phytools plotTree.wBars coloring the bars by our continuous trait

Today, a phytools user writes:

“I am trying to use the command plotTree.wBars in phytools to add bars as tips of a circular tree. The bars represent values of a continuous trait assigned to the tips. I was wondering if it is possible to also color these bars according to their values?”

The answer is, of course, yes. Let's see a couple of different examples.

For this example, I'll use a phylogeny and dataset for liolaemid lizards from Esquerre et al. (2019).

Luke Harmon & I use these data in our upcoming book, so the easiest way to get them is from the book website.

## load phytools
library(phytools)
## read tree
liol.tree<-read.nexus(file="http://www.phytools.org/Rbook/7/Liolaemidae.MCC.nex")
liol.tree

## 
## Phylogenetic tree with 258 tips and 257 internal nodes.
## 
## Tip labels:
##   Ctenoblepharys_adspersa, Liolaemus_abaucan, Liolaemus_koslowskyi, Liolaemus_albiceps, Liolaemus_irregularis, Liolaemus_ornatus, ...
## 
## Rooted; includes branch lengths.

## read data
liol.data<-read.csv(file="http://www.phytools.org/Rbook/7/Liolaemidae.data.csv",
    row.names=1)
head(liol.data)

##                         parity_mode max_altitude temperature
## Ctenoblepharys_adspersa           O          750       23.05
## Liolaemus_abaucan                 O         2600       20.20
## Liolaemus_albiceps                V         4020       12.38
## Liolaemus_andinus                 V         4900       11.40
## Liolaemus_annectens               V         4688        5.10
## Liolaemus_anomalus                O         1400       23.78

I'm going to first plot the tree with max_altitude using plotTree.wBars. The only trick I'm adding is that I'm going to steal some internal calculations (for scale=NULL) to add a scale bar to the tree using polygon. If we had set a specific value for scale, we'd use that instead of the variable ss.

max_altitude<-setNames(liol.data$max_altitude,rownames(liol.data))
plotTree.wBars(liol.tree,max_altitude,type="fan",lwd=1,
    tip.label=TRUE,fsize=0.3)
ss<-0.3*max(nodeHeights(liol.tree))/diff(range(max_altitude))
h<-0.95*par()$usr[2]
bb<-3000
polygon(x=-h+c(0,0,ss*bb,ss*bb),y=h+c(1,0,0,1),col="grey")
text(x=-h+ss*bb,y=h+0.5,paste(bb,"m",sep=""),pos=4)

Next, let's do it again, but this time coloring all the bars greater than or equal to 2000m with one color (blue) and all those below 2000m another (red). For this one, I will set the plotTree.wBars argument col which specifies the fill color of our polygons. The easiest way for me to do this here is by creating a vector col containing only the color blue (palette()[4] in this case), and then substituting palette()[2] for all max_altitude<2000.

max_altitude<-setNames(liol.data$max_altitude,rownames(liol.data))
col<-setNames(rep(palette()[4],length(max_altitude)),names(max_altitude))
col[max_altitude<2000]<-palette()[2]
plotTree.wBars(liol.tree,max_altitude,type="fan",lwd=1,
    tip.label=TRUE,fsize=0.3,col=col)
ss<-0.3*max(nodeHeights(liol.tree))/diff(range(max_altitude))
h<-0.9*par()$usr[2]
bb<-2000
polygon(x=-h+c(0,0,ss*bb,ss*bb),y=h+c(1,0,0,1),col=palette()[2])
text(x=-h+ss*bb,y=h+0.5,paste("<2000m",sep=""),pos=4,cex=0.8)
h<-0.9*par()$usr[2]
bb<-2000
polygon(x=-h+c(0,0,ss*bb,ss*bb),y=0.95*h+c(1,0,0,1),col=palette()[4])
text(x=-h+ss*bb,y=0.95*h+0.5,paste(">2000m",sep=""),pos=4,cex=0.8)

Lastly, let's do it based on a continuous color gradient. I guess this should be the hardest, but I'm pretty sure we can figure it out!

In this case, what I decided to do was to first discretize my continuous character (altitude) by normalizing it to the interval [0,1] and then multiplying this quantity by 1,000 (actually, 999 + 1). After rounding, I have created an index vector between 1 & 1,000. I then create my color ramp with 1,000 colors and pull out those that I want to use in my graph. I chose the topo.colors palette, but I could have picked any.

ind<-round((max_altitude-min(max_altitude))/diff(range(max_altitude))*999)+1
cols<-setNames(topo.colors(n=1000)[ind],names(ind))
plotTree.wBars(liol.tree,max_altitude,type="fan",lwd=1,
    tip.label=TRUE,fsize=0.3,col=cols)
ss<-0.3*max(nodeHeights(liol.tree))/diff(range(max_altitude))
h<-0.9*par()$usr[2]
bb<-10000
add.color.bar(bb*ss,topo.colors(n=1000),title="",
    subtitle="max altitude (m)",lims=range(max_altitude),
    x=-h,y=h,prompt=FALSE,fsize=0.8)

That's it folks!