Accounting for population structure and data quality in demographic inference with linkage disequilibrium methods

Area of Science:

• Population genetics
• Bioinformatics

Background:

• Demographic inference commonly uses panmictic models, which may not reflect complex natural population structures.
• Suboptimal genotyping data quality can impact the accuracy of population genetic analyses.

Purpose of the Study:

• To introduce two novel software tools, GONE2 and currentNe2, for estimating effective population size (N_e).
• To address limitations in existing methods by accommodating complex population structures and imperfect data.

Main Methods:

• Developed GONE2 for inferring recent N_e changes with genetic maps, and currentNe2 for contemporary N_e estimation without genetic maps.
• Utilized SNP data from single population samples.
• Incorporated methods to analyze population structure (F_ST, migration rate, subpopulation number) and handle data imperfections like genotyping errors and low sequencing depth.

Main Results:

• Validated GONE2 and currentNe2 using simulations and laboratory populations of *Drosophila melanogaster* across various demographic scenarios.
• Extended analyses to diverse species populations.
• Demonstrated that neglecting population subdivision often results in underestimation of effective population size (N_e).

Conclusions:

• The developed software tools provide robust estimates of effective population size (N_e) even with complex population structures and suboptimal data.
• Accounting for population subdivision is crucial for accurate demographic inference.