Genetics & the Jews

By Razib Khan | June 6, 2010 1:20 am

The 2,000 year dance between the Jewish people and Western civilization has spawned many questions of scholarly interest. A relatively minor point, though not trivial, has been the issue of the biological relatedness of the Jewish people, and their relatedness to the nations among whom they were resident. This particular point became more starkly relevant with a scientific understanding of human genealogy and genetic relationship in the 18th and especially 19th centuries, but its root can be traced back to antiquity. Jews are not simply a set of individuals who espouse a belief in the God of the Jews, or hold to the laws of the God of the Jews. Rather, one aspect of Jewish identity is its collective component whereby the adherents of the Jewish religion also conceive of themselves as a particular nation or tribe, and therefore bound together by a chain of biological descent. Ergo, the traditional assertion that one is a Jew if one’s mother is a Jew.

Of course these issues can not be understood except in light of a complex historically contingent sequence of events. Our understanding of what it means to be Jewish today, or the understanding of Jews themselves as to their own identity, is the outcome of a long process where self-identified Jews interacted with the broader milieu, as well as evolving in situ. In other words, the Jewish people and the seeds of the Jewish Diaspora were shaped by developments within and without the Jewish culture, and these developments left an impact on the genes of the Jewish people. Contemporary groups outside the “Jewish mainstream,” such as the Beta Israel, Bene Israel and the Karaites, but with an acknowledged connection to Judaism, are windows into other faces of being Jewish besides that of Rabbinical Judaism.

And yet it is descents of the adherents of Rabbinical Judaism, the Judaism of the Pharisees, which we think of when we think of Jews (even the non-Orthodox traditions emerged out of a cultural milieu where Orthodox Judaism was normative). The vast majority of the Jews of the world trace their lineage back to the groups who organized their lives around not just the Bible, but also the Talmud, and subsequently commentaries and rulings by rabbis who were trained in the Talmud. Today these Jews fall into three broad groups, the Ashkenazim, Sephardim and Mizrahim. The Ashkenazim are rather easy to define, as they are the Jews of Central Europe who have been so prominent over the past few centuries. Though it seems likely that in the first millennium their ancestors were to be found along the Rhine, more recently their center of gravity has been in Central & Eastern Europe, in particular Poland and Lithuania. The Sephardim were originally the Jews of Spain, but after their expulsion in 1492 they settled in the Ottoman Empire, and to a lesser extent in other regions of Europe such as the Netherlands. A major confounding issue with the modern Sephardim is that in the Ottoman lands they encountered and interacted with preexistent Jewish communities, who often maintained a distinctive identity subsequent to the influx of the Sephardim. Though in most cases, such as in Morocco and Syria, the Sephardim became culturally dominant and assimilated the indigenous Jewish community into their identity (though they often abandoned Ladino, the language they brought from Spain, for the local lingua franca), in other cases two distinct Jewish communities were coexistent down down to the modern era (e.g. Greece). Finally, the Mizrahim are Jews of the East or Oriental Jews, those Jews whose ancestors hail from Muslim lands where the Sephardim were never a presence. To a great extent the Mizrahim identity is a recent catchall constructed to identify a real dividing line between those groups which are the products of the Sephardic-indigenous synthesis, such as the Moroccan Jews, and those which are not, such as the Yemeni Jews. Often all non-Ashkenazi Jews are referred to as Sephardic because of a common religious liturgy which binds them.

But naturally it gets more complicated than this. Between the rise of Islam and Christianity as the dominant religious civilizations in which Jews were embedded and the Enlightenment Rabbinical Judaism had established a modus vivendi. Jews were a corporate entity, a minority subordinate to the majority, whose relationship with the majority was mediated through eminent individuals who spoke for and had power over the community. Though often fraught in the execution in the abstract the position of Jews within pre-modern political units was not controversial; Jews were subjects with obligations, often a useful minority for potentates. They were not citizens with rights and responsibilities. Over the past few centuries that has obviously changed. The French Revolution and the emergence of the idea of a nation-state where all citizens have equal rights and responsibilities before the law, along with a scientific concept of race, complicated the Jewish relationship with the societies in which they were resident, particularly in Europe (though pan-Turk and pan-Arab nationalism were analogous and resulted in similar problems of identity). Despite phenomena such the Spanish fixation on “cleanliness of blood”, as well the Jews self-conception as the descendants of Israel, it was in the 19th century that the idea of a Jewish race with very specific and determined biological qualities which were heritable came to the fore. The Nazi total extermination program stood in contrast to previous assaults on the existence of Jewish community, where conversion to Christianity, and assimilation more broadly, were plausible goals. The Nazis aimed to eliminate not just the culture of the Jews, but their very biological existence. Ironically assimilated European Jews themselves internalized this sense of their racial/national distinctiveness, evident even in those with no religious aspect of Jewish identity at all such as Sigmund Freud. This explains the secular nature of the original Zionist project, whose aim was to create a national homeland for the Jews as a people, and so normalize them as a nation among nations, rather than being among the nations (this was a project which religious and assimilationist Jews initially opposed).

