Human Demographic Processes and Genetic Variation as Revealed by mtDNA Simulations

Humans' ability for rapid dispersal and adaptation has allowed us to colonize diverse geographic and climatic regions of the planet, creating a complex evolutionary history. This complexity can be understood, at least partially, by modeling the underlying demographic parameters in the evolutionary process. In this study, we analyze a model of human evolution in which population size, gene flow (GF), and time are varied. Specifically, we simulate mitochondrial DNA for 42 demographic scenarios, represented by 42 parameter combinations, to describe the initial dispersal of modern humans out of Africa. The analyses include three values for colonization size (CS; 1%, 10%, and 30% of the African population), seven values for rate of GF (10(-6)-0.5), and two values for time of colonization (50,000 and 100,000 years ago). We then estimate summary statistics for the simulated data sets to calculate the percent of explained variation by each parameter and to identify which parameter combinations generate distinct differences in genetic variation, that is, which demographic scenarios can be distinguished from each other. On the basis of these results, we make recommendations about which summary statistics to use according to the parameter of interest. Our results show that CS, GF, and their interaction have the largest effect on genetic variation under our model of human evolution. Comparison with empirical data suggests that 1% of the existing African mitochondrial genetic variation left and colonized the rest of the world (i.e., CS = 1%) and bidirectional GF continued at a level of ∼10 individuals per generation (i.e., GF = 10(-3)) after the initial colonization. Our study serves as a model to bridge the gap between the use of simulations for theoretical population genetics and empirical data analysis such as approximate bayesian computation approaches and is, thus, applicable to the study of molecular evolution in any organism.