Simulation

Stabilizing Selection Out-of-Africa

Interactive simulation of GWAS effect sizes and allele frequency under selection and demographic parameters.

Selection induces a relationship between GWAS effect sizes for the selected trait and minor allele frequencies, neutral drift reduces this relationship, and bottlenecks accelerate drift. In this simulation, a trait is under stabilizing selection in the pre-OOA ancestral population, with the selection scenario changing after OOA. The top panel shows the relationship between squared GWAS effect sizes and allele frequency (both in the respective populations) over time.

The second panel shows the proportion of (European) GWAS effect size variance attributable to African versus European allele frequencies under different adaptation scenarios:
Divergence in Europe: Selection in post-OOA Europe is uncorrelated with ancient selection (local adaptation).
Divergence in Africa: Selection in post-OOA Africa is uncorrelated with ancient selection (local adaptation).
Africa Neutral: Neutrality in post-OOA Africa but selection in Europe persists.
Selection in Both: Selection in post-OOA Africa and post-OOA Europe persists.
Both Neutral: Neutrality in post-OOA Africa and in post-OOA Europe.
Europe Neutral: Selection in post-OOA Africa persists but selection in post-OOA Europe stops.

In general, we find that more of the European GWAS effect size variance is explained by African frequencies when there is neutrality in Europe or convergent selection in Europe/Africa because the European bottleneck introduces a substantial amount of random drift variance.

The bottom panel shows the effective population size (Ne) over time

Simulation inspired by results in Rossen et al. (2026).