This model is based on the published model (GitHub), with few simplifications and changes to ensure straight-forward and quick usage.
Background
The t haplotype is a genetic element in house mice, a meiotic driver, which means that it manipulates sperm such that it is transmitted at increased chance to progeny (roughly 90 rather than 50 percent, known as “drive”).
However, carrying t homozygously (notated as t/t), i.e. on both chromosome pairs, is associated with lethality or male sterility. Furthermore, +/t (carrying one t copy) males produce sperm that, while 90% t and 10% +, is not very competitive against +/+ male sperm. This means that in matings between one female and multiple males, +/t males sire fewer offspring than their +/+ competitors. In summary, t is fittest when it is at low frequency in a population and when few females mate with more than one male. In the opposite case, t is at risk of extinction. How does evolution shape t further under these circumstances? We found that t-carriers are more dispersive, meaning they are more likely to emigrate from their populations. We built simulation models to understand whether t are more likely to evolve increased dipsersal given their traits. The full model can be found on GitHub and the associated paper on bioRxiv. A condensed version of the model for playing around with it can be found below.
The online model
The model starts with 1000 mice on random patches, which differ in density, with random dispersal propensities. After 1000 ticks (time intervals with mating and dispersing etc), half of the mice become +/t and from then on the two genotypes compete. There is mutation and inheritance, meaning whatever dispersal propensity leads to higher fitness is selected and rises in frequency, which can be monitored with the graphs in the simulation. Note that t is plotted as 0 until t are in the population.
Press setup, then go to run the simulation. Play with variables below the plots, but best ignore others for just a quick look at this model.
Default settings that were adapted for quicker simulation times that do not depend on averaging days and days of computing time to see results:
World size more than halfed (browser version is much slower and hence fewer mice should be simulated)
tcome in immediately to make it more fun to watch.
max ticks are 10000, which usually gives results, but model runs beyond that.
“full mutations” are enabled, i.e. there is an additional chance to evolve a phenotype independent of the current phenotype, e.g. go from never dispersing to always dispersing, to more quickly / more often explore the range of possible phenotypes.
Only D0 (intercept, i.e. dispersal propensity) is plotted and loci are set to 1. Loci = 2 takes too long to evolve for this presentation.
P_inc refers to probability of incremental mutation (small ± change from former dispersal propensity) and P_full is the probability of a mutation anywhere from 0 to 1 (0 to 100%) propensiy