Carol,
you have set the value called ‘micro-threshold’ to zero, this will lead to problems in the calculation of the likelihood during the run,
the default is 200 for the singlestep and the multistep model, it seems that a threshold of 10 works also for your data, but:
- I suggest to not use the singlestep mutation model, because it is __glacially slow__ for real runs and does not deliver ‘better’ estimates than the Brownian mutation model (see Beerli 2007).
- You need to adjust the upper prior boundary for Theta and M, see also attached parmfile
Peter