Compare and Contrast: Looking at figures in a new light
Is it possible to put yourself in another person’s shoes and see the world from their point-of-view? When you are trying to communicate the results of your research, it is necessary to consider how others see things, both figuratively and literally.
Since many of us are visual learners and gravitate towards images in articles or presentations or websites, it is important to make the pictures that we present are accessible to all. One thing to consider is how well an image will show up for someone who cannot distinguish well between colors. As many as 8% of males and 0.5% of females of northern european descent have some variation of color blindness (Deeb 2005, as cited in Wong 2011). If you see colors in the normal range, it is very difficult to imagine what this experience would be like – and I know I am guilty of having not considered this for numerous presentations. But, there are steps one can take to understand the challenges presented by some colors and make your visual data presentation as accessible as possible.
There are a few tools that I have found recently along these lines. The first is a website called vischeck.com. This site allows you to upload pictures (png or jpg) and check how they would appear to someone with any of three color deficiencies: deuteranope (red/green), protanope (red/green), or tritanope (blue/yellow).
Next, try to use data analysis tools that allow you to have as much control over color settings as possible. Although there are more ‘user-friendly’ programs for statistics available, a statistical tool like R gives you lots and lots of options for controlling how your figures look. A good rule is to try to keep contrast high between colors.
I wrote about heatmaps before, and they can be an effective tool for showing differences in gene expression among samples. In the heatmaps below, the grid represents the expression of 19 genes (each row is a gene) for 14 samples (each column is a sample). [The 'trees' along the left and top of the heatmap indicate how the genes and samples cluster based on similarity of gene expression.] Some common colors used for heatmaps are shown in the figure below. At the top of the figure is the red/green combination, in which red indicates high gene expression and green indicates low gene expression. How this heatmap would appear with a red/green color deficiency is simulated to the right. The default heatmap colors in R range from yellow to red (middle left of the figure), where yellow represents low gene expression, and red represents high gene expression. Using a range of reds and blues (bottom left of the figure; red is high gene expression, blue is low) provides a broader range of colors, better contrast, and may give better resolution for those with red/green color deficiency (simulation on bottom right). Because of the use of red, this is still may not be the optimal design.
If needed, in R you can use a package called RColorBrewer, which provides a number of color palettes to choose from – a number of these correspond to colors you can choose specifically to be ‘colorblind safe’ at the website http://colorbrewer2.org/ . This color selection tool is originally intended to facilitate color selection for cartography design, but can work for other uses as well.
A nice short article about how to make good color figures for publication & presentation:
Wong, Bang (2011) Points of View: Color Blindness. Nature Methods 8(6):441.