Skip to main content

R for beginners and intermediate users 3: plotting with colours

For my third post on my R tutorials for beginners and intermediate users, I shall finally touch on the subject matter that prompted me to start these tutorials - plotting with group structures in colour.

If you are familiar with R, then you may have noticed that assigning group structure is not all that straightforward. You can have a dataset that may have a column specifically for group structure such as this:

B0 B1 B2 Family
Acrocanthosaurus 0.308 -0.00329 3.28E-05 Allosauroidea
Allosaurus 0.302 -0.00285 2.04E-05 Allosauroidea
Archaeopteryx 0.142 -0.000871 2.98E-06 Aves
Bambiraptor 0.182 -0.00161 1.10E-05 Dromaeosauridae
Baryonychid 0.189 -0.00238 2.20E-05 Basal_Tetanurae
Carcharodontosaurus 0.369 -0.00502 5.82E-05 Allosauroidea
Carnotaurus 0.312 -0.00324 2.94E-05 Neoceratosauria
Ceratosaurus 0.377 -0.00522 6.07E-05 Neoceratosauria
Citipati 0.278 -0.00119 5.08E-06 Oviraptorosauria
Coelophysis 0.351 -0.00921 1.63E-04 Coelophysoidea
Compsognathus 0.165 -0.00160 1.44E-05 Basal_Coelurosauria
Daspletosaurus 0.276 -0.00272 1.98E-05 Tyrannosauroidea
.
.
.

and you'd hope that there is an intuitive and easy way of specifying colour or grouping structure based on this last column. The short answer is, yes, there is. But when I was a novice, it was not necessarily easy to grasp.

But before we go into that, let's review simple plotting in R.

Let's say we want to plot the first 2 principal components from a principal components analysis (PCA) based on the complete dataset introduced briefly above. I just wanted to show PCA plots instead of the original variables because I wanted to represent all three variables in two dimensions. So assuming that we have the PCA results and we call the scores along the first two PCs, PC1 and PC2 respectively, then we can plot them like this:

  plot(PC1, PC2)

and it would give you a default plot that looks like this:


The default plot setting is as shown above, with simple open circles. If you don't like the symbols then you can use the pch argument to change it the way you like it:

  plot(PC1, PC2, pch=19)

Here, I use pch=19 which is just a filled circle:


Now let's start adding group structure using colour.

First, we must assign the group structure. Let's say that the data table above is called data, then the column containing group information is either data[,4] or data$Family. We need to use this information to assign grouping structure to the plot. One way of achieving this is to assign the contents of this column (or vector in R) as a "factor" and call the R object something like family:

  family <- as.factor(data[,4])

A factor in R is basically like categorical data so each unique entry, like "Allosauroidea" or "Tyrannosauroidea", is not just a random line of text but is treated like categories within a factor variable. Now family is an R object of class factor with 12 levels, i.e, the individual clades. So the object family acts as our grouping structure. We can use this grouping structure to assign colours with the col argument to our plot like this:

  plot(PC1, PC2, pch=19, col=family)

and the resulting plot is as follows.


In this plot, the data points have been separated into 12 levels each with a corresponding colour. However, if you look closely, there are only eight colours in the above plot, black, red, green, blue, cyan, magenta, yellow and gray. This is because the default colour palette in R is set at these eight colours, and any new colour assignment exceeding this number would just recycle these eight colours in order. For instance, in our case, the first to the eighth clades in family Allosauroidea to Ornithomimosauria get assigned the colours from black to gray, and again for the ninth group onwards, i.e., outgroup = black, Oviraptorosauria=red, Therizinosauroidea=green, and Tyrannosauroidea=blue. We can resolve that problem by using a different palette.

There are several built-in colour palettes that can give your plot quite different looks. You can have a look at some of these by typing help(palette) or help(colors) in R. A few examples are: rainbow, heat.colors, topo.colors, and terrain.colors. In order to change the palette from default to any of these colour palettes, we have to first create a colour palette by one of the following:

  col.rainbow <- rainbow(12)
  col.topo <- topo.colors(12)
  col.terrain <- terrain.colors(12)

where the "12" within the parentheses indicates the number of colours in the palette. And to assign one of these colour palettes for instance col.rainbow as the global colour palette just type:

  palette(col.rainbow)

Once you've done that, you can retype the plotting function:

  plot(PC1, PC2, pch=19, col=family)

but this time, the colour scheme would be different:


Next, if we use palette(col.topo) instead:


and palette(col.terrain):


All these different colour palettes are pretty and useful in different ways, but maybe they're not exactly what you're looking for, or maybe you want to control what colours to use for each group. For that, you'd have to specify your own colours. To specify the colours you want, you can just directly specify colour in the plotting function:

  plot(PC1, PC2, pch=19, col="red")

which gives a plot like this:


But this shows just one colour, which is not bad if you're going to overlay each group using different colours every time, for instance using points()- maybe I'll expand on points() on a later post. You can also specify a vector of colours if you want to use more than one colour.

  plot(PC1, PC2, pch=19, col=c("red","black"))

and the resulting plot would look like this:


It's not readily obvious, but this plot just alternates between red and black throughout the whole data set; Acrocanthosaurus = red, Allosaurus = black, Archaeopteryx = red, Bambiraptor = black, and so on. Here's the same plot with taxa names superimposed using text():


The same thing would happen if you provide a vector of 12 colours to reflect our 12 clades; the 12 colours would alternate through the list of 42 taxa instead of through the 12 clades. There are two ways of resolving this (maybe there are more, but I know of two).

