Stochastic Linear Ornstein-Uhlenbeck Comparative Hypotheses
An implementation of a phylogenetic comparative method. It can fit univariate among-species Ornstein-Uhlenbeck models of phenotypic trait evolution, where the trait evolves towards a primary optimum. The optimum can be modelled as a single parameter, as multiple discrete regimes on the phylogenetic tree, and/or with continuous covariates. See also Hansen (1997) doi:10.2307/2411186, Butler & King (2004) doi:10.1086/426002, Hansen et al. (2008) doi:10.1111/j.1558-5646.2008.00412.x.
- Removed 'background' vignette
- Added a
NEWS.md file to track changes to the package.
- Breaking changes: several variable names in function arguments and model outputs have been changed for clarity.
- Documentation improved.
- Model summary print changed
- Draft for long-form manual/vignette begun.
slouch 2.0.0 marks a near-complete rewrite of the entire package.
- Slouch now uses the phylogenetic tree format from package
- Model outputs are returned as a composite object, which consequentely is programmable.
- Implementation of explicit Brownian-motion models, including intercept-only models, regression with direct-effect covariates, random-effect (trend) covariates, regime-dependent trends and options to estimate
- Slouch now allows estimation of parameters using numerical optimization techniques.
- Computation in general is much faster, due to better memory management and use of Cholesky decomposition in regression coefficient estimation.
- Docstrings begun.
- Fixed several bugs when calculating the residual variance-covariance matrix, in particular for random-effect models with non-ultrametric trees.
- Fixed incorrect calculation of
tia for non-ultrametric trees.
- Fitch algorithm for ancestral state character estimation.
ouchtree phylogeny format.