-color divergent colormap to use when AssayCenterRows is TRUE.

The following slots refer to general plotting parameters:

ShowDimNames, a character vector specifying the dimensions for which to display names. This can contain zero or more of "Rows" and "Columns". Defaults to "Rows".
LegendPosition, string specifying the position of the legend on the plot. Defaults to "Bottom" but can also be "Right".
LegendDirection, string specifying the orientation of the legend on the plot for continuous covariates. Defaults to "Horizontal" but can also be "Vertical".

The following slots control the effect of the transmitted selection from another panel:

SelectionEffect, a string specifying the selection effect. This should be one of "Color" (the default), where all selected points change to the specified color; "Restrict", where all non-selected points are not plotted.
SelectionColor, a string specifying the color to use for selected points when SelectionEffect="Color". Defaults to "red".

The following slots control some aspects of the user interface:

DataBoxOpen, a logical scalar indicating whether the data parameter box should be open. Defaults to FALSE.
VisualBoxOpen, a logical scalar indicating whether the visual parameter box should be open. Defaults to FALSE.

In addition, this class inherits all slots from its parent Panel class.

ComplexHeatmapPlot(...) creates an instance of a ComplexHeatmapPlot class, where any slot and its value can be passed to ... as a named argument.

In the following code snippets, x is an instance of a ComplexHeatmapPlot class. Refer to the documentation for each method for more details on the remaining arguments.

For setting up data values:

.cacheCommonInfo(x) adds a "ComplexHeatmapPlot" entry containing valid.assay.names, a character vector of valid assay names; discrete.assay.names, a character vector of valid assay names with discrete atomic values; continuous.assay.names, a character vector of valid assay names with continuous atomic values; valid.colData.names, a character vector of names of columns in colData that are valid; discrete.colData.names, a character vector of names for columns in colData with discrete atomic values; continuous.colData.names, a character vector of names of columns in colData with continuous atomic values; valid.rowData.names, a character vector of names of columns in rowData that are valid; discrete.rowData.names, a character vector of names for columns in rowData with discrete atomic values; continuous.rowData.names, a character vector of names of columns in rowData with continuous atomic values. Valid assay names are defined as those that are non-empty, i.e., not ""; valid columns in colData and rowData are defined as those that contain atomic values. This will also call the equivalent Panel method.
.refineParameters(x, se) replaces any NA value in "Assay" with the first valid assay name; and NA value in "CustomRowsText" with the first row name. This will also call the equivalent Panel method for further refinements to x. If no valid column metadata fields are available, NULL is returned instead.

For defining the interface:

.defineInterface(x, se, select_info) defines the user interface for manipulating all slots described above and in the parent classes. TODO It will also create a data parameter box that can respond to specialized .defineDataInterface, and a visual parameter box and a selection parameter box both specific to the ComplexHeatmapPlot panel. This will override the Panel method.
.defineDataInterface(x, se, select_info) returns a list of interface elements for manipulating all slots described above.
.defineSelectionEffectInterface(x) returns a list of interface elements for controlling the multiple selection effect.
.defineOutput(x) returns a UI element for a brushable plot.
.panelColor(x) will return the specified default color for this panel class.
.hideInterface(x, field) returns a logical scalar indicating whether the interface element corresponding to field should be hidden. This returns TRUE for the selection history ("SelectionHistory"), otherwise it dispatches to the Panel method.

For generating the output:

.generateOutput(x, se, all_memory, all_contents) returns a list containing contents, a data.frame with one row per point currently present in the plot; plot, a ggplot object; commands, a list of character vector containing the R commands required to generate contents and plot; and varname, a string containing the name of the variable in commands that was used to obtain contents.
.exportOutput(x, se, all_memory, all_contents) will create a PDF file containing the current plot, and return a string containing the path to that PDF. This assumes that the plot field returned by .generateOutput is a Heatmap object.

For monitoring reactive expressions:

.createObservers(x, se, input, session, pObjects, rObjects) sets up observers for all slots described above and in the parent classes. This will also call the equivalent Panel method.
.renderOutput(x, se, output, pObjects, rObjects) will add a rendered plot element to output. The reactive expression will add the contents of the plot to pObjects$contents and the relevant commands to pObjects$commands. This will also call the equivalent Panel method to render the panel information text boxes.

For defining the panel name:

.fullName(x) will return "Complex heatmap".

For documentation:

.definePanelTour(x) returns an data.frame containing a panel-specific tour.

Kevin Rue-Albrecht

Panel, for the immediate parent class.

#################
# For end-users #
#################

x <- ComplexHeatmapPlot()
x[["ShowDimNames"]]
x[["ShowDimNames"]] <- c("Rows", "Columns")

##################
# For developers #
##################

library(scater)
sce <- mockSCE()
sce <- logNormCounts(sce)

old_cd <- colData(sce)
colData(sce) <- NULL

# Spits out a NULL and a warning if there is nothing to plot.
sce0 <- .cacheCommonInfo(x, sce)
.refineParameters(x, sce0)

# Replaces the default with something sensible.
colData(sce) <- old_cd
sce0 <- .cacheCommonInfo(x, sce)
.refineParameters(x, sce0)