It could be useful for predictive modelling to allow users to seed the simulation with a set of individuals with pre-set traits. This is particularly relevant for neutral genetic loci, as researchers may be interested in setting an initial population structure after their empirical genetic data.