Symbolize a phylolm fit (Gaussian PGLS, v0.21.4 first slice)

Builds a symbolized_model from a phylolm object. v0.21.4 covers BM and Pagel's lambda evolutionary models; OU / EB / kappa / delta / trend produce a symbolized_model but carry the model name in metadata$phylo_model for downstream diagnostics rather than full template support.

Usage

# S3 method for class 'phylolm'
symbolize(fit, symbols = NULL, units = NULL, context = NULL, data = NULL, ...)

Arguments

fit

A fitted statistical model object.

symbols

Optional named character vector mapping variable names to user-supplied LaTeX symbols, e.g. c(body_mass = "W_i", temperature = "T_i").

units

Optional named character vector mapping variable names to units, e.g. c(body_mass = "g", temperature = "C").

context

Optional short character description of the model, e.g. "avian body-size location-scale model".

data

Optional data frame. If NULL, falls back to fit$model.frame (phylolm stores it there).

...

Reserved for method-specific extra arguments.

Value

A symbolized_model object.

Details

phylolm fits the PGLS marginal form y = X*beta + e, e ~ N(0, sigma^2 * C(alpha)) where C(alpha) is a tree-derived correlation matrix. There is no explicit u_p random-effect — the phylogenetic signal lives entirely in the residual covariance. metadata$phylo_representation is set to "pgls_marginal" so downstream renderers can phrase this correctly.

Confidence intervals

phylolm reports Wald-type standard errors via vcov(fit). The confint_low / confint_high columns of fixed_effects are computed as estimate +/- 1.96 * std_error, and ci_method is set to "wald".

References

Ho, L.S.T. & Ané, C. (2014). A linear-time algorithm for Gaussian and non-Gaussian trait evolution models. Systematic Biology, 63(3), 397-408.

Tung Ho, L.S. & Ané, C. (2014). Intrinsic inference difficulties for trait evolution with Ornstein-Uhlenbeck models. Methods in Ecology and Evolution, 5(11), 1133-1146.