problems running popABC for real data

[Luca Antoni]

Dear Dr. Lopes,

I'm currently trying to use popABC on a real dataset which includes 4 populations genotyped at 17 microsatellite markers. Unfortunately the runs stop when simulating the trees, after creating a certain number of trees, which varies from time to time. Maybe sometime it hits values outside the range or something like that. Not sure. Anyway I tried to change the priors with no luck. Every run stops at a different point but I never managed to hit even 100 trees. Any idea on what could cause this? The initial setting parameters I used were (but I tried smaller ranges as well):

Effective population size (for all Ne and Neanc):

-uniform(10,1000000)

Time events:

-uniform(10,1000000)

Migration rates(for all current and ancestral):

-uniform(1e-05,0.9)

Mutation rates:

-lognormal(-3.5,0.5,0,2)

Recombination rates:

-No recombination

Hope you can help me out with this.

Thanks in advance,
Luca Antoni