With the Holocaust, and the post-World War II rejection of racial nationhood, the often pseudo-scientific practice of measuring and categorizing people according to skull metrics, and more legitimately blood groups, fell into disrepute. Some scholars began to reconfigure the Jews not as a biological descent group, but as a religious ideology or confession which eventually became an ethnic identity. The most extreme proponents of the cultural model presumed that Jewish groups emerged through cultural diffusion and religious proselytization. The Jews of Poland were Poles who adopted Rabbinical Judaism. The Jews of Morocco were Arabs or Berbers who adopted Rabbinical Judaism. And so forth. In other words this school transformed Jewishness into what the German Reform movement had attempted, making of Jews just another religious confession with no ethnic connotations (and therefore entailing a reinterpretation of some aspects of Chosen Peoplehood).

But the pendulum has swung back, in part thanks to the rise of genetic science, and in part broader currents in the Jewish world. In regards to the second I will note that the American Reform movement has pulled back from its more aggressive accommodations with the sensibilities of gentiles. Of particular relevance for the topic at hand, Reform Judaism has reversed its rejections of the idea of Jewish nationhood. I suspect this is in large part because American Jews, and Jews in Western nations more generally, feel less need to prove that they belong by aligning themselves self-consciously to mainstream conceptions of religious identity as anti-Semitism has declined.

And now we come to genetics. The genetics of Jews are a large set of related fields. Much of it is motivated by medical considerations, in particular “Jewish diseases” such as Tay-Sachs. Though the ultimate aim of much research is to clarify population stratification in association studies, over the past few years there has been a great deal of light shed on the possible origins of and the relationships of Jews to each other and other populations. Originally the focus was on uniparental lineages, male and female markers passed through the Y chromosome and mtDNA respectively. The general results of these were that both the extreme scenarios of total replacement and pure cultural diffusion are false. On the one hand Jews across the world by and large share unexpected genetic affinity which one would not predict from geography, but only from their common religious-ethnic identity as Jews. But Jews also cluster geographically in a way that is reminiscent of the gentile populations among whom they have settled, suggesting either independent evolution after an initial separation and/or admixture with the local populations.

jewpc2One of the most popular posts on this weblog focuses on the differences between Ashkenazi Jews and gentiles, in particular peoples of European descent. The figure to the left illustrates that white Americans who are gentile or Jewish are rather easy to distinguish genetically from each other. That Jews exhibit a particularly distinctive genetic signature may not be all that surprising, considering that medical geneticists have long known that there are diseases which are biologically rooted and heavily overrepresented among this population. Distinctive traits imply distinctive genes. And the demographic history of the Jewish people as attested to in the literary records can be fitted rather easily within the framework of many of the results coming out of the genetic studies.

But what about the issues I mooted above in regards to the divisions among the Diasporic Jewish community? A new paper in The American Journal of Human Genetics takes a stab at attempting establish a set of relations between different Jewish communities, as well as other populations which they may have admixed with. Abraham’s Children in the Genome Era: Major Jewish Diaspora Populations Comprise Distinct Genetic Clusters with Shared Middle Eastern Ancestry:

