Simulating phylogenies

A great advantage of TiPS, besides its computational efficiency, is that it can generate phylogenies from the population dynamics trajectories using a backward-in-time approach.

For this, we may need a vector of sampling dates.

For the SIR example, these are stored here:

dates <- system.file("extdata", "SIR-dates.txt", package = "TiPS")

The simulate_tree function simulates a phylogeny from a trajectory object and uses a set of sampling dates:

sir_tree <- simulate_tree(
  simuResults = traj_dm,
  dates = dates,
  deme = c("I"), # the type of individuals that contribute to the phylogeny
  sampled = c(I = 1), # the type of individuals that are sampled and their proportion of sampling
  root = "I", # type of individual at the root of the tree
  isFullTrajectory = FALSE, # deads do not generate leaves
  nTrials = 5,
  addInfos = FALSE) # additional info for each node

The sampled option can be used for labeled phylogenies (i.e. with multiple host types) but it requires specifying the proportion of each label type. The root option indicates the state of the individual initiating the dynamics. See the next section for details.

The complete phylogeny can be obtained (therefore neglecting the sampling dates) with the option isFullTrajectory.

Finally, some runs may fail to simulate a phylogeny from a trajectory for stochastic reasons. The nTrials parameter indicates the number of unsuccessful trials allowed before giving up.

The simulated phylogeny can be visualised using:

ape::plot.phylo(sir_tree, cex = .5)

Writing the tree in an output file with the option outFile is possible. If a file name is given as input, by default the tree is written in a Newick format. To write the simulated tree in a Nexus format, the option format must be specified and fixed to Nexus (option format = "nexus").

sir_tree <- simulate_tree(
  simuResults = traj_dm,
  dates = dates,
  deme = c("I"),
  sampled = c(I = 1),
  root = "I",
  isFullTrajectory = FALSE,
  nTrials = 5,
  addInfos = FALSE,
  outFile = "sir_tree.nexus",
  format = "nexus"
)