Tracing very long-term kinship networks using SOCSIM

Abstract

While each individual has 10 billion ancestors a thousand years ago, these are not distinct and the actual number of distinct ancestors is much smaller. A female ('Mitochondrial Eve') and a male ancestor ('Y-chromosome Adam') of all humans certainly existed, possibly about 100,000 years ago, and a most recent common ancestor ('MRCA') of all humans existed much more recently. I use the SOCSIM micro simulation program to examine the patterns of descent over periods of several centuries of an initial population using as indicators: the proportion of these people without any living descendants by the end year of the analysis; the mean value and variability in the number of their distinct descendants; and the distribution of genetic contribution (ie the expected proportion of the DNA of individuals in the initial population found in the later population). About three-quarters of those born in the past have no descendant, mainly because they did not reach the age of reproduction. With the initial population sizes used here, about 4,000 people, after about 500 years the number of descendants of all of those who have any descendant becomes close to the size of the total number of descendants, confirming that even in this time-scale, a person is either the ancestor of everyone, or of no-one. However, the genetic contribution that those in the initial population make to later generations does not exhibit a similar tendency to uniformity. Issues such as the sensitivity of simulation results, which are inevitably based on smaller numbers than real human breeding group sizes, and the need to modify conventional measures of generational replacement to cases with multiple lines of descent are also considered. © 2004 Max-Planck-Gesellschaft.