Profile-likelihood curve(s) for per-trait phylogenetic signal

For each requested trait, sweep over a grid of candidate values of the 2-component phylogenetic signal \(H^2_t = \sigma^2_{\text{phy},t} / (\sigma^2_{\text{phy},t} + \sigma^2_{\text{non},t})\), refit the model under each constraint, and return the LR curve. Currently supports only the simple 2-component case (phylo_diag + species-level diag); 3-component decompositions error with a note.

Usage

profile_phylo_signal(
  fit,
  trait_idx = NULL,
  n_grid = 21L,
  grid_extent = 4,
  conf_level = 0.95
)

Arguments

fit

A fit returned by gllvmTMB().

trait_idx

Integer vector of trait indices, or NULL for all.

n_grid

Integer; number of grid points per trait. Default 21.

grid_extent

Numeric; total grid width as a multiple of a robust scale (default 4).

conf_level

Confidence level. Default 0.95.

Value

A data.frame of class profile_phylo_signal, profile_derived, data.frame.

Details

Parallel CURVE endpoint to profile_ci_phylo_signal().

See also

profile_ci_phylo_signal(), plot.profile_derived().