For more than a century, Jews and non-Jews alike have tried to define the relatedness of contemporary Jewish people. Previous genetic studies of blood group and serum markers suggested that Jewish groups had Middle Eastern origin with greater genetic similarity between paired Jewish populations. However, these and successor studies of monoallelic Y chromosomal and mitochondrial genetic markers did not resolve the issues of within and between-group Jewish genetic identity. Here, genome-wide analysis of seven Jewish groups (Iranian, Iraqi, Syrian, Italian, Turkish, Greek, and Ashkenazi) and comparison with non-Jewish groups demonstrated distinctive Jewish population clusters, each with shared Middle Eastern ancestry, proximity to contemporary Middle Eastern populations, and variable degrees of European and North African admixture. Two major groups were identified by principal component, phylogenetic, and identity by descent (IBD) analysis: Middle Eastern Jews and European/Syrian Jews. The IBD segment sharing and the proximity of European Jews to each other and to southern European populations suggested similar origins for European Jewry and refuted large-scale genetic contributions of Central and Eastern European and Slavic populations to the formation of Ashkenazi Jewry. Rapid decay of IBD in Ashkenazi Jewish genomes was consistent with a severe bottleneck followed by large expansion, such as occurred with the so-called demographic miracle of population expansion from 50,000 people at the beginning of the 15th century to 5,000,000 people at the beginning of the 19th century. Thus, this study demonstrates that European/Syrian and Middle Eastern Jews represent a series of geographical isolates or clusters woven together by shared IBD genetic threads.

The major limitation of this study that I can see is that two very numerous and interesting groups of non-Ashkenazi Jews, Moroccan and Yemenis, were not included. Yemenis in particular are of interest because there is some historical reference to kings of Yemen who adhered to the Jewish religion, and so implicitly may have brought over substantial numbers of South Arabians to the religion. The studies I have seen about the genetics of Yemeni Jews are mixed in regards to whether they exhibit more affinity with other Jews, or with non-Jewish Yemenis. But set next to the treasure trove of results that’s a minor complaint. Quick review of the groups in the study:

Ashkenazi – easy, Jews of Central Europe

Iraqi Jews – Mizrahi, presumably Jews who descend from the Babylonian community which dates back to the First Exile

Iranian Jews – Mizrahi, should be derived from the Babylonian Jewish community. For most of history after the conquest of Babylon by the Persians Mesopotamia and the Iranian heartland were integrated into one political unit. The the division between Mesopotamia and Iran was fixed after the Ottomans managed to hold what became Iraq against the attempts by the Safavid dynasty of Persia to reclaim it in the 16th century.

Syrian Jews – Sephardic, but a compound of ancient Levantine Jews who date back to Roman antiquity and post-1492 Sephardim. The native Syrian liturgical tradition apparently persisted down into the modern period before its recent extinction

Turkish Jews – Sephardic, but a compound of Anatolian Jews who date back to Roman antiquity and post-1492 Sephardim

Greek Jews – Mostly Sephardic, a compound of Greek Jews who date back to Roman antiquity and post-1492 Sephardim (note that Greece was part of the Ottoman Empire until the 19th century)

Italian Jews – I believe this study classes them as Sephardic, but the origin and nature of this group is ambiguous. The Jewish community of Italy may date back to Roman antiquity, and so lay outside of the Ashkenazi-Sephardic dichotomy, but operationally it has been influenced by the pan-Mediterranean peregrinations of the Sephardic Diaspora

In fact the last point, that different Jewish communities have interacted and influenced each other, is a general truth. Persecutions of Jews during the medieval period as far away as Germany and Spain resulted in infusions of new migrants into the Jewish community of Kerala in South India as an extreme case. Just as there was an Islamic world which stretched from the Atlantic to the borders of China, and a Christian world which spanned Spain and Russia, so the Jewish world stretched from its heart in Central Europe and the Middle East, all the way to far flung outposts such as Kaifeng in North China and Kerala in South India. But powerful streams of cultural interconnectedness do not necessarily entail a great deal of gene flow.

Let’s go to the results as illustrated in the figures. First a table which shows pairwise genetic distances between the Jewish populations enumerated above and selected groups from the HGDP database. The numbers above the diagonal represent Fst, in other words the proportion of genetic variation within the total population as defined by the row-column pair which is between population. The bigger the number, the greater the genetic distance between the two populations.


Since I had to shrink the figure some, here’s the text which describes the gist of these results:

These findings demonstrated that the most distant and differentiated of the Jewish populations were Iranian Jews followed by Iraqi Jews (average FST to all other Jewish populations 0.016 and 0.011, respectively). The closest genetic distance was between Greek and Turkish Sephardic Jews (FST = 0.001) who, in turn, were close to Italian, Syrian, and Ashkenazi Jews. Thus, two major groups were identifiable that could be characterized as Middle Eastern Jews and European/Syrian Jews, an observation that was supported by pairwise FST and by phylogenetic tree analysis….

