bidirectional migration

[glaudelrio]

I am having trouble to set the migration band when considering bidirectional migration. How should I set it?

Furthermore, if I want to estimate migration direction, should I add A > B, and B > A in the same analysis as different migration bands, or should I add A > B in one run, and B > A in other run? How can I access which migration direction is more probable?

Thank you so much!

[igronau]

In general, you should be able to define two directional migration bands A-->B and B-->A in the same model and then see which one has a higher rate. However, you should note that there are often cases where the data cannot really distinguish between migration in one direction and migration in the other, and then the inference can be driven by somewhat arbitrary artifacts. This is mostly the case when you're trying to model migration between two sister populations that are of similar size. When modeling gene flow between populations that are not sisters in the phylogeny, the direction of migration is often inferred quite reliably, because it results in a discernible footprint on the local the local genealogies.

[glaudelrio]

Dear Dr. Gronau, thank you so much for your response! That's really helpful!
Best wishes