Introducing the plotting capabilities of the `phylosignal` package

The phylosignal package comes with functions designed to plot trait values and phylogeny together. These functions are generics for phylo4d objects. This document present how to use them.

Data

First, we load the package phylosignal and some others. We will also use the dataset carnivora from adephylo.

library(ape)
library(adephylo)
library(phylobase)
library(phylosignal)
data(carni19)

Here is a phylogenetic tree of 19 carnivora species.

tre <- read.tree(text = carni19$tre)

And we create a dataframe of 3 traits for the 19 carnivora species. - Body mass - Random values - Simulated values under a Brownian Motion model along the tree

dat <- data.frame(carni19$bm)
dat$random <- rnorm(dim(dat)[1], sd = 10)
dat$bm <- rTraitCont(tre)

We can combine phylogeny and traits into a phylo4d object.

p4d <- phylo4d(tre, dat)

Basics

Once we have a phylo4d object, we can plot it… There are three plotting functions: barplot, dotplot and gridplot. These functions are actually wrappers of the function multiplot.phylo4d.

barplot(p4d)

dotplot(p4d)

gridplot(p4d)

Each of these functions can be used with one of the three tree styles: phylogram, cladogram and fan. For example, here is a dotplot with a cladogram.

dotplot(p4d, tree.type = "cladogram")

And here a gridplot with a fan tree.

gridplot(p4d, tree.type = "fan", tip.cex = 0.6, show.trait = FALSE)

Select the ratio of the plot occupied by the tree.

barplot(p4d, tree.ratio = 0.5)

Control which traits to plot and their order.

barplot(p4d, trait = c("bm", "carni19.bm"))

Add simple error bars.

mat.e <- matrix(abs(rnorm(19 * 3, 0, 0.5)), ncol = 3,
                dimnames = list(tipLabels(p4d), names(tdata(p4d))))
barplot(p4d, error.bar.sup = mat.e, error.bar.inf = mat.e)

It is also possible to open a fan tree with a specified angle.

barplot(p4d, tree.type = "fan", tip.cex = 0.6, tree.open.angle = 160, trait.cex = 0.6)

Colors

It’s easy to color bars ‘by species’ with a vector.

barplot(p4d, bar.col = rainbow(19))

And for a finer control, one can use a matrix. Here, negative values in red.

mat.col <- ifelse(tdata(p4d, "tip") < 0, "red", "grey35")
barplot(p4d, center = FALSE, bar.col = mat.col)

Clearly identify traits with colored backgrounds:

barplot(p4d, trait.bg.col = c("#F6CED8", "#CED8F6", "#CEF6CE"), bar.col = "grey35")

For gridplots, cells are colored with a color palette, using the cell.col argument.

gridplot(p4d, tree.type = "fan", tree.ratio = 0.5,
         show.trait = FALSE, show.tip = FALSE,
         cell.col = terrain.colors(100))

Combine arguments for sophisticated plots

tip.col <- rep(1, nTips(p4d))
tip.col[(mat.col[, 2] == "red") | (mat.col[, 3] == "red")] <- 2
barplot(p4d, center = FALSE, trait.bg.col = c("#F6CED8", "#CED8F6", "#CEF6CE"),
        bar.col = mat.col, tip.col = tip.col, trait.font = c(1, 2, 2))

And control many other things! See ?multiplot.phylo4d for more informations.

Introducing the plotting capabilities of the phylosignal package

Data

Basics

Colors

.

.

Introducing the plotting capabilities of the `phylosignal` package