The Turkish and Greek communities were operationally nearly unified until the independence of Greece 150 years ago, so the small distance makes sense. It is notable that the distinction in terms of genetic distances maps onto that between the Roman and Persian Empires, where two Jewish communities emerged with different loci, Mesopotamia and the Palestine-Alexandria axis, respectively. Syrian Jews, who were within the boundaries of the Roman Empire, are more similar to European Jews than Iraqi Jews to their east. Though this may be due in part to the influx of Sephardim from Spain within the past few hundred years.

But Fst numbers can be hard to interpret in a gestalt fashion. So let’s look at PCA plots. They filtered the SNPs for the most ancestrally informative ones; i.e., ones which exhibit lots of between population difference. With these SNPs they extracted the largest independent components of variation. Note that the difference between PC 2 and 3 is small in magnitude, and so both are of interest. First, here are the Jewish groups in aggregate as they relate to other HGDP populations:


No surprise. Jews span Europeans and Middle Easterners. But let’s drill down to a finer grain. They also used the PopRes data set, which from what I recall is a bit more cosmopolitan than the HGDP one. I’ve added some clarifying labels.


The above changes nothing really in how we understand the relationships of Jews, in particular Ashkenazi Jews, to Europeans. Roughly, Jewish genetic relatedness to European groups tracks how strongly influenced by Rome a region was. Jews are closest to Italians, least close to Finns and Russians. Also, remember to be careful about PCA plots; from what I can gather these dimensions fall out of the set of SNPs designed to maximize between population differences between the Jewish groups so as to increase the power to distinguish Jewish clusters.

Going back to the HGDP sample, you see similar patterns.



Iranian and Iraqi Jews, Jews who were not touched by the Sephardic Diaspora, or, the Roman Empire, are distinct from the Jewish groups to their west. In fact it is interesting to observe that the various Levantine Arab groups are rather close to Syrian Jews when set next to the Iraqi and Iranian Jews, at least in total genome content.

Another way to look at the variation is through Structure, where there are K ancestral groups, and individual genomes are conceived of as a synthesis of K groups.


The Structure plot confirms that Ashkenazi Jews are more European than other Jewish groups, and Iranian Jews the least European. This influence of geography, or isolation by distance, shows up in other studies. But it should be weaker or non-existent in a perfectly cosmopolitan Jewish Diaspora where distance is no consideration. This model seems false. The most plausible explanation for the patterns here, supported by uniparental lineages, is that local Jewish populations have admixed with surrounding populations. Of course it could be that Ashkenazi Jews went through a population bottleneck and became a highly endogamous inbred community, so that genetic drift resulted in their uniqueness. But in that case they should show up as distinctive as the Kalash of Pakistan, who may be thought to have formed their own “micro-race” through genetic isolation.

Switching back to a big-picture summary of the genetic relationships, here’s a phylogenetic tree which was generated with the Fst numbers above. I think these should be viewed with caution, as trees like this are sharp and discontinuous by their nature. Even the authors observe that these Jewish groups, as well as human populations in general, have been characterized by gene flow and admixture over time, so that the assumptions which underly some of these representations are idealizations. Trees are invariably claimed to be robust, and yet somehow I’ve seen a really wide range of trees across different studies contingent upon the marker set or technique for the same set of populations quite often.


Finally, the authors examined the degree of identity by descent (IBD) across the genome of the Jewish groups. IBDjust refers to the fact that a region of the genome is identical with another because they’re descended from the same original copy. Siblings for example have huge regions of the genome identical by descent because each parent contributes one half of the offsprings’ genome. Over the generations the correlations of genetic variants across a physical strand are broken up by recombination. If two individuals who are putatively not related have long regions of the genome which are identical by descent that suggests that they share a recent common ancestor whose genomic contribution hasn’t been diluted by too much time and recombination.

Figure 3 of this paper summarizes the main IBD results. In panel A the red bars are Jewish-Jewish comparisons, yellow Jewish-non-Jewish, and blue non-Jewish-non-Jewish. Panel C plots the genetic relationships adduced from the IBD results on a 2-D plane.


