Changelog

heritable 0.2.0

• Added a glmmTMB model backend for Gaussian (identity-link) models #19. All broad- and narrow-sense estimators (H2_*() / h2_*(), including Standard, Cullis, Oakey, Piepho, and Delta) are supported and reproduce the lme4 backend exactly. Non-Gaussian glmmTMB models fail with an informative error.
• Added reliability (r², the coefficient of determination) as an additional metric #24. h2_Reliability() / H2_Reliability() return the overall (mean) reliability r̄², h2_Reliability_by_genotype() / H2_Reliability_by_genotype() return the per-genotype values, and H2_Reliability_parameters() computes them from variance components. The measure is also available via h2(method = "Reliability") and H2(method = "Reliability") (not part of the default method set).
• h2() and H2() now fail with an informative error for non-Gaussian lme4 models (e.g. glmer()) instead of a cryptic method-dispatch error #48.
• Added narrow-sense heritability estimation through h2() and the h2_*() estimator functions for models with a known additive genetic covariance structure.
• Added a consistent interface across broad-sense (H2()), narrow-sense (h2()), and their Standard, Cullis, Oakey, Piepho, and Delta estimators.
• Added support for more complex asreml and lme4 random-effect structures, including known genomic relationship matrices and structured residuals.
• Added richer heritable objects that retain the fitted model and calculation metadata for downstream methods.

heritable 0.1.1

• Update Maintainer to Dr. Emi Tanaka
• Refactored asreml integration #37, #32
  • Added var_comp.asreml() for extracting and assembling variance components for asreml model objects
• Refactored H2() for asremlmodels so that it now takes arguments like marginal, stratification
  • Improved handling of vm() specification
• Refactored H2_standard()
• Compute confidence intervals for heritable objects #30