Hi,
The files you have selected must be from independent runs of BEAST
from the same XML file. If the log files differ an error will occur
stating that the number of columns in the first file does not match that
of the second file.
My questions are:
A. Can I combine log files from multiple runs with different MCMC chain length (other parameters are exactly the same)
B. Can I combine log files from multiple runs with
some having PS/SS marginal likelihood estimation (MLE) but others not?
C. Can I combine log files from multiple runs with different initial values of some priors or different tuning values of some operators?
I have tried combining the log files from multiple runs which differ in the ways I mentioned above, and it seemed to work. But because of the statement in the tutorial, I would like to make sure that I am not doing something wrong.
Cheers,
Michael