Important bug in Seq-Gen...

Andrew Rambaut andrew.rambaut at remove_thiszoo.ox.ac.uk
Wed Dec 16 07:50:00 EST 1998


Dear All,

Unfortunately we have discovered a serious problem with all
versions of Seq-Gen prior to 1.1 (the new one). The problem
only affects simulations done with gamma model of rate heterogeneity
(either continuous or discrete). The problem is that within a 
set of replicate runs, the sites are not being reallocated rates
at random causing each site to have the same rate throughout all
the replicates. This will cause a set of simulations to be more
similar to each other than they should. This would affect parametric
bootstrap type tests (though phylogenetic performance tests would
probably be less affected).

Please download the new version (1.1) from our web-site:
<http://evolve.zoo.ox.ac.uk>.

If this affects your analyses then my deepest apologies. It is
almost impossible to be certain that software does not have
bugs in it. This one was particularly nasty because the results
seemed fine (it was only when comparing to seperate runs with
identical input that the error was noticed). 

Please feel free to email me if you have any queries. 
Andrew

=====================================================================
  Andrew Rambaut,           EMAIL - Andrew.Rambaut at zoology.ox.ac.uk
  Zoology Department,         WWW - http://evolve.zoo.ox.ac.uk/
  University of Oxford,
  South Parks Road, Oxford, England
=====================================================================




More information about the Mol-evol mailing list