brownie.lite: Likelihood test for rate variation
Description
This function takes a modified "phylo" object with a mapped binary or multistate trait (see read.simmap) and data for a single continuously valued character. It then fits the Brownian rate variation ("noncensored") model of O'Meara et al. (2006; Evolution). This is also the basic model implemented in Brian O'Meara's "Brownie" program.Usage
brownie.lite(tree, x, maxit=2000, test="chisq", nsim=100, se=NULL)
Arguments
x
a vector of tip values for species; names(x) should be the species names.
maxit
an optional integer value indicating the maximum number of iterations for optimization - may need to be increased for large trees.
test
an optional string indicating the method for hypothesis testing - options are "chisq" or "simulation".
nsim
number of simulations (only used if test="simulation").
se
a vector containing the standard errors for each estimated mean in x.
Value
- a list with the following components:
- sig2.singleis the rate for a single rate model - this is usually the "null" model.
- a.singleis the estimated state at the root node for the single rate model.
- var.singlevariance on the single rate estimator - obtained from the Hessian.
- logL1log-likelihood of the single-rate model.
- k1number of parameters in the single rate model (always 2).
- sig2.multipleis a length p (for p rates) vector of BM rates from the multi-rate model.
- a.multipleis the estimated state at the root node for the multi-rate model.
- var.multiplep x p variance-covariance matrix for the p rates - the square-roots of the diagonals should give the standard error for each rate.
- logL.multiplelog-likelihood of the multi-rate model.
- k2number of parameters in the multi-rate model (p+1).
- P.chisqP-value for a likelihood ratio test against the $\chi^2$ distribution; or
- P.simP-value for a likelihood ratio test agains a simulated null distribution.
- convergencelogical value indicating if the likelihood optimization converged.
Details
Sampling error in the estimation of species means can also be accounted for by assigning the vector se with the species specific sampling errors for x.References
O'Meara, B. C., C. Ane, M. J. Sanderson, and P. C. Wainwright. (2006) Testing for different rates of continuous trait evolution using likelihood. Evolution, 60, 922--933.