Jewish groups share a lot of the genome identical by descent. Additionally, there’s a general agreement with the other results as to which groups are close to each other. They note in the text that the segments identical by descent among Jews are rather small, which implies  that recombination has broken up the large blocks. So that means that a high proportion of Jewish-Jewish IDB is a function more of many common ancestors deep in the past, rather than a few more recent common ancestors. Ashkenazi Jews in particular exhibit increased sharing of the genome across short blocks as opposed to longer ones, suggestive of a demographic expansion from a small population. Genic regions were was also moderately enriched around the loci which were IDB, a possible indication of functional commonalities across Jewish populations. If you’re interested in genes which Jews tend to share IDB, here they are a list:


After all that where are we? I think this section of the discussion addresses the broad brush findings:

The Middle Eastern populations were formed by Jews in the Babylonian and Persian empires who are thought to have remained geographically continuous in those locales. In contrast, the other Jewish populations were formed more recently from Jews who migrated or were expelled from Palestine and from individuals who were converted to Judaism during Hellenic-Hasmonean times, when proselytism was a common Jewish practice. During Greco-Roman times, recorded mass conversions led to 6 million people practicing Judaism in Roman times or up to 10% of the population of the Roman Empire. Thus, the genetic proximity of these European/Syrian Jewish populations, including Ashkenazi Jews, to each other and to French, Northern Italian, and Sardinian populations favors the idea of non-Semitic Mediterranean ancestry in the formation of the European/ Syrian Jewish groups and is incompatible with theories that Ashkenazi Jews are for the most part the direct lineal descendants of converted Khazars or Slavs. The genetic proximity of Ashkenazi Jews to southern European populations has been observed in several other recent studies.

Early history matters, and what these findings point to is that a division between western and eastern Jews which falls along the lines of Roman-Persian political division exists today even after 2,000 years. In terms of both culture and genetics there is “first mover” advantage. Even though only a minority of the population of the United States is of English origin, the vast majority of Americans speak English, and adhere to cultural traditions of English provenance. Similarly, admixture events early in the history of a group may have an outsized effect contingent upon later variations in population size.

Focusing more on specific cultural and historical parameters the authors note that what was Jewish in the time of Augustus was very different from what was Jewish in the time of Charlemagne. By the time of Charlemagne the Judaism of the Pharisees had marginalized other groups (excepting to some extent the Karaites). In the time of Augustus Jews were divided between different sects and persuasions, and there was a welter of diversity. Additionally, in the marketplace of Roman religion Jews were a moderately entrepreneurial group. The dynasty of Herod himself was of convert origin. There was a wide spectrum of Jewish religious practice and belief, from the near monastic isolation of the Essenes, to the engaged but separatist Pharisees, and finally to the wide range of more syncretistic practices which fall under the rubric of “Hellenistic Judaism.” Many scholars assert that it was from the last sector which Christianity finally arose as a Jewish sect, and that Christianity eventually absorbed all the other forms of Hellenistic Judaism. Judaism of the Pharisees, which became Rabbinical Judaism, and more recently Judaism qua Judaism, was shaped in large part by having to accommodate and placate the dominant Christian and Islamic religious cultures in which it was integrated by the early medieval period. Conversion to Judaism from Christianity or Islam was often a capital crime (though conversion from Christianity to Judaism was not forbidden in Muslim lands, while presumably conversion from Islam to Judaism in Christian lands would not have been, though few Muslims lived in Christian lands). So after 500 A.D. it seems that what may have occurred was that a Jewish Diaspora characterized by geographically determined genetic diversity, despite some common original Levantine origin, was genetically isolated from surrounding populations. This explains why there seems relatively little influx of Slavic genes into the Ashkenazim despite their long sojourn within Poland-Lithuania and later the Russian Empire. In contrast, the Roman Jewish community was already large in the days of Julius Caesar, and presumably intermarried with the urban proletariat of diverse origins.In an ironic twist these data suggest that modern Jews, in particular the Ashkenazim, but to a lesser extent the Sephardim as well, share common ancestry with gentile Europeans due to the unconstrained character of the pagan Greco-Roman world which Jews were to a great extent strident critics of. Contra Tertullian Athens had much to do with Jerusalem.

Image Credit: Wikipedia

Citation: Atzmon, G., Hao, L., Pe’er, I., Velez, C., Pearlman, A., Palamara, P., Morrow, B., Friedman, E., Oddoux, C., & Burns, E. (2010). Abraham’s Children in the Genome Era: Major Jewish Diaspora Populations Comprise Distinct Genetic Clusters with Shared Middle Eastern Ancestry The American Journal of Human Genetics DOI: 10.1016/j.ajhg.2010.04.015

CATEGORIZED UNDER: Genetics, Genomics, History

Comments (55)

  1. dpwes

    PopRes NE = 20+ Poles, 6 Russians, 1 Finn, 1 Latvian, and 1 Ukrainian.

