Heatmap

Create a highly customizable heatmap using the ComplexHeatmap package.

Usage

splendidHeatmap(
  this_matrix,
  importance_values,
  these_samples_metadata,
  max_number_of_features_per_group = 10,
  splitColumnName = "pathology",
  metadataColumns = c("pathology"),
  numericMetadataColumns = NULL,
  numericMetadataMax = NULL,
  prioritize_ordering_on_numeric = TRUE,
  custom_colours = NULL,
  legend_direction = "horizontal",
  legend_position = "bottom",
  legend_row = 3,
  legend_col = 3,
  fontSizeGene = 6,
  metadataBarHeight = 1.5,
  leftStackedWidth = 4,
  metadataBarFontsize = 5,
  groupNames = NULL
)

Arguments

this_matrix

A data frame with column Tumor_Sample_Barcode and a column for each feature. Can be binary. Expected to not contain negative values.

importance_values

Provide a data frame of feature (in rows) by group (in columns) with numeric values representative of feature importance. Can be obtained from rf$inportance or basis(NMF).

these_samples_metadata

Data frame containing metadata for your samples.

max_number_of_features_per_group

Optional argument to indicate how many features from each group to be considered for display. Default is 10.

splitColumnName

Optional argument to indicate which metadata column to split on. Default is set to pathology.

metadataColumns

A vector containing the categorical column names you want to plot below.

numericMetadataColumns

A vector containing the numeric columns you want to plot below.

numericMetadataMax

A numeric vector of cutoffs to apply to numeric columns above.

prioritize_ordering_on_numeric

Logical argument specifying whether to sort on numeric metadata first or other metadata columns. Default is TRUE (sort on numeric metadata, then on other columns).

custom_colours

Provide named vector (or named list of vectors) containing custom annotation colours if you do not want to use standartized pallette.

legend_direction

Optional argument to indicate whether legend should be in horizontal (default) or vertical position.

legend_position

Optional argument to indicate where the legend should be drawn. The default is set to bottom, but can also accept top, right, and left.

legend_row

Fiddle with these to widen or narrow your legend (default 3).

legend_col

Fiddle with these to widen or narrow your legend (default 3).

fontSizeGene

Font size for gene labels (default 6).

metadataBarHeight

Optional argument to adjust the height of bar with annotations. The default is 1.5

leftStackedWidth

Optional argument to control how wide should the stacked plot on the left be. The default is 4.

metadataBarFontsize

Optional argument to control for the font size of metadata annotations. The default is 5.

groupNames

optional vector of group names to be displayed above heatmap. Should be the same length as the number of groups that will be shown. Default is NULL (no labels).

Value

Nothing

Details

Make an heatmap that is looking cute using ComplexHeatmap. The metadata is expected to follow the structure and column naming used in GAMBL. If you provide your own non-GAMBL samples and metadata, you must include at least the columns with names corresponding to annotation tracks and column "Tumor_Sample_Barcode". showing sample ids. The metadata can contain numeric columns, which will be plotted as numeric variables in the annotation. The feature matrix is supplied in `this_matrix` argument. and is expected to have samples in rows, and features in columns. The argument importance_values is similar to the widths of NMF object or importance values for feature/group from RF models. It is also expected to have column names (having names of the groups that will be shown on heatmap) and rownames (corresponding to feature ids).

Examples

if (FALSE) { # \dontrun{
splendidHeatmap(this_matrix = data,
                importance_values = rf$importance[,c(1:3)],
                these_samples_metadata = MASTER.METADATA,
                splitColumnName = "pathology",
                metadataColumns = c("cohort",
                                    "pathology",
                                    "sex",
                                    ".",
                                    "COO_consensus",
                                    "DHITsig_consensus",
                                    "seq_type"),
                numericMetadataColumns = ".",
                numericMetadataMax = 0.7,
                custom_colours = custom_colours)
} # }