Package 'rphylo'

Derive empirical frequencies of states

Description

Compute empirical frequencies of compound states

Usage

get_emp_freq(charfile)

Arguments

charfile

data.frame with taxa as rownames, two binary trait columns

Value

A numeric vector of length 4 giving the relative frequencies of compound states '00', '01', '10', and '11'.

Examples

get_emp_freq(tdata2)

---

Print method for dep_model objects

Description

Nicely formats MCMC output for dep_model objects.

Usage

## S3 method for class 'dep_model'
print(x, ...)

Arguments

x	An object of class 'dep_model'.
...	additional arguments (ignored)

Value

The object 'x', invisibly.

Examples

fit <- run_dep_model(
  ttree2,
  charfile = tdata2,
  burnin      = 5,
  iterations  = 50,
  sample_freq = 10,
  prior       = list(type = "exponential", mean=10),
  hp          = NULL,
  revjump     = FALSE,
  recon_nodes = NULL,
  tags        = NULL,
  pis         = c(1,1,1,1),
  seed        = 42
 )

fit   # triggers print.dep_model()

---

Run a dependent discrete-character model

Description

Performs Bayesian inference of correlated evolution between two binary traits, on one or more phylogenetic trees using MCMC or RJMCMC.

Usage

run_dep_model(
  trees,
  charfile,
  burnin,
  iterations,
  sample_freq = 1000L,
  tags = NULL,
  recon_nodes = NULL,
  revjump = FALSE,
  hp = NULL,
  hp_dev = HP_DEV_DEFAULT,
  prior = list(type = "exponential", mean = 10),
  pis = c(1, 1, 1, 1),
  seed = NULL,
  verbose = TRUE
)

Arguments

trees	multiPhylo object (trees should already be scaled if desired)
charfile	data.frame with taxa as rownames, two binary trait columns
burnin	integer number of iterations to discard before sampling
iterations	integer number of post-burnin iterations
sample_freq	integer record one sample every sample_freq iterations
tags	named list of character vectors defining nodes by their descendant taxa. Example: list(root_node = c("Taxon1","Taxon2","Taxon3"))
recon_nodes	character vector of tag names to reconstruct ancestral states for. Must be names present in tags. Set to NULL to skip ASR.
revjump	logical: use RJMCMC (default FALSE)
hp	hyperprior bounds list or NULL. If non-NULL, the parameters of 'prior' are themselves estimated. Format depends on prior type: exponential: list(mean = c(min, max)); uniform: list(min = c(min, max), max = c(min, max)); gamma: list(shape = c(min, max), scale = c(min, max)); lognormal: list(location = c(min, max), scale = c(min, max)).
hp_dev	numeric: proposal SD for hyperprior moves (default 1.0)
prior	list specifying prior on rate values. Supported forms include: list(type="exponential", mean=10); list(type="uniform", min=0, max=100)
pis	numeric vector of length 4 giving root state prior weights for states 00, 01, 10, 11; or "stationary". Weights need not sum to 1 (they are normalised internally). Default: c(1,1,1,1) matches 'pis none' in BayesTraits. Use uniform distribution: c(0.25,0.25,0.25,0.25). Use empirical distribution: get_emp_freq(charfile). Use stationary distribution: "stationary".
seed	integer random seed (NULL = random)
verbose	logical print progress (default TRUE)

Value

object of class "dep_model" containing: $log_lh numeric vector of sampled log-likelihoods;

$rates data.frame of sampled full rate vectors (8 columns);

$n_classes integer vector of sampled rate class counts;

$mapping list of sampled mapping vectors (mapping between rate values and rate classes);

$prior_samples data.frame of sampled prior parameters (NULL if no hp);

$anc named list of ASR matrices [n_samples x 4] per recon node, or NULL if recon_nodes = NULL.

NA rows indicate the node was not monophyletic on that tree.

$settings list of run settings.

Examples

tree <- ttree2
chars <- tdata2

fit <- run_dep_model(ttree2,
                   charfile    = tdata2,
                   burnin      = 10,
                   iterations  = 30,
                   sample_freq = 10,
                   prior       = list(type = "exponential", mean=10),
                   hp          = NULL,
                   revjump     = TRUE,
                   recon_nodes = c("root_node"),
                   tags        =list(root_node = ttree2[[1]]$tip.label),
                   pis         = get_emp_freq(tdata2),
                   seed        = 42
                   )

---

Scale branch lengths of phylogenetic trees

Description

Rescales all branch lengths in a 'multiPhylo' object so that the mean branch length across all trees equals 'target_mean_bl'.

Usage

scale_trees(multiphylo, target_mean_bl = 0.1)

Arguments

multiphylo	An object of class 'multiPhylo'.
target_mean_bl	target mean branch length. Default: 0.1.

Value

A list with:

trees

A 'multiPhylo' object with rescaled branch lengths.

scalar

The scalar multiplier applied to all branch lengths.

Examples

scale_trees(ttree2)

---

Example trait dataset

Description

Binary trait data corresponding to 'ttree'.

Usage

tdata

Format

A data frame with taxa as row names and 2 binary trait columns.

Source

Simulated data.

---

Example trait dataset (60 taxa)

Description

Binary trait data corresponding to 'ttree2'.

Usage

tdata2

Format

A data frame with taxa as row names and two binary trait columns.

Source

Simulated data.

---

Example phylogenetic tree

Description

A small phylogenetic tree for demonstrating rphylo analyses.

Usage

ttree

Format

An object of class 'phylo'.

Source

Simulated data.

---

Example phylogenetic trees (500 trees)

Description

500 phylogenetic trees for demonstrating rphylo analyses.

Usage

ttree2

Format

An object of class 'multiPhylo' containing 500 trees.

Source

Simulated data.

---