Dear Francisco, you can use methods such as Msvar (Girod et al. 2011 ; doi : 10.1534/genetics.110.121764), or more classical approaches are Approximate Bayesian Computation methods, if you have a nice sampling and different sound hypotheses to compare (for instance in european bats, see Razgour et al. 2013 ; doi : 10.1111/ele.12158).

Hi Francisco

There used to be a very simple tool called "Bottleneck" (Cornuet and Luikart, 1996), but not sure whether it is still available.

Pablo

Thank you for the suggestions. I'm using ABC, but because I would like to test it in a specific range, I'm not sure how I can do it. I was thinking that maybe splitting the population in those areas and consider different colonization time and population sizes for each subgroup can help.

To be honest, I never had to actually implement these methods myself, and mostly read about these. The demographic parameters you are talking about (population size, time of divergence) are precisely the ones estimated in ABC procedures in population genetics - or can be derived from the estimated parameters - but I do not understand what you mean by "test it in a specific range". If you mean a geographical range, I think there is no such models implemented in this kind of computations; at most you can provide various relationships between your sampled regions as prior in different scenarios (see Fig. S6 of the aforementioned paper by Razgour, for instance).

Allright Francisco

So you have decided to use ABC. Now, you should start with a simple model. If your goal is to know whether a population is actually under expansion, you don't need to split the population into different colonization areas at different times. Keep it simple. Consider just one population of size Ne which is growing in size at a certain rate or that is has a stable size. Indeed, for coalescence issues you might consider another (ghost) population your population is derived from at a certain time in the past (but note that you do not need to sample this ghost population).

On the other hand, according to Haasl et al (2014), microsatellites are much more effective if you summary statistics is the SFS.

Good luck

Pablo

Dear