search

CompEvol / beast2

Bayesian Evolutionary Analysis by Sampling Trees
www.beast2.org
GNU Lesser General Public License v2.1
240 stars 84 forks source link

ClusterTree has no attribute to allow rescaling the tree to appropriate units #179

Closed alexeid closed 10 years ago

alexeid commented 10 years ago

Often you want to start with a reasonable tree, and so a UPGMA tree is a good starting place. However UPGMA trees are in units of substitutions. If your MCMC is on a time-tree in units of years or millions of years, then the scale could be very different and calculation of the likelihood could have numerical issues starting with a tree in units of substitutions. So an attribute should be added to ClusterTree taking a clockRate, which all divergence times are divided by to convert into times.

alexeid commented 10 years ago

This issue is fixed with https://github.com/CompEvol/beast2/commit/f1ea5e2c1fb9b8552e3eef975d171bc3f85543f0

tgvaughan commented 10 years ago

Hi Alexei, you can close these issues automatically by saying something like "closes issue #179" in the commit message itself (although I often forget).