mlskygrid: Maximum likelihood non-parametric estimation of effective...

In emvolz-phylodynamics/mlesky: Maximum likelihood estimation of effective population size through time, growth rates, and phylodynamic regressions

Maximum likelihood non-parametric estimation of effective population size through time

Description

Maximum likelihood non-parametric estimation of effective population size through time

Usage

mlskygrid(
  tre,
  sampleTimes = NULL,
  res = 25,
  tau = 1,
  tau_lower = NULL,
  tau_upper = NULL,
  tau_tol = 0.001,
  ncross = 5,
  ncpu = 1,
  quiet = TRUE,
  NeStartTimeBeforePresent = Inf,
  ne0 = NULL,
  adapt_time_axis = FALSE,
  model = 1,
  formula = NULL,
  data = NULL
)

Arguments

tre	A dated phylogeny in ape::phylo or treedater format (see documentation for ape). This can also be a multiPhylo or list of trees, in which case each is treated as a clade sampled from within the same population. In this case the sampleTimes vector should be supplied so that clades can be aligned in time.
sampleTimes	An optional named vector of sample times for each taxon. Names should correspond to tip labels in trees. This is required if providing a list of trees or covariates.
res	Length of time axis over which to estimate Ne(t) (integer). If NULL, will heuristically search for a good value
tau	Precision parameter. Larger values generate smoother trajectories of Ne(t). If NULL, will optimize using cross-validation.
tau_lower	Lower bound for precision parameter if estimating
tau_upper	Upper bound for precision parameter if estimating
tau_tol	Optimization tolerance when optimizing tau by cross-validation
ncross	Number of folds in cross-validation
ncpu	If doing cross-validation, each fold will be handled in parallel if ncpu > 1 (see parallel package)
quiet	Provide verbose output from optimizer?
NeStartTimeBeforePresent	If <Inf, will only estimate Ne(t) between the most recent sample and this time before the most recent sample
ne0	Vector of length *res* giving starting conditions of Ne(t) for optimization, or a single value which will be used as rep(ne0,res)
adapt_time_axis	If TRUE will choose Ne(t) change points in periods with high frequency of phylogenetic branching
model	Model to use, can be 1 (=skykappa model), 2 (=skygrid model) or 3 (=skygrowth model)
formula	Formula for use of covariates. The left hand side should be one of 'diffLogNe', 'logNe' or 'diffDiffLogNe'. For example, if modeling the effect of a single covariate x on growth of Ne, an appropriate formula may be ‘diffLogNe ~ x - 1' where ’-1' specifies that an intercept is not estimated.
data	For use of covariates, data.frame must include 'time'

Value

A fitted model including effective size through time

emvolz-phylodynamics/mlesky documentation built on Feb. 13, 2025, 1:47 p.m.