Inferring Tree topologies

[MPR]

Hi-
I have a very simple question about the option to test all possible
tree topologies. I read the previous post on this topic which was
very helpful, but I am still unsure how to actually code this in
the .prs file. We have three populations and thus three possible
topologies we'd like to compare. We'd like to do this automatically
rather than re-running the program for each. Is this option 0 or
option 3? If it is option 3, what is meant by "choose a model
marker?" I apologize if this is in the documentation somewhere, but I
have been unable to figure it out. Also, in the previous thread on
this topic, you mention options 1 and 4, can you elaborate on what
those are?

Thanks!

[Joao Sollari Lopes]

Hi,

You're right, the user-guide doesn't do a very good job on explain the
different options when setting the topology. In the .prs files you
have the following legend:

>topology: 0 - uniform distribution;
1 - choose topology from a list;
2 - specify topology manually [e.g. ((Pop1,Pop2)Pop3)
-> 1 2 2 3];
3 - uniform distribution (and choose a Model marker);
4 - choose topology from a list (and choose a Model
marker);
5 - specify topology manually (and choose a Model
marker).


- If you want the the program to sample from an uniform distribution,
thus, letting the model vary automatically you should choose option 0.
- If you want to choose the topology from the list in the end of the
user-guide use option 2.
- If you want to set the topollogy manually (following the format
disclosed) you should choose option 3 followed by the reverse
branching sequence.
- Option 3, 4 and 5 do the same as 0, 1 and 2, respectively, but they
also let you choose a marker (a digit) to mark all the simulations.
The reason these options are available is in case you need to identify
different sets of simulations.

In your particular scenario you can just use option 0. This option
will produce simulations with a parameter (top) that will be either 1,
2 or 3 depending on the topology assumed for the particular
simulation. You can verify to what these topologies correspond in the
list of topologies in the annexes of the User-guide.

Hope that helps,
Joao

[MPR]

Thank you, this was very helpful. I had been looking only at the
legend for the "toypriors.prs" file and hadn't even realized there
were additional examples. I think I have it now, just to make sure,
let me run this by you:

My ultimate goal is to compare the three possible topologies given
three populations, make a choice, and then consider individual
parameters (like Ne1, etc.) within that particular topology.

To that end, I plan to run the program 3 times. Each time, I will
select a different topology. I will do this by selecting option 1 and
choosing either 1, 2, or 3 (the numbers corresponding to topologies
given in the annex file). After, I have done this, I can use
something like calmod to distinguish between the potential topologies
and so on.




On Mar 16, 5:45 am, Joao Sollari Lopes <j.sollari.lo...@gmail.com>
wrote:

[Joao Sollari Lopes]

Yep, that seems about right!

Or you can just use option 0 and let the the topology vary within the
same set of simulations. This way you don't need to consider different
simulated sets. Again, the parameter 'top' will save the information
on which topology was selected for the particular simulation.

Joao