Plot cell-type deconvolution heatmap

Usage

get_celltype_heatmap(
  x,
  group_column = NULL,
  samples = NULL,
  cell_types = NULL,
  top_n = 20,
  transform = c("none", "zscore", "log1p"),
  cluster_rows = TRUE,
  cluster_columns = TRUE,
  show_values = FALSE,
  font_size = 11,
  return_type = c("plot", "matrix", "both")
)

Arguments

x: A VISTA object.
group_column: Optional grouping column from sample_info(x). If provided and cluster_columns = FALSE, samples are ordered by this group.
samples: Optional character vector of samples to include.
cell_types: Optional character vector of cell types to include.
top_n: Number of top cell types by mean score when cell_types is NULL.
transform: One of "none", "zscore", or "log1p".
cluster_rows: Logical; hierarchical cluster cell types.
cluster_columns: Logical; hierarchical cluster samples.
show_values: Logical; overlay numeric values on tiles.
font_size: Base font size.
return_type: One of "plot", "matrix", or "both".

Value

A ggplot object, matrix, or list depending on return_type.

Examples

mat <- matrix(rpois(20, lambda = 20), nrow = 5)
rownames(mat) <- paste0("gene", seq_len(5))
colnames(mat) <- paste0("sample", seq_len(4))
se <- SummarizedExperiment::SummarizedExperiment(
  assays = list(norm_counts = mat),
  colData = S4Vectors::DataFrame(
    cond = c("A", "A", "B", "B"),
    row.names = colnames(mat)
  ),
  rowData = S4Vectors::DataFrame(
    gene_id = rownames(mat),
    row.names = rownames(mat)
  )
)
v <- as_vista(se, group_column = "cond")
md <- S4Vectors::metadata(v)
md$cell_fractions <- data.frame(
  fibroblast = c(0.2, 0.3, 0.4, 0.5),
  epithelial = c(0.8, 0.7, 0.6, 0.5),
  row.names = colnames(mat)
)
S4Vectors::metadata(v) <- md
get_celltype_heatmap(v, group_column = "cond")