sim.bdtypes.stt.taxa: sim.bdtypes.stt.taxa: Simulating multitype birth-death trees with a fixed number of tips sampled through time.

Description

sim.bdtypes.stt.taxa simulates trees on n tips sampled through time under a multitype birth-death process.

Usage

sim.bdtypes.stt.taxa(n,lambdavector,deathvector,
sampprobvector,init=-1,EI=FALSE,eliminate=0)

Arguments

Number of sampled tips.

lambdavector

Matrix of dimension kxk, where k is the number of different states. The entry (i,j) is the rate with which an individual in state i gives rise to a new lineage of state j.

deathvector

Vector of dimension k, the entry i is the death rate of an individual in state i.

sampprobvector

Vector of dimension k, the entry i is the probability of an individual in state i being sampled upon death, i.e. being included into the final tree.

init

Default is -1, meaning the initial individual is in a random state (which is chosen from the equilibrium distribution of states). If init>0, then the initial state is 'init'.

If EI=TRUE a model with two types, namely exposed and infectious individuals, is assumed. Infectious individuals transmit and give rise to exposed individuals with rate lambdavector[2,1], and exposed individuals become infectious with rate lambdavector[1,2]. Exposed individuals have a 0 death rate and cannot be sampled. For an example simulation see below.

eliminate

Only relevant if EI=TRUE. Under EI=TRUE all sampled tips are in state 2. If eliminate>0, the first eliminate tips are marked with state 1. This facilitates further analysis, e.g. we now can easily prune these first eliminate tips to mimic no sampling close to the epidemic outbreak.

Value

out

Phylogenetic tree with 'n' sampled tips. In out$states, the states for the tips are stored.

References

T. Stadler, S. Bonhoeffer. Uncovering epidemiological dynamics in heterogeneous host populations using phylogenetic methods. Phil. Trans. Roy. Soc. B, 368 (1614): 20120198, 2013.

Examples

Run this code

# NOT RUN {
# Simulate two trees with 10 tips
set.seed(1)
n<-10
lambda <- rbind(c(2,1),c(3,4))
mu <- c(1,1)
sampprob <-c(0.5,0.5)
numbsim<-2
trees<-lapply(rep(n,numbsim),sim.bdtypes.stt.taxa,
lambdavector=lambda,deathvector=mu,sampprobvector=sampprob)

# Testing the model with exposed class (EI = TRUE)
set.seed(2)
# simulate tree with expected incubation period of 14 days, 
# infectious period of 7 days, and R0 of 1.5:
mu <- c(0,1/7)
lambda <- rbind(c(0,1/14),c(1.5/7,0))
# sampling probability of infectious individuals is 0.35:
sampprob <-c(0,0.35)
# we stop once we have 20 samples:
n <- 20
# we simulate one tree:
numbsim<-1
# We mark first eliminate=10 tips such that we can easily drop them later
# (if deleting these 10 tips, we mimic no sampling close to the outbreak)
trees<-lapply(rep(n,numbsim),sim.bdtypes.stt.taxa,lambdavector=lambda,deathvector=mu,
sampprobvector=sampprob,EI=TRUE,eliminate=10)
# }