Package index • gllvmTMB

Start here

Fit a stacked-trait GLLVM and load beginner-friendly examples

gllvmTMB(): Fit multivariate response models from wide or long data
traits(): Wide-format trait marker for the gllvmTMB() formula LHS
gllvmTMBcontrol(): Control parameters for gllvmTMB()
miss_control(): Missing-data control for gllvmTMB()
impute_model(): Predictor model for a missing covariate used inside mi()

Core covariance keywords

Start with the ordinary decomposition, diagonal baseline, or full covariance before adding a relationship source. The grid has three modes; the one-shared-variance case is now indep(…, common = TRUE) (the soft-deprecated scalar() is under Deprecated below)

latent(): Latent-factor (reduced-rank) random effect: latent(0 + trait | g, d = K)
indep(): Per-trait marginal variance: indep(0 + trait | g)
dep(): Full unstructured trait covariance: dep(0 + trait | g)

Advanced formula keywords and shorthands

First-class mode-dispatch wrappers; the bar form (e.g. phylo(0 + trait | species, mode = …)) is current, while the legacy bare-alias form is soft-deprecated. Public articles prefer the explicit core keywords

phylo(): Phylogenetic random effect: lme4-bar mode-dispatch wrapper
spatial(): Spatial random field: lme4-bar mode-dispatch wrapper

Relatedness and spatial helpers

Build relationship matrices, sparse precisions, meshes, and spatial diagnostics

pedigree_to_A(): Pedigree to additive-genetic relatedness matrix A
pedigree_to_Ainv_sparse(): Sparse pedigree inverse-A via Henderson-Quaas (MCMCglmm-free)
make_cross_kernel(): Build a cross-lineage relatedness kernel
profile_cross_rho(): Profile fixed cross-lineage rho values
profile_cross_rho_ci(): Profile sensitivity interval for the cross-lineage rho
diagnose_kernel_separability(): Diagnose fixed-kernel separability before fitting
make_mesh() plot(<sdmTMBmesh>): Construct an SPDE mesh for gllvmTMB
add_utm_columns() get_crs(): Add UTM coordinates to a data frame
plot_anisotropy() plot_anisotropy2(): Plot anisotropy from a gllvmTMB model

Response families

Match response support and sampling process before adding covariance. Several exported compatibility constructors still stop explicitly rather than fitting an unsupported model; the response-family article states the current reader-facing boundary.

Beta() lognormal() gengamma() gamma_mix() lognormal_mix() nbinom2_mix() nbinom2() nbinom1() truncated_poisson() truncated_nbinom2() truncated_nbinom1() student() tweedie() censored_poisson() delta_gamma() delta_gamma_mix() delta_gengamma() delta_lognormal() delta_lognormal_mix() delta_truncated_nbinom2() delta_truncated_nbinom1() delta_poisson_link_gamma() delta_poisson_link_lognormal() betabinomial() delta_beta(): Additional families
ordinal_probit(): Ordinal-probit threshold family for the multivariate engine
multinomial(): Multinomial (baseline-category logit) family

Predictor-model families (modelled missing predictors)

Not response families: these tag the submodel for a categorical or ordered predictor in the modelled missing-predictor workflow (impute_model() / mi()), not a response distribution.

categorical(): Baseline-category softmax unordered categorical missing-predictor family
cumulative_logit(): Cumulative-logit ordered categorical missing-predictor family

Source-specific covariance keywords

Add a pedigree, phylogeny, spatial field, supplied kernel, or known-variance structure only when the study design requires it

animal_indep(): Independent per-trait animal-model random intercepts: animal_indep(0 + trait | id)
animal_dep(): Full unstructured animal-model trait covariance: animal_dep(0 + trait | id)
animal_latent(): Reduced-rank animal-model latent factors: animal_latent(id, d = K)
pedigree_to_A(): Pedigree to additive-genetic relatedness matrix A
pedigree_to_Ainv_sparse(): Sparse pedigree inverse-A via Henderson-Quaas (MCMCglmm-free)
phylo_indep(): Per-trait phylogenetic marginal variance: phylo_indep(0 + trait | species)
phylo_dep(): Full unstructured phylogenetic trait covariance: phylo_dep(0 + trait | species)
phylo_latent(): Reduced-rank phylogenetic latent factors: phylo_latent(species, d = K)
spatial_indep(): Per-trait spatial marginal field: spatial_indep(0 + trait | coords)
spatial_dep(): Full unstructured spatial trait covariance: spatial_dep(0 + trait | coords)
spatial_latent(): Reduced-rank spatial latent factors: spatial_latent(0 + trait | coords, d = K)
kernel_latent() kernel_indep() kernel_dep(): Generic dense-kernel covariance keywords
meta_V(): Known-V meta-analytic random effect (canonical name): meta_V(V = V)
block_V(): Build a block-diagonal sampling-variance matrix V

