Generate a trait from ordinated empirical data

[TGuillerme]

Generate a trait from ordinated empirical data

This simulation template generates tree with a Brownian Motion trait with a decreasing variance for each trait to look like ordinated data.

What does it do?

What function does it uses?	yes/no	comments
Uses make.bd.params	✅	using crude.bd.est
Uses make.traits	✅
Uses make.modifications	❌
Uses make.events	❌

bd.params

High speciation parameters to create a bushy tree.

## Load empirical data from the dispRity package
library(dispRity)
data(BeckLee_tree)
data(BeckLee_mat99)

## Estimating birth death parameters from an empirical tree
my_bd.params <- crude.bd.est(BeckLee_tree, method = "estimate")

traits

Generating a 97 dimensional process with decreasing variance per dimensions.

## Getting the number of dimensions
n_dimensions <- ncol(BeckLee_mat99)

## Calculating the variance per dimensions
variance_per_dimension <- apply(BeckLee_mat99, 2, var)/max(apply(BeckLee_mat99, 2, var))

## Creating the variance-covariance matrix (no covariance)
scale_matrix <- matrix(0, ncol = n_dimensions, nrow = n_dimensions)
diag(scale_matrix) <- variance_per_dimension

## Creating the 97D brownian motion trait
my_traits <- make.traits(process = BM.process, n = n_dimensions,
                         process.args = list(Sigma = scale_matrix))

A running example

## My favorite seed
set.seed(9)

## Some stopping rules
my_stop.rule <- list(max.time = 140)

## The simulation
my_simulation <- treats(stop.rule = my_stop.rule,
                        bd.params = my_bd.params,
                        traits    = my_traits)

## The tree
plot(my_simulation)

## Show the simulated variance
simulated_variance <- apply(my_simulation$data, 2, var)/max(apply(my_simulation$data, 2, var))
plot(simulated_variance, ylab = "scaled_variance")
lines(y = variance_per_dimension, x = 1:97)
legend(x = "topright", pch = c(21, NA), lty = c(NA, 1), legend = c("simulated", "observed"))

Reference

If you use this template in a publication, please cite:

• treats
• R