Skip to contents

Reconstruct ancestral embeddings under Brownian Motion (pure-R reference)

Source: R/reference.R
reconstruct_ancestral_bm_r.Rd

Estimates the root state with the GLS estimator and then computes the conditional expectation of every internal node. With `obs_sigma = 0` the resulting node estimates match `phytools::fastAnc`.

Usage

reconstruct_ancestral_bm_r(
  tree,
  z_obs,
  rate = NULL,
  z_anc = NULL,
  obs_sigma = 0
)

Arguments

tree

A `phylo` object.

z_obs

Numeric matrix (`S x D`) of observed tip embeddings.

rate

Optional per-dimension BM rate. If `NULL`, estimated via ML under `obs_sigma = 0`.

z_anc

Optional per-dimension root state. If `NULL`, estimated by GLS.

obs_sigma

Observation noise sd added to the diagonal of the tip covariance matrix.

Value

List with `z_anc` (length `D`), `rate` (length `D`), and `node_estimates` (`n_nodes x D`). Row `u` of `node_estimates` corresponds to ape node number `n_tips + u`.