P
paleobuddy
ANALYTICS
Velocity0.0
Simulating Diversification Dynamics
paleobuddy can now simulate trait-dependent diversification, not just birth-death.
phylogeneticsdiversificationfossil-recordsimulationresearch-software
◆Current state
paleobuddy simulates diversification, fossil records and phylogenetic trees, with rates that can be arbitrary functions of time — its founding idea, implemented through rexp.var() generalizing exponential and Weibull draws. The 1.1.0 release adds state-dependent speciation and extinction simulation at roughly MuHiSSE generality, and lets simulations stop at a target number of extant species instead of conditioning on time.
◆Where it's heading
Releases track the maintainer's publications rather than a product cadence — 1.0.0 accompanied the MEE manuscript, 1.0.0.1 exists purely as a Zenodo citation anchor, and 1.1.0 is stated as going with a paper on SSE model accuracy for trees including fossil data. That framing sets the direction: the package grows whichever capability the next study needs to test. The stated SSE limits, no quantitative traits and no cladogenetic transitions, mark exactly where that boundary currently sits.
◆Prediction
Quantitative traits and cladogenetic transitions are named as missing, which makes them the obvious next targets, though on this history the timing will follow a paper rather than a roadmap.
◆Recent moves
- 1y ago
paleobuddy 1.1.0 adds SSE trait-dependent simulation
⚡ SPARKThe first release to add a new model class rather than refine the birth-death core: state-dependent speciation and extinction, roughly as general as MuHiSSE. Simulating to a target number of extant species also frees users from conditioning on time, which changes how comparable simulated datasets can be made.
View source ↗ - 3y ago
paleobuddy 1.0.0.1: Zenodo citation release
Explicitly a dummy release cut after Zenodo integration so the accepted manuscript had something to cite. No code change.
View source ↗ - 4y ago
paleobuddy 1.0.0: first release with time-varying rates
The initial release, built around rexp.var() generalizing exponential and Weibull sampling to accept any function of time as a rate, with bd.sim() applying that flexibility to diversification. That time-varying core is what the later SSE work extends.
View source ↗