  2. Michelle


    I think the HUGE population of Syrian Jews would be very surprised to find that “The native Syrian liturgical tradition apparently persisted down into the modern period before its recent extinction.” Their liturgical tradition is most certainly alive and well!

  3. bioIgnoramus

    Are there suitable data to test the proposition that Palestinians – or perhaps subgroups within the Palestintans – are proportionately more heavily descended from “ancient Levantine Jews who date back to Roman antiquity” than are, on average, Jewish Israelis?

  4. Answerthis

    exactly..THIS is the question that should be asked and answered…….

    Are there suitable data to test the proposition that Palestinians – or perhaps subgroups within the Palestintans – are proportionately more heavily descended from “ancient Levantine Jews who date back to Roman antiquity” than are, on average, Jewish Israelis?

  5. Are there suitable data to test the proposition that Palestinians

    ancient DNA extracted from jewish burials? the issue here is that i think “jew” was a much more open and fluid category during the antique period than today.

  6. michelle, please read this:

    please clarify your comment with more detail as you may know more than this entry in wikipedia, or, if you don’t know more than this, never comment again because you’re comment had very little value-add.

  7. note to commenters: please stay on topic and non-conspiratorial. and non-stupid. otherwise your comment won’t be published anyhow.

  8. Aspen

    Dear Mr. Khan:

    I note that the title of your article has been changed from what it was when originally posted. (The original title is still reflected in the link.) I am very glad to see the change, but wish that you would openly address the problematic nature of the original title and why you made the alteration, rather than merely using the adjustable capabilities of the electronic publishing medium to try to cover the error.

    (Edit, added after seeing comment #8 published: I believe this is on-topic, and germane to the discussion. While not addressing the specific genetically-focused content of your paper, there is cultural context around those same specific topics that make this a pertinent issue. I would very much like to see you take the opportunity that you have here to recognize that.)

  9. Mohammed Lebbadi

    This is all very interesting, but I think it ignores something extremely important which some have referred to (the origin of the Palestinians): We know for sure that Judaism was a dominant religion in the Maghreb at one time, before Islam. Most Maghrebian Jews were Islamized and later many Sepharadic Jews were Islamized. There are Muslim and Jewish families in Morocco, especially so-called Berbers, who share the same names and we have Muslim families who have Sepharadic names. Something else, we also have common saints. My question is: why not extend the study to include whole populations, not just Jews, to see how the groups are related?

  10. We know for sure that Judaism was a dominant religion in the Maghreb at one time, before Islam.

    this is false. christianity was the dominant organized religion (though there were certainly jews and your ultimate point would be interesting to investigate, and there are indicates within this paper that there may be a non-trivial impact of north african ancestry in sephardic jews; control-f mozabite).

    aspen, i determine what is on topic, you do not. the question is not on topic and further inquiries will be tagged as spam. my title was meant as an implicit pointer to 19th century german discussions on the topic (marx and baeur). i didn’t want the conversation to veer into talking about the evolution of the term ‘the jewish question’ so i didn’t elaborate. this obviously was a problem because people kept talking about other usages of the term in the comments (specifically, the nazi one). i didn’t want to have to talk about the issue in the comments after others began to focus on the nazi use of the term because i’d rather focus on the ancient history and genetics. so i changed the title.

    nevertheless, you still want to talk about it, as do many others whose comments i’m not going to publish. you can keep submitting the comments, but i’m going to tag the comments as spam, just as i’m doing with the other patently crazy stuff that’s coming in right now. the history of the term and the relationship of jews to modern nation-states has some relationship to this post, but i’m going to censor most of the comments which focus on it because i’d rather hear more about genetic and ancient historical questions, which most people don’t talk much about (either through ignorance or lack of interest, i don’t know).

  11. Derin

    Arthur Kostler has dealt with this issue decades ago.

    Koestler a “Jew” shows in his book The Thirteenth Tribe, that so called “Jews” are not who they say they are, but rather descendants of converts to Judaism from deep within eastern Europe.

    Some distance from the African mainland where the original Hebrew peoples had their origins and trials.