Report-ready extractors

Recover covariance, correlation, ordination, repeatability, communality, phylogenetic signal, and variance proportions

extract_Sigma(): Extract the implied trait covariance / correlation at one tier
extract_Gamma(): Extract a cross-lineage Gamma block
extract_coevolution_modules(): Extract cross-lineage coevolutionary modules
predict_cross_covariance(): Predict pair-specific cross-lineage covariance
extract_Sigma_table(): Extract a report-ready table of Sigma or correlation entries
compare_Sigma_table(): Compare fitted Sigma-table rows with a known truth matrix
extract_Omega(): Total trait covariance Omega summed across requested tiers
extract_communality(): Communality of each trait
extract_correlations(): Extract cross-trait correlations
extract_cross_correlations(): Cross-family correlations between a nominal (multinomial) trait and partners
extract_residual_cov() extract_residual_cor(): Implied trait covariance or correlation
extract_repeatability(): Per-trait repeatability with confidence intervals
extract_phylo_signal(): Phylogenetic-signal proportions per trait
extract_proportions(): Per-trait proportion-of-variance decomposition across all model components
extract_ordination(): Extract ordination scores and loadings from a fitted multivariate model
extract_loadings(): Extract the trait loading matrix
extract_lv_effects(): Predictor effects on latent-score means
extract_rotated_loadings_table(): Extract a tidy rotated loading table
extract_cutpoints(): Extract ordinal-probit cutpoints from a fitted gllvmTMB model

Methods and plots on fitted models

Standard summaries, predictions, simulations, confidence intervals, and current plotting entry point

print(<gllvmTMB_multi>) summary(<gllvmTMB_multi>) print(<summary.gllvmTMB_multi>) logLik(<gllvmTMB_multi>) nobs(<gllvmTMB_multi>): Methods on a fitted gllvmTMB model
print(<gllvmTMB_Sigma_phy_slope>): Print an augmented latent-slope Sigma extraction
confint(<gllvmTMB_multi>): Confidence intervals for a fitted gllvmTMB model
tidy(<gllvmTMB_multi>): Tidy a fitted gllvmTMB model
simulate(<gllvmTMB_multi>): Simulate new responses from a fitted gllvmTMB model
predict(<gllvmTMB_multi>): Predict from a fitted gllvmTMB model
predict_missing(): Predict the masked (missing) response cells of a gllvmTMB fit
imputed(): Extract fitted missing-predictor values from a gllvmTMB fit
phylo_signal_mi(): Phylogenetic-signal diagnostic for a modelled missing predictor
plot(<gllvmTMB_multi>): Plot a fitted multivariate gllvmTMB model
plot_correlations(): Plot pairwise trait correlations with intervals
plot_Sigma_table(): Plot report-ready Sigma table rows
plot_Sigma_heatmap(): Plot Sigma-table rows as a trait-by-trait heatmap
plot_Sigma_comparison(): Plot Sigma-table estimates against a known truth matrix

First-line diagnostics and uncertainty

Start here before interpreting covariance, ordination, or interval estimates

screen_gllvmTMB(): Screen candidate responses before fitting a gllvmTMB model
screen_control(): Control pre-fit response screening
screen_table(): Extract tables from a pre-fit screen
check_gllvmTMB(): Check convergence, Hessian, gradients, and interval readiness
gllvmTMB_diagnose(): Diagnose a fitted model and suggest next actions
predictive_check(): Fitted-model predictive checks for a multivariate gllvmTMB fit
diagnostic_table(): Extract report-ready tables from diagnostic objects
residuals(<gllvmTMB_multi>): Diagnostic residuals for a multivariate gllvmTMB fit
bootstrap_Sigma(): Bootstrap covariance, correlation, communality, and ICC summaries
confint_inspect(): Inspect profile confidence-interval shape for a fitted model
profile_phylo_signal(): Profile-likelihood curve(s) for per-trait phylogenetic signal
plot(<profile_derived>): Plot a profile-derived LR curve

Advanced validation utilities

Slower diagnostics for investigating a fitted model; none is a calibration certificate

sanity_multi(): Print a quick convergence and parameter sanity report
check_auto_residual(): Check whether link_residual = "auto" is coherent for this fit
gllvmTMB_check_consistency(): Run an advanced Laplace-consistency check
compare_dep_vs_two_psi(): Canonical likelihood-based cross-check for the paired phylogenetic decomposition
compare_indep_vs_two_psi(): Cheap diagonal cross-check for the paired phylogenetic decomposition (large T)
profile_targets(): List profile-ready confidence-interval targets for a fitted model
tmbprofile_wrapper(): Profile-likelihood CI for one parameter or linear combination

Simulation helpers

Generate small stacked-trait fixtures for examples and known-data-generating checks

simulate_site_trait(): Simulate a functional-biogeography GLLVM dataset (sites × species × traits)
simulate_unit_trait(): Simulate a generic stacked-trait GLLVM dataset (units x observations x traits)

