MASTER version 1.8

This release provides a new simulation type, PopulationFunctionFromMaster, which provides a way to use MASTER in coalescent analyses. These population functions aren't yet parametric from BEAST's point of view so the possibilities are currently limited as far as inference goes, but this already makes it easy to draw coalescent trees from MASTER-simulated trajectories using BEAST's RandomTree class.

As always, MASTER-1.8.0.zip is the stand-alone copy, while MASTER.package.v1.8.0.zip is the BEAST package that requires a working BEAST installation.

Bug fixes

This release fixes two post-processing and end-conditioning bugs introduced during the v1.4 -> v1.5 transition. Anything involving post-processing of InheritanceEnsembles will have been affected by the first bug, while the second caused problems for LeafCountEndConditions with includeExtant="true" and individual populations specified.

Please update ASAP!

The stand-alone release is MASTER-1.7.1.zip, while MASTER.package.v1.7.1.zip is for use in conjunction with an existing installation of BEAST 2 and can be installed directly from BEAUti's package manager.

MASTER version 1.7

This release provides an improved (and now documented) BeastTreeFromMaster simulation type. This allows MASTER to be used to simulate trees for use in BEAST 2. Possible applications include initialisation of the tree state in a phylogenetic MCMC chain or simulation of a tree that can be used by BEAST's SimulatedAlignment class to produce a simulated data set.

MASTER version 1.6

This release provides a couple of new features to do with conditioning and sampling.

1. Inheritance simulations (Trajectories and Ensembles) can now include a LeafCountPostSimCondition that forces the simulation to be repeated until the generated inheritance graph has the desired number of leaves. This can be an exact match, or a lower or upper bound, and the counting can be restricted to leaves associated with particular populations. See examples/SIRSampleTreeConditioned.xml for an example.
2. The LineageSampler has undergone some modifications. It is now possible to restrict sampling to leaves associated with particular populations. Heterochronous sampling is also now supported.

The stand-alone release is MASTER-1.6.0.zip, while MASTER.package.v1.6.0.zip is for use in conjunction with an existing installation of beast 2.

As always, please raise an issue if you notice a problem (either with the software or the documentation).

Bug fixes

This is a patch release that fixes two severe bugs introduced during the 1.4 -> 1.5 transition:

1. The LineageSampler post-processor was incorrectly labeled, breaking any script that used this.
2. The NewickOutput and NexusOutput outputs did not listen to any of their attributes besides the output file name. Most importantly, this meant that trees could not be generated from backward-time processes like the coalescent.

Additionally, the examples/ directory now contains an example use of LineageSampler.

(Note the binary for the standalone release is MASTER-1.5.2.zip. The other file is a package add-on for BEAST 2, and needs a working BEAST installation to run. Visit www.beast2.org for more details.)

MASTER version 1.5.1

This release provides a number of improvements, both visible and invisible:

1. Code-base has been streamlined (~3000 fewer lines of code)
2. Input file syntax has been improved. (It is now unnecessary to use InheritanceModel and InheritanceReaction when specifying models and reactions for InheritanceTrajectory and InheritanceEnsemble simulations. See examples for details.)
3. Piecewise constant variation in reaction rates is now possible. Specify this by setting rate="RATE1:TIME1, RATE2:TIME2, ..." in the <reaction> element. If TIME1>0 the reaction rate will be zero until the simulation time reaches TIME1.

Beware that as with any large update, there may be bugs. Please raise an issue if you find that something is broken.

Enjoy!

Bug fix

Fixed a node annotation bug in the LineageSampler post-processor. Note that this release has been repackaged in a way that will hopefully neutralize the "runs but does nothing" problem that some OS X users have been experiencing.

Range elements for complex model construction.

New element has been added which allows for large numbers of reactions with equivalent rates to be added. Documentation coming soon.

This release also includes a number of bug fixes to the graph post-processing code.

New leaf count inheritance simulation end condition.

A new end condition has been added (and documented) which allows users to condition on the number of leaves in a tree.

New inheritance graph post-processing interface.

New interface is documented on the wiki. Will include a dedicated tutorial at some point.