  12. but rather descendants of converts to Judaism from deep within eastern Europe.

    these data falsify this claim in its strong form (that ashkenazi jews are predominantly descended from eastern european converts). i’m going to stop publishing people who make assertions which have been falsified by the data i reviewed in this very post.

  13. probably futile, but if you’re going to leave an insulting comment, feel free to just email me. the comment won’t be seen by anyone else.

  14. Aspen

    Mr Khan:

    While authorial intent may engage with reader interpretation of the text, said intent does not strictly define nor limit such interpretation. You are, of course, entitled to your opinion and to direct the flow of discussions resulting from your published article, but I disagree that this point is inherently off-topic; in fact, regarding that difference of opinion, I refer you to your own words above:

    Of course these issues can not be understood except in light of a complex historically contingent sequence of events.

    Exactly so. You state that your original title was meant to reference Bauer’s 1843 book and Marx’s 1844 essay in response. In return, I point out that both works addressed religious, sociopolitical, and economic issues, not genetic ones. How then does that make your first title any more applicable to the subject than subsequent remarks referring to the extended and more current historical context?

    Finally, a closer reading of your article leaves me with a number of questions regarding your attempt to establish a connection between the sociocultural and religious definitions of group identity and the examination of genetic inheritance between specifically-defined population groups and external comparators in the cited paper. The issues surrounding racial and ethnic identity are deep and complex ones, and mention of the semantics needed in defining the terms and appropriate context for a specific discussion here appears to me to be largely lacking, which is unfortunate.

    With all due respect, I believe I’ll disengage from this discussion now, and go read and analyze the original paper as cited instead. I suspect I’ll have a better and more historically and scientifically satisfying experience in doing so.

  15. Elisabeth

    Your outright dismissal of the highly offensive nature of your original title (a title that can still be seen in the link) as unimportant is extremely bothersome. It should be very simple to avoid titling your piece on Jewish genetics in a way that makes a clear link to the attempted genocide of the Jewish race. An acknowledgement of error and understanding of the offense you caused should be the natural response to realizing said error. Pretending that it doesn’t matter at all, or that the hurt you caused is less important than the discussion you hope to provoke with this article, is precisely the wrong stance to take. I sincerely hope you rethink it and realize the mistake you made.

  16. elisabeth, you incorrectly characterized mohommed’s assertion. as for you second point, obviously it was a stupid title since i’m spending time talking about the title instead of the content of the post. for the purposes of this post i had in mind 19th century usage of the term, not the 20th century one. but many did not obviously. so i changed it. end of story.

  17. Leslie Schwartz

    Really interesting, and a very broad, encompassing study.

    I wonder about the genetic distance or similarity with the earliest identified ancestors of the Jewish people, for example, Abram who was a Sumerian, and lived approximately 3800 years ago.

    Also, Jacob, and the twelve tribes of the ancient nation. Presuming any of these persons or populations existed, there might be something indicative of that found in the DNA from ancient burials.

    Is there any genetic evidence for the existence of the twelve tribes, and were
    1) they descendants of a single ancestor
    2) otherwise twelve unrelated tribes of that geography, or
    3) possibly the descendants of Abram’s entourage of 380 Sumerian soldiers and their families?

    And given that Abram was Sumerian, and presumably a member of the nobility, do studies of Sumerian burials and their DNA give evidence of the genetic distance or similarity to the Jewish Middle Eastern population of 3000 years ago?



  18. Dan Ross

    It would be very interesting to compare these groups to the genetics of the Samaritans. The assumption is that Samaritans are the most continuous and (until the 20th century) isolated population from ancient Jewry (assuming that southern Judeans and northern Israelites were genetically similar), although like Jews they circulated around the Roman Empire and may have had some admixture at that time. This might also shed light on your assumption that Iranian and Iraqi Jews descend from the original Babylonian Exile, which is not at all certain. And, since the Samaritan community predates the Druze by at least 800 years, it could also shed light on the relationships of Druze, Palestinians, Bedouin and Middle Eastern Jews.

    It would not be correct to lump together the Kaifeng and Cochin (Kerala) communities as “far-flung outposts.” Kaifeng’s Jews were indeed an outpost of the Middle Eastern/Persian trading network until they lost touch with mainstream Jewry. Cochin’s origins are more mysterious, with presumably significant genetic admixture prior to their reencounter with Middle Eastern Jews.

