Skip to contents

Heatmap with sensible defaults for RNA-seq expression data

Source: R/quickmap.R
quickmap.Rd

Generate a heatmap using pheatmap with sensible defaults for RNA-seq. quickmap() will also attempt to perform vectorized scaling, clustering, and distance calculations with functions from the Rfast, fastcluster and rdist packages in order to speed up calculations for large gene expression matrices.

Usage

quickmap(
  mat,
  diverging_palette = TRUE,
  fix_extreme = FALSE,
  thresh = 1.96,
  removeVar = NULL,
  ...
)

Arguments

mat

numeric matrix to be passed onto pheatmap function

diverging_palette

logical. Default(TRUE). Sets the color scale to a diverging palette (blue -> white -> red). If FALSE, set the color to a continuous color palette viridis::magma(), useful for un-scaled expression data.

fix_extreme

logical. Should the extreme values at the ends of a diverging palette be fixed colors? Default FALSE.

thresh

If fix_extreme=TRUE then at what value should the color scale be fixed? Since scaling is implied by the use of this argument, the default value is set to 1.96 (positive and negative).

removeVar

If not NULL, remove this proportion of features based on the variance across rows. Default NULL. Note, if removeVar is used the lowest variance features are removed prior to scaling the input data

...

args to be passed to pheatmap() function

Value

pheatmap object. See ?pheatmap::pheatmap() for details

Details

The default arguments to pheatmap::pheatmap() are:

    • scale = "row"

    • show_rownames = FALSE

    • border_col = NA

    • cluster_rows = TRUE

    • cluster_cols = TRUE

    • color = colorRampPalette(c("dodgerblue3", "grey99", "firebrick3"))(30) for diverging_palette = TRUE

    • color = rev(viridisLite::magma(n = 30)) for diverging_palette = FALSE

    • treeheight_row = 0

    • clustering_distance_rows = "correlation"

    • clustering_distance_cols = "euclidean"

    • clustering_method = "ward.D2"

    • angle_col = 315

    You can pass in additional arguments or simply override the defaults as well.

Examples

# display heatmap of scaled count data and add title to the plot
quickmap(GSE161650_lc, main = "THZ1 vs DMSO")


# Remove 90% lowest variance features and fix color scale
quickmap(
  GSE161650_lc,
  removeVar = 0.9,
  main = "THZ1 vs DMSO",
  fix_extreme = TRUE,
  thresh = 1
)
#> Removing 90% lowest variance features...