Statistical tests for the neutrality hypothesis, such as Ewens' sampling theory, traditionally assume panmixia (random mating within a population).
Real-world populations often exhibit structure, consisting of numerous local populations with limited gene flow between them.
This subdivision can violate the assumptions of standard neutrality tests, potentially leading to inaccurate conclusions about evolutionary processes.
Purpose of the Study:
To propose a method for applying statistical tests of the neutrality hypothesis to structured populations with weak gene flow.
To investigate the statistical independence of tests applied to samples from individual local populations (demes).
To assess the potential for increased statistical power by combining results from multiple demes.
Main Methods:
Simulations were used to evaluate the statistical properties of applying Ewens' neutrality tests to subpopulations.
The study analyzed the correlations in allele frequencies among demes under low levels of gene flow.
A method for combining test results from different demes was developed and applied.
Main Results:
At low levels of gene flow, migration between demes acts similarly to mutation, introducing new alleles.
Correlations in allele frequencies among demes were found to be sufficiently small at low migration rates, allowing for statistically independent application of Ewens' tests within each deme.
Combining results from multiple demes increased the statistical power to detect deviations from neutrality.
Conclusions:
Population subdivision must be considered when testing the neutrality hypothesis, especially at low gene flow.
The proposed method of applying and combining Ewens' tests across demes provides a statistically sound approach for analyzing structured populations.
This approach enhances the ability to detect deviations from neutrality in subdivided populations, as demonstrated with salamander species data.