This simulation was modeled after a physical experiment conducted by Ozan Kiratli. Drosophila melanogaster populations from Maine, Pennsylvania, and Florida were sampled and crossed to create PA-ME and PA-FL hybrids in a 3:1 ratio. Preliminary findings suggest that the three original populations are equally differentiated from each other (Fig. 1). The populations were subsequently placed under summer and fall selection. Some hybrid populations were more responsive, adapting faster than the original populations. The next step, and where this experiment begins, is to simulate the empirical experiment. in order to determine the extent of the effect that migration and local adaptation have on population structure and allele frequency.

The first stage of the simulation is to create a null model influenced solely by genetic drift, without selection or migration. Three sub-populations would be allowed to diverge until they had reached similar levels of differentiation as those of the wild populations, measured with FST values. SLiM (Haller and Messer, 2019) was used to simulate three sub-populations of individuals for a range of generations. Other parameters such as recombination rate, mutation rate, and number of loci were chosen to be as close to the D. melanogaster genome as possible while simultaneously scaled down to keep a reasonable runtime.

At this stage in the simulation we can say for sure that there are factors other than genetic drift affecting the sub-populations. The next step in creating the simulations would be to apply fitness effects and create migration events to gain a better understanding of whether migration, local adaptation, or both are responsible for the differences among the populations.