Simulations to create a distribution of Y-DNA branch sizes of the Ashkenazi Jews

The data

The distribution of sizes of branches, ordered by size.

The models

These models describe the Jews arriving in Polish-Lithuanian Europe. To find a realistic model, the models should follow a set of observed characteristics. The most important are:
As a measure of the amount of Jews in recent times we used the values of 700.000 in 1620 (see Demographic history of Poland), 1.000.000 in 1770 (see History of Jews in Poland) and 9.000.000 in 1890 (see Historical Jewish population comparisons)
This simulation routines were previously used for the paper on 2000 years genetic variability in Flanders, Brabant and Limburg.

In the first diagram of each model one can see the size of the Ashkenazi as function of time. In this simulation 400 male line ancestors are arriving near the year 800. In green the amount of arriving male line ancestors is indicated. In the piechart the sizes of the present simulated haplogroups is indicated. They are ordered in size. The number is the order or the simulated male line ancestor. Most of the male line ancestors will have no male line descendants after many years.

Model 1

Model input: This is the (most simple) model to describe the Jews arriving in Polish-Lithuanian Europe (400 people arriving at the same time and constant grow rate in the period until 1620). In this model the distribution of the size of the Ashkenazim does not follow the observed distribution.

Simulation results: In this simulation we have descendants of 84 male line ancestors.

Comparison with observations: In the observations we have 8 male line ancestors that have 50% of the Ashkenazi descendants. The diversity of sizes of the different simulated branches is insufficient. This simulation does not describe the distribution sufficiently.


Model 2

Model input: In this model the the growth rate of the people is not stable, but has periods of no growth and stronger growth.

Simulation results: In this simulation we have descendants of 84 male line ancestors.

Comparison with the observations: In this model the distribution is a little bit better than in the previous model, but the fundamental difference between model and data is still present.

Model 3

Model input: In this model we use less ancestors to start with and a larger population growth.

Simulation results: The time of initial arrival of the first Ashkenazi Jews is later. The number of branches is less than the number of Ashkenazi branches and the small branches are lacking.

Comparison with the observations: This model does not have the amount of present branches. It does not describe the diversity of the amount and size of the branches.

Model 4

Model input: In this model the 400 people are arriving over a period of time in the area where the population growth is strong. This can be seen by the green dots in the first diagram.

Simulation results: In this simulation we have descendants of 75 male line ancestors.

Comparison with the observations: This model can describe the observed distribution of haplogroups. The numbers describe the arrival times of the groups. As one would expect the present largest groups are the groups that arrived first. The groups that arrived last are the relative small groups.

Model 5

Model input: In this model we assume that we have a difference in status in the community. This might be the case if sons of a elite group in the community have more chance to have descendants.

Simulation results: It is possible to create parameters to realize the distributions. The differences in population growth should be large and the continuation of the differences should last (in the simulated models) about 400 years.

Comparison with the observations: It is possible to make a model to fit the distribution, but the required parameters are unlikely realistic.

General Characteristics of the model

The model has the following characteristics: The model has a few parameters: In the model only males that reach the age of becoming a father are calculated in the model; for this reason we the term "potential father" in the model. This means that women and children that do not reach this age are not part of the model. To make the numbers comparable with e.g. population size we should use a ratio of population size to Y-DNA ancestors.

The simplest model

In the simplest model we have a population with a stable population growth over a long period of time. In the document of paper on 2000 years genetic variability in Flanders, Brabant and Limburg it was shown that, as a first approximation, one can use the value of 1.02 for a population growth in the Neolitic period and 1.06 in the period 8000-2000 ybp for the population of R1b in Europe. These two values give the following characteristics of number of parameters:

typical for period??8000-2000ybpshort periods or special groups
population growth in 30 years1.011.021.061.20
percentage of potential fathers that has a present day descendant0.010.020.060.20
mean number of generations of a person that lived once to a shared ancestor of a living person 6032103
ratio of population size to Y-DNA found at a moment in time4502257525
average number of SNP markers not shared by a living person (using 90years/SNP)201031
time to double the population size2010 years1050 years360 years114 years
time to 10-fold the population size6950 years3480 years1180 years380 years

Best model

The model that best describes the distribution of branchsizes is Model 4.

It gives an arrival of Ashkenazi Jews over a long period of time, and an effective population growth that is about 1.25 per generation (using an average generation of 30 years).