Most existing methods for estimating effective population size over time have low accuracy for the recent past. In contrast, methods such as IBDNe that use segments of identity by descent (IBD) between individuals in a population-based sample can estimate recent effective population size as a function of time, with high accuracy for the past 3-200 generations.1
Many current-day American populations are admixtures of African, European and/or Native American ancestry, and the ancestry of an individual’s genetic material can be estimated at each point the genome. Here we extend IBDNe to utilize these local ancestry calls in order to estimate the ancestry-specific recent effective population sizes for admixed populations. This allows one to estimate the past effective sizes of the ancestral African, European, and Native American populations, as well as the founding sizes at the time of colonization and the post-admixture effective population sizes.
We demonstrate the efficacy of our method on simulated admixed data, and we apply it to admixed American populations from the 1000 Genomes Project and to African American samples from two US cities. We estimate that the pre-colonization ancestral population sizes were 1-3 orders of magnitude larger than the effective population sizes immediately after colonization. In most cases, population sizes rebounded quickly after colonization. We also estimate that prior to the colonization events, the growth rates of the Native American ancestral populations were 2-4% per generation, which are similar to the growth rates estimated for the European and African ancestral populations over the same time periods.