Skip to contents

Generic group-with-known-covariance random effect (brms-style)

Source: R/brms-sugar.R
gr.Rd

A user-friendly wrapper for any random effect whose covariance is a user-supplied matrix. Mirrors brms's (1 | gr(species, cov = M)). Inside a gllvmTMB() formula, this desugars to propto(0 + group | trait). The actual covariance matrix used by the engine is the top-level phylo_vcv = argument of gllvmTMB() when one is supplied; otherwise the formula-level cov = matrix is used. When both are given, the top-level phylo_vcv = wins and cov = is ignored. Same convention as phylo() and phylo_rr().

Usage

gr(group, cov = NULL)

Arguments

group

An unquoted column name giving the grouping factor.

cov

Optional covariance matrix. Used when no top-level phylo_vcv is passed to gllvmTMB(); ignored when phylo_vcv is also supplied.

Value

A formula marker; never evaluated.

Examples

if (requireNamespace("ape", quietly = TRUE)) {
  set.seed(1)
  tree <- ape::rcoal(60); tree$tip.label <- paste0("sp", seq_len(60))
  sim <- simulate_site_trait(
    n_sites = 1, n_species = 60, n_traits = 3,
    mean_species_per_site = 60,
    Cphy = ape::vcv(tree, corr = TRUE),
    sigma2_phy = rep(1, 3), sigma2_eps = 0.2, seed = 1
  )
  df <- sim$data
  df$species <- factor(tree$tip.label[as.integer(df$species)],
                       levels = tree$tip.label)
  Cphy <- ape::vcv(tree, corr = TRUE)
  # \donttest{
  # gr() is a deprecated alias and prints a message; new code can use
  # phylo_indep(0 + trait | species, tree = tree, common = TRUE).
  fit <- gllvmTMB(value ~ 0 + trait + gr(species),
                  data      = df,
                  trait     = "trait",
                  unit      = "species",
                  cluster   = "species",
                  phylo_vcv = Cphy)
  # }
}
#> ! Formula keyword `gr()` is a deprecated alias; use `phylo_indep()` for new
#>   code.
#> ℹ `phylo_indep(0 + trait | species, common = TRUE)` =/=> `gr(species)`
#> → Aliases will be dropped at the next minor release. See `?phylo_indep`.