What steps will reproduce the problem?
1. Run a discrete trait analysis
What is the expected output? What do you see instead?
Expect that the log file will contain samples from the posterior for geographic
transition rates. Instead, the log file contains the rate and indicator
variables used by BSSVS, and an additional step is required to transform this
into the rate posteriors.
It's probably dangerous to expose users to the BSSVS rate and indicator
variables, as this makes it possible for users to treat the BSSVS rate variable
as the actual transition rate - which it is not.
What version of the product are you using? On what operating system?
BEAST_CLASSIC v1.1.8 on GNu/Linux.
Please provide any additional information below.
Original issue reported on code.google.com by tgvaughan on 15 May 2015 at 12:41
Original issue reported on code.google.com by
tgvaughan
on 15 May 2015 at 12:41