I posted a request last week on evoldir for advice regarding modelling options
for exploring the impacts of post-glacial range expansion on the spatial
distribution of genetic diversity under a variety of dispersal models.  I
received numerous helpful replies which I have coalated below.

Here is the original post:

Dear colleagues,

I am hoping someone on the EvolDir site will be able to help me out.  I wish to
simulate the impacts of glaciation on the spatial distribution of genetic
diversity, but am uncertain what programs are availble for such a purpose. 
Specifially, I wish to simulate the impacts of a population bottleneck
(maintained for 'x' generations) and subsequent population expansion on the
spatial distribution of genetic diversity (Ho, R) for a multiple microsatellite
markers.  Supposedly, this is too complicated a process for EASYPOP.  Ideally
the simulation program would be able to model either a one dimensional stepping
stone model, and/or leptokurtic dispersal.

I have no leads at this time, so any advice is greatly appreciated.

Best wishes,

Dan HasselmanDan Hasselman
Marine Gene Probe Lab
Dalhousie University
Halifax, NS
B3J 4J1
dhasselm@dal.ca

Here are the responses, which I have edited to maintain anonymity:

1) You could try splatche (v. 1.1) (for SPatiaL And Temporal Coalescences in
Heterogenous Environment), developed by Excoeffier and colleagues

2) You can do this in SimCoal2 (Excoffier's group;
http://cmpg.unibe.ch/software/simcoal2/) or Dick Hudson's ms (upon which
SimCoal is partly written).

3) I think IBDSim is the simulation program you're looking for.
http://kimura.univ-montp2.fr/~rousset/IBDSim.html

4)One simulation program that will likely do what you want is called  SIMDIV
that was written by Jody Hey.  There is not a paper describing  the program,
but you can find documentation, and download the program,  from Jody Hey's
website (http://lifesci.rutgers.edu/~heylab/).  You  can specify many
parameters including: time since divergence, Ne of  populations before and
after divergence, growth rates of populations  before and after divergence,
etc... It is fairly comprehensive.

5) See Carr and Marshall. 2008. Intraspecific phylogeographic genomics from
multiple complete mtDNA genomes in Atlantic cod.  Genetics 180: 381-389.

6)I list a bunch of simulation programs in
http://simupop.sourceforge.net/Main/Features (people may have other
suggestions) and you may find one of them  (perhaps Bottleneck?) usable. If
none of them works for your application, please be assured that simuPOP can
certainly simulate the evolutionary process that you need, provided that you
know exactly what to simulate, and you or someone in your group can program in
Python. Your email did not describe your simulation in detail, but I guess you
can use a demographic model with bottleneck and migration
(http://simupop.sourceforge.net/manual/build/4_userGuide.html#demographic-changes),
and self-defined statistics
(http://simupop.sourceforge.net/manual/build/4_userGuide.html#statistics-calculation-operator-stat
and
http://simupop.sourceforge.net/manual/build/6_userGuide.html#output-statistics
because I am not sure what your Ho and R are). The landscape genetics example
in the simuPOP online cookbook
(http://simupop.sourceforge.net/cookbook/pmwiki.php/Cookbook/LandscapeGenetics)
might help. Please feel free to join the simuPOP mailinglist and discuss your
simulation there.

7)It sounds like what you're interested in is very similar to:
Jesus, F. F., J. Wilkins, and J. Wakeley, 2006. Expected coalescence times and
segregating sites in a model of glacial cycles. Genetics and Molecular Research
5:466-474. The software is available at John Wakeley's website.  Its in C and
you'll need to know some of the language in order to get it up and running.

8)Hickerson MJ has published a few papers on this in Evolution over the  last 3
years that might help.

9)If you have a Mac, you can try my own software "Marlin" which provides  a
graphical user interface for creating spatially explicit  simulations, and
which can only model colonizations with different  dispersal models. Otherwise,
I would recommend QuantiNemo, upon which  my program is based, but which is
also available for Windows and Linux.

dhasselm@dal.ca