First, you can prepare a vector of colour names such as "red", "black" or "burlywood" so that members of the same clade are assigned the same colour, for instance Allosauroidea = red and Tyrannosauroidea = blue. Naturally, you'd have a vector of length 42 (or number of taxa) with 12 different colour names (corresponding to the number of groups). You can view all the available colour names in R by typing colors(). Instead of typing all 42 colour assignments straight into the plotting function, we can create an R object in advance. The simplest way to do this is to create an extra column in your dataset in Excel containing the appropriate colour names. If this column is called Colours within your dataset data then you can call it up as data$Colours, or if it was the fifth column, then data[,5]. And because we'd want R to recognise the colour names as characters it's probably best if we use the as.character( ) function:

  colour <- as.character(data$Colours)

If you don't use as.character( ) for instance like this:

  colour <- data$Colours

then colour is not a character vector but a factor vector. Plotting using this factor vector will not produce a plot of your preferred colours but based on whatever your palette setting is. So if it is set on default, then you'd just get eight colours.

Using the character vector instead of the factor vector as the colour argument we can plot again:

  plot(PC1, PC2, pch=19, col=colour)

and get this plot:


and with text superimposed:


You can now clearly see that each clade has its own colour and every single member of a given clade is of a same colour. For instance, all tyrannosauroids are in blue, while allosauroids are in red.

The second way of getting this plot is to create a character vector of 12 colour names:

  col.list <- c("red","slategray","seagreen",....,"blue")

and set this vector as the colour palette:

  palette(col.list)

If you plot using the factor vector family as the grouping structure:

  plot(PC1, PC2, pch=19, col=family)

you should get the same plot as the one using the character vector colour for col=colour shown above.

I think that is about it for colour plotting, and I hope I did not make it too confusing. In summary, you can specify colour schemes either 1) by referring to a character vector (containing the colour names; e.g., colour) in the plotting function, or 2) by setting a character vector (e.g. col.list) as the colour palette and specifying a factor vector (e.g. family) as the col argument in your plotting function.

In a future post, I plan to expand on the points( ) function I briefly mentioned, but also touch on how to superimpose text using the text( ) function

Comments

Unknown said…
How would you create a legend according to each colour used in the plot?
Raptor's Nest said…
Thanks for that question! I'll make a post on that soon.
Dr.Thermophile said…
Thank you! I've been making myself crazy trying to figure out how to do this.
Unknown said…
This is so helpful! Thank you for the clear explanation
leah said…
Thank you! that was exactly what i was looking for
Unknown said…
Thank you sooo much!!!!! <3
DimShumFlava said…
Hi. I've an interesting scenario: What if there were certain groups that you wanted to share the same color but others not ?

For example, you wanted Allosauroidea and Tyrannosauroidea colored in as yellow, but all remaining groups to have unique colors.

Furthermore, what if you have multiple data sets and you wanted to generate multiple plots... and you want Allosauroidea and Tyrannosauroidea across all of the multiple data sets to be yellow... AND the groups are in a different order in each data set?

Thank you, I hope what I've described makes sense !
Raptor's Nest said…
Hi DimShumFlava,

Sorry for the late response, I haven't been checking my blog much recently...

The simplest way to assign colours to groups is by preparing a character vector of colour names for each taxon, and just assigning "yellow" for allosauroid and tyrannosauroid taxa while assigning unique colours for all other groups. This can be done for each dataset separately, and can be done in either R or some spreadsheet application like Excel.

If you were asking how to change the colours when using factor levels, then I suppose you can create a character vector of colour names for each level of the factor variable, so 11 different colours in our case (12 unique groups - 1 for the duplicated colour, yellow), making sure that the factor levels and colour vector match up in order so that you're assigning the correct colour to the desired groups. Then you can set the colour palette using your colour vector. You can see the levels of the factor variable, e.g., family, with the levels() function, levels(family), and create your colour vector using this as a guide.

The default in R for factor orders is based on the order when the factor is coerced into a character vector, so by alphabetical order. This means that even across datasets, as long as you've got the same factor variable with the same number of levels, the colour information should be consistent.

Does this answer your question?

Popular posts from this blog

The difference between Lion and Tiger skulls

A quick divergence from my usual dinosaurs, and I shall talk about big cats today. This is because to my greatest delight, I had discovered today a wonderful book. It is called The Felidæ of Rancho La Brea (Merriam and Stock 1932, Carnegie Institution of Washington publication, no. 422). As the title suggests it goes into details of felids from the Rancho La Brea, in particular Smilodon californicus (probably synonymous with S. fatalis ), but also the American Cave Lion, Panthera atrox . The book is full of detailed descriptions, numerous measurements and beautiful figures. However, what really got me excited was, in their description and comparative anatomy of P. atrox , Merriam and Stock (1932) provide identification criteria for the Lion and Tiger, a translation of the one devised by the French palaeontologist Marcelin Boule in 1906. I have forever been looking for a set of rules for identifying lions and tigers and ultimately had to come up with a set of my own with a lot of help

Hind limb proportions do not support the validity of Nanotyrannus

While it was not the main focus of their paper, Persons and Currie (2016) , in a recent paper in Scientific Reports hinted at the possibility of Nanotyrannus lancensis being a valid taxon distinct from Tyrannosaurus rex , using deviations from a regression model of lower leg length on femur length. Similar to encephalisation quotients , Persons and Currie devised a score (cursorial-limb-proportion; CLP) based on the difference between the observed lower leg length and the predicted lower leg length (from a regression model) expressed as a percentage of the observed value. The idea behind this is pretty simple in that if the observed lower leg length value is higher than that predicted for its size (femur length), then that taxon gets a high CLP score. I don't particularly like this sort of data characterisation (a straightforward regression [albeit with phylogeny, e.g. pGLS] would do the job well), but nonetheless, Persons and Currie found that when applied to Nanotyrannus , it