Ashkenazi Jewish Genetics

Most Jews are either Ashkenazi Jews or Sephardic Jews, and most American Jews are Ashkenazi Jews, typically with origins someplace in Eastern Europe immediately prior to immigration to the United States.

A new paper looks at the genomes of Ashkenazi Jews today in order to determine their population history using computerized tools and a greatly expanded set of Jewish and non-Jewish reference genomes that has accumulated over the last few years. 

This is necessary because the historical record provides very little meaningful or reliable insight into the process by which the Jewish diaspora resulted in migrations of Jews from the Middle East to Europe. The gap in the historical record is particularly acute in the several century long gap period after the composition of the Babylonian Talmud and the fall of the Roman Empire, and before the pogroms directed at Jews in Europe which roughly coincided with the Crusades. Yet, this gap period was apparently the period of Ashkenazi Jewish ethnogenesis and the source of the population bottleneck that is a defining feature of Ashkenazi Jewish population genetics.

The quaint term for this time period in historical circles is the "dark ages" and the dearth of historical sources from that era deserve that name, even though we do know a fair amount about that time period now by a variety of means. This time period overlaps with the expansion of the Islamic empire, the Slavic expansion in Eastern Europe, the "migration period" of (mostly) Germanic tribes in Europe, and Anglo-Saxon and Viking raids into and migration of England (which is arguably part of the migration period).

The genetic analysis is not straightforward because (1) there are considerable similarities between the genetic profiles of the suspected source populations, (2) there is little ancient DNA to bear on the question very directly, and (3) modern populations are not great proxies of historical populations at the time that the relevant stages of the migrations happened, particularly in the Levant and Slavic Europe. Indeed, the latest results come with considerable acknowledged uncertainty, although the broad outlines of the analysis are probably correct.

The West Hunter blog has a summary of the findings of this open access paper that fairly states the conclusions of the paper at least as well as I could describe them myself:

When they analyze the origins of the European component of Ashkenazi ancestry, they conclude that most is southern – probably Italian, but that smaller amounts originated from (probably) Western Europe and (more certainly) Eastern Europe: and in that temporal order. They conclude that the Italian admixture slightly predated a late medieval founder event. Different methods came up with somewhat different estimates for the total amount of European ancestry: the local ancestry inference (LAI) approach came up with 53% European, while the GLOBETROTTER analysis came up with an estimate of 67% European ancestry (after calibration by simulations). In their best guess, they split the difference and go for 60% European.

To sum up, their model is that a population from the Levant mixed with Italians, and shortly thereafter moved to the Rhineland (the founding bottleneck), perhaps mixing to some degree with the local Europeans there, and certainly mixing some with Slavic types when they moved to the Polish-Lithuanian Commonwealth.

How do their conclusions differ from those in the last report? Previously they were thinking that the bottleneck was around 1350, a product of the Black Death and savage persecution – now they’re talking the original settlement in the Rhineland. Previously they had a somewhat lower estimate of European ancestry (~48%, now 60%). I thought these two conclusions likely a couple of years ago. 

The big new point, important if correct, is that the admixture with Italians is relatively recent – too recent to have happened back in Roman times. In their model, this main admixture event is 25-55 generations ago, while the founding bottleneck is 25-35 generations ago. It’s not impossible that the admixture happened at the same time as the founding.

About 15%-25% of Ashkenazi Jewish ancestry, which is a quarter to a almost half or so of the total European ancestry, is Eastern European.  I've also seen other sources conclude that the paternal line is more strongly Levantine than Jewish maternal ancestry.

As Razib Khan has noted in comments to his post on the same paper, the Black Plague hypothesis was always suspect as a source of the population bottleneck that is clearly apparent in Ashkenazi Jewish population genetics, because it didn't kill a large enough percentage of the population to be likely to have caused the observed bottleneck effects.

This is also as good a post as any to observe that all Jewish populations in the world from Ashkenazi and Sephardic Jews to Jews in Yemen, India and elsewhere, are significantly admixed with local populations. Endogamy in the founding populations of Jews in the Jewish diaspora has almost always been modest in the founding period, even though these barriers to admixture ossified with time. There are probably far fewer "pure blooded" Jews with exclusively Levantine origins and descended from Jews in their Iron Age kingdoms (I'm not convinced that there are any), than there are "pure blooded" Native Americans who lack European or African admixture, for example.