Loadings (Lambda) and confirmatory factor analysis

Extract, rotate, and constrain the loading matrix

rotate_loadings(): Rotate the loadings of a fitted multivariate model
plot_rotated_loadings(): Plot a rotated loading matrix
compare_loadings(): Compare two loading matrices after Procrustes alignment
suggest_lambda_constraint(): Suggest a lambda_constraint matrix for a reduced-rank fit
suggest_lambda_constraints(): Compare several lambda_constraint suggestions
confirmatory_lambda(): Build a confirmatory lambda_constraint matrix from group membership
loading_ci(): Confidence intervals on individual entries of the loading matrix
loading_profile(): Profile-likelihood curve(s) for entries of the loading matrix
flag_unreliable_loadings(): Flag loadings whose CI overlaps a "biologically negligible" region
plot_loadings_confidence_eye(): Confidence Eye plot for per-species loading uncertainty
plot(<profile_loadings>): Plot a profile_loadings object (LR U-shape per entry)

Profile-likelihood confidence intervals

Target-specific likelihood-profile routes and fallbacks; inspect each target and do not assume universal calibration

profile_ci_phylo_signal(): Profile-likelihood CI for per-trait phylogenetic signal H^2

Experimental Julia bridge

Optional point-estimation bridge; use only after reading the explicit method limitations, and do not assume R/TMB feature parity

gllvm_julia_fit(): Fit a GLLVM with the Julia engine (GLLVM.jl bridge_fit).
gllvm_julia_setup(): Set up JuliaCall and load the GLLVM.jl engine (once per R session).
logLik(<gllvmTMB_julia>) print(<gllvmTMB_julia>) coef(<gllvmTMB_julia>) predict(<gllvmTMB_julia>) fitted(<gllvmTMB_julia>) residuals(<gllvmTMB_julia>) simulate(<gllvmTMB_julia>) confint(<gllvmTMB_julia>) summary(<gllvmTMB_julia>) print(<summary.gllvmTMB_julia>): Methods for Julia bridge fits

Deprecated covariance functions

Soft-deprecated as of 0.5.0; only the functions listed here are deprecated, and the broader animal, phylogenetic, spatial, and kernel APIs remain current. The scalar family (one shared variance) is now indep(…, common = TRUE); the *_unique diagonal companion folds into latent(); the standalone slope keywords are superseded by the reduced-rank latent / indep / dep slope forms. Existing calls keep working and warn once per session.

scalar(): One shared trait variance: scalar(0 + trait | g)
animal_scalar(): Single-shared-variance animal-model random effect: animal_scalar(id)
phylo_scalar(): Single-shared-variance phylogenetic random effect: phylo_scalar(species)
spatial_scalar(): One shared spatial-field variance across traits: spatial_scalar(0 + trait | coords)
kernel_scalar(): Dense-kernel one-shared-variance covariance: kernel_scalar()
unique_keyword unique deprecated: Trait-specific unique variance: unique(0 + trait | g)
animal_unique() deprecated: Per-trait independent animal-model random intercepts: animal_unique(id)
phylo_unique() deprecated: Per-trait independent phylogenetic random intercepts: phylo_unique(species)
spatial_unique() deprecated: Per-trait independent spatial random fields: spatial_unique(0 + trait | coords)
kernel_unique() deprecated: Deprecated alias: kernel_unique()
animal_slope(): Animal-model random slope on a continuous covariate
phylo_slope(): Phylogenetic random slope on a continuous covariate

Deprecated and compatibility aliases

Soft-deprecated spellings and gllvm-style accessors retained for migration; new code should prefer the canonical keywords and extract_*() functions

gr(): Generic group-with-known-covariance random effect (brms-style)
meta(): Known sampling-error term (brms-style, deprecated alias)
meta_known_V(): Known-V meta-analytic random effect (deprecated alias): meta_known_V(V = V)
phylo_rr(): Reduced-rank phylogenetic random effect (PGLLVM)
spde(): Spatial reduced-rank Gaussian random field per trait (Matérn, SPDE/GMRF)
gllvmTMB_wide(): Fit a GLLVM from a wide unit × trait matrix
extract_Sigma_B(): Between-site covariance matrix Sigma_B (backward-compat wrapper)
extract_Sigma_W(): Within-site covariance matrix Sigma_W (backward-compat wrapper)
extract_ICC_site(): Site / individual-level ICC per trait
extract_residual_split(): Separate OLRE residual variance from the distribution-specific latent residual
VP(): Variance partition by source
Beta() lognormal() gengamma() gamma_mix() lognormal_mix() nbinom2_mix() nbinom2() nbinom1() truncated_poisson() truncated_nbinom2() truncated_nbinom1() student() tweedie() censored_poisson() delta_gamma() delta_gamma_mix() delta_gengamma() delta_lognormal() delta_lognormal_mix() delta_truncated_nbinom2() delta_truncated_nbinom1() delta_poisson_link_gamma() delta_poisson_link_lognormal() betabinomial() delta_beta(): Additional families
getLV(): Extract latent-variable scores from a fitted multivariate model
getLoadings(): Extract the loading matrix from a fitted multivariate model
getResidualCov() getResidualCor(): Extract implied trait covariance or correlation
ordiplot(): Draw a two-axis ordination plot for a fitted multivariate model