SmoothTree

This library implements:

• Larget's conditional clade distribution (CCD) [@larget2013]. This is a distribution over cladograms (labeled phylogenetic trees without meaningful branch lengths) derived from a collection of observed trees X under the assumption of conditional independence of disjoint subtrees. It is the maximum entropy distribution over tree topologies subject to the constraint of matching observed marginal split frequencies in X (see @szollosi2013).
• Aldous' beta-splitting distribution over cladograms [@aldous1996].
• Arbitrary Markov branching models (MBMs), of which the above are special cases. Crucially, the library implements a sparsely represented MBM that is obtained as the posterior of a beta-splitting (Dirichlet) prior distribution and an observed collection of splits (represented by a CCD). This can be seen as a smoothed CCD, which spans the whole tree space (the CCD does not generally cover the whole tree space)
• Efficient simulation routines for the multi-species coalescent (MSC) model.
• A likelihood-free expectation propagation algorithm [@barthelme2014] for approximate Bayesian inference of species trees from gene tree distributions.

Examples

Coming soon.

References

[@larget2013] Larget, Bret. "The estimation of tree posterior probabilities using conditional clade probability distributions." Systematic biology 62.4 (2013): 501-511.

[@aldous1996] Aldous, David. "Probability distributions on cladograms." Random discrete structures. Springer, New York, NY, 1996. 1-18.

[@barthelme2014] Barthelmé, Simon, and Nicolas Chopin. "Expectation propagation for likelihood-free inference." Journal of the American Statistical Association 109.505 (2014): 315-333.