Skip to contents

heat_map_check_colours() simulates how the colours of a heatmap look with the three common colour-vision deficiencies - deuteranopia and protanopia (red-green, about 1 in 12 men of northern European descent) and tritanopia (blue-yellow) - and reports the colours that become hard to tell apart: two levels of an annotation, of the values of non-numeric cells, of stacked bars or of branch colours, and values far apart on a colour scale that end up looking alike. The deficiencies are simulated as Machado, Oliveira and Fernandes (2009) model them (at full severity, in linear RGB, as colorspace::simulate_cvd() does), and colours count as hard to tell apart when their difference in CIELAB (Delta E) falls below min_difference.

Usage

heat_map_check_colours(
  x,
  deficiency = c("deuteranopia", "protanopia", "tritanopia"),
  min_difference = 10
)

Arguments

x

a heatmap (an object of class heat_map), or colours (a vector, named by what they stand for).

deficiency

the colour-vision deficiencies to simulate, all three by default.

min_difference

the smallest difference (CIELAB Delta E) of two colours that count as easy to tell apart, 10 by default. Colours that differ by less than this in normal vision too are reported as such.

Value

a data.frame of class heat_map_colour_check with one row per pair of colours that are hard to tell apart: element (e.g. "annotation cell_type" or "colour scale"), vision ("normal" or the deficiency), the two levels (first and second; positions along the scale for colour scales), their colours and their difference. It prints as a summary.

Details

Qualitative palettes of more than four or five colours rarely stay apart for every deficiency: colour-blind readers see colours along fewer dimensions. Name the levels in the legend and beside the tracks, or choose colours that differ in lightness as well as hue, e.g. the first eight colours of palette.colors(palette = "Okabe-Ito") (Okabe and Ito 2008; R 4.0.0 or later), given to annotation_colours or cell_col_palette. Colour scales such as "viridis" or "Blue-Red 3" change in lightness from end to end, so they stay readable.

With options(HeatmapR.check_colours = TRUE), heat_map() checks every heatmap it draws and gives a message when colours are hard to tell apart.

References

Machado GM, Oliveira MM and Fernandes LAF (2009). A physiologically-based model for simulation of color vision deficiency. IEEE Transactions on Visualization and Computer Graphics 15(6), 1291-1298. doi:10.1109/TVCG.2009.113

Okabe M and Ito K (2008). Color universal design (CUD): how to make figures and presentations that are friendly to colorblind people. https://jfly.uni-koeln.de/color/

See also

heat_map_describe for alternative text

Author

Dillon Hammill (dillon.hammill21@gmail.com)

Examples

# The red and green of matplotlib's palette: apart for most readers, close
# for red-green colour-blind ones
heat_map_check_colours(c(case = "#D62728", control = "#2CA02C", other = "#1F77B4"))
#> 1 pair of colours is hard to tell apart (colours), e.g. case and control of colours with deuteranopia. The first eight colours of palette.colors(palette = "Okabe-Ito") stay apart for colour-blind readers.
#> 
#> colours:
#>   case and control (deuteranopia, Delta E 7.3)

hm <- heat_map(
  mtcars,
  scale = "column",
  annotation_y = data.frame(
    cylinders = factor(mtcars$cyl),
    row.names = rownames(mtcars)
  ),
  annotation_colours = list(cylinders = c("#D62728", "#2CA02C", "#1F77B4")),
  popup = FALSE
)
#> Applying range scaling to each column...

heat_map_check_colours(hm)
#> 1 pair of colours is hard to tell apart (annotation cylinders), e.g. 4 and 6 of annotation cylinders with deuteranopia. The first eight colours of palette.colors(palette = "Okabe-Ito") stay apart for colour-blind readers.
#> 
#> annotation cylinders:
#>   4 and 6 (deuteranopia, Delta E 7.3)