can ihs/ehh be applied on specific geographic region population?

Question

[evo_genomics]

Can ihs/ehh be applied on a specific geographic region population (i.e African Population) for selection estimation? Or can it be applied to mixture of all populations (Asian, African Europe America)?

Which of these two methods will be more logical?

Thanks

Answer 1

[Zev.Kronenberg]

Within population measures are more reasonable.

[GabrielMontenegro]

Yes, using homogeneous population is more sensible, however you should also take into account the number of samples in your dataset. Simulations have shown that these tests maintain power with sample sizes of ca. 40 chromosomes or above. If you have similar populations (say geographically similar as you asked) you can lump them together to increase power, however if they are different is better to use them separately.

[Zev.Kronenberg]

Two comments:

1. "similar" is a slippery slope.
2. You also lose power by lumping together populations that do not share the same selective sweeps.