SLIM scripts to simulate population divergence
We simulated four models with two populations that diverged from the same ancestral population and are identical in all population genetic parameters. We modified the simulation from Rousselle et al. and simulated 1500 coding sequences of 500bp length, separated by non-coding regions and re-scaled the population genetic parameter to an initial population size of 10k (mutation rate of 5.468928e-08, recombination rate of 6.604764e-06). The neutral mutations were initialized in ~1/3 of the coding sequence and deleterious mutations were initialized ~2/3, resembling the structure of a codon. Deleterious mutations were simulation with a selection coefficient s, a relative measure if fitness, drawn from the same reflected gamma distribution (mean of -2.5 and shape of 0.35) and a dominance coefficient of 0.1. All models start with an initial burn-in phase of 50000 generations to generate genetic diversity and a stable population. The models vary in either having immediate population size changes followed by constant population size (models A, B) or having exponential growth (models C, D). We further distinguished between a split directly after the burn-in phase (models B, D) or at a later timepoint, following the PSMC’ analysis interpretation (models A, C).