| Research into the genetic history of Homo sapiens benefits both from advances in the methods of typing more of the human genome and from computational methods that provide ways to make inferences about past population history using modern samples. I present here a method of estimating parameter values and comparing alternate models of population history that uses Bayesian statistical methods and is capable of combining data sets from various regions of the genome of various types, including Single Nucleotide Polymorphism, Short Tandem Repeat and sequence data from the mitochondria, the Y chromosome and the autosomes. This method is encoded in a software, named REJECTOR, which compares observed data to data produced via coalescent simulation, producing posterior distributions of the parameter values whose statistics most closely approach the observed. REJECTOR can analyze data using a variety of commonly-used summary statistics, and includes methods of calculating summary statistic values on linked autosomal systems such as the linked autosomal systems, termed SNPSTRs, that I typed and analyzed in this project. The rejection algorithm method implemented in REJECTOR allows for computationally efficient assessment of models of human genetics history, including inferences of such parameters as population size, divergence time, migration rate, and time of migration start. Complex and realistic models of population divergence followed by isolation and then migration can be calculated in REJECTOR. In this study I applied such a model to three African populations, estimating parameter values for population sizes and migration rates. The posterior distributions I thus obtained indicate a lower migration rate from foragers to their agriculturalist and pastoralist neighbors than vice versa.; I compared models of recent African origin to multiregional models of Acheulean expansion, using RETECTOR to estimate divergence time, population sizes and migration rates for Africa and Europe. I used a sequence data set composed of regions of the autosomes, the mitochondrial genome, X chromosome and Y chromosome. Population sizes and migration rates were estimated with precision, but there was less power to detect changes in time of divergence, though there is a tendency to favour more recent divergence times in the autosomes and Y chromosome. |