    Separately, this analysis leaves me wondering about the “demographic miracle” of Ashkenazi Jews between the 15th and 19th centuries. If not due to significant intermarriage or conversion, how can it be explained?

  19. Presuming any of these persons or populations existed

    key issue. i don’t think we can go much further than the mid-first millennium BCE, so abraham, etc., is somewhat moot. additionally, the scholarship that i’ve seen suggests that abram’s sumerian origins may either be a translation error (a different ur than that of sumeria), or, an attempt to establish a mythological deep tie to mesopotamia during the first exile.

    i’d also lean toward the proposition that by the year early roman period that middle eastern component of the modern jewish ancestry could be traced back to several distinct middle eastern groups (i assume that the babylonian exiles intermarried with the locals and the anti-miscegenation stuff in their lit is trying to cover this up).

  20. Mohammed Lebbadi

    Mr Khan, you say it’s false, but how come our queen before Islam was Jewish, Al Kaheena, she united all the Berber tribes against the Arab invaders, how come there are no Christian traditions here if what you say is correct?

  21. but how come our queen before Islam was Jewish

    because some berbers were jewish. and was see a queen of all the berber tribes? my impression was she headed a particular confederacy, which itself was only a subset of the berber groups.

    how come there are no Christian traditions here if what you say is correct?

    the usual explanation i have read is that christianity was overly reliant on its institutional strength in urban and coastal areas, and once the arabs undermined the patronage network it wasn’t robust, especially in the face of persistent division between donatists and catholics. literary records observe a christian community in north africa up until the 11th century. by analogy, there were certainly more jews in iran in 1800 than zoroastrians, but zoroastrians were once the majority, not jews. why did jews persist? the usual explanation is that jews were not dependent on elite institutions and so could persist in the face of islamic dominance, while zoroastrians could not.

    certainly north africa had other groups besides christians. jews, manichees, as well as pagans, down to the time of the arab conquests. but from all i have read christians were the largest component of the organized religious (pagans may have been larger, though these are guesswork). you may be correct of course, but in my estimation the probability of there being more jews than christians in north africa in the late 7th century when the maghreb was conquered is low enough to reject out of hand.

  22. Heroic posting. My head’s about to explode from all the info. But in a good way!

  23. Ming

    These are the most cryptic graphs I’ve ever seen!

  24. pconroy

    Great analysis Razib – as usual!

    Notice that the Ashkenazi Jews (AJs) are very close to the Italians, and Italian Jews – which confirms earlier suspicions that Ashkenazi Jews got their start in Rome and surrounds and then relocated to the trading zone between the Roman provinces of Gaul and the Germanic lands. Remember Cologne and other places started basically as trading posts on the frontier.

  25. Check out the publications of Dr. Karl Skorecki, nephrologist at Rambam in Haifa who discovered the Cohen gene variant on the Y chromosome. Or even better, invite him to speak. What a treat.

  26. Mohammed Lebbadi

    Thanks, Mr Khan, for giving consideration to my remarks. What we know here is that there are no Christian traces outside the European enclaves which were scattered all around but quite isolated. And since Muslims didn’t force any group to convert, as shown by their presence in Iberia, I am wondering why there is no mention by historians of any Christian minorities who persisted, while on the contrary there’s plenty about Berber Jewish groups. I am no specialist, just wondering, and I think it would be interesting to do a genetic study of the whole population, let’s say in Morocco and Yemen for example, and of Jews originating in these countries to see how they are related.

  27. Thus, the genetic proximity of these European/Syrian Jewish populations, including Ashkenazi Jews, to each other and to French, Northern Italian, and Sardinian populations favors the idea of non-Semitic Mediterranean ancestry in the formation of the European/ Syrian Jewish groups and is incompatible with theories that Ashkenazi Jews are for the most part the direct lineal descendants of converted Khazars or Slavs. The genetic proximity of Ashkenazi Jews to southern European populations has been observed in several other recent studies.

    I’m not sure I’m following how they’ve discredited the Khazar theory. (At least, that’ s part of what I think the author is trying to hint at. Although they don’t explicitly say that, so I may be mistaken. But anyways….) If you look at this PC plot…
    … the Ashkenazi appear close to the Adygei. (The Adygei, being a Caucasian population, serving as a proxy for the Khazars.) Although this doesn’t prove Khazar theory, it still would seem to leave the question open.

    Or am I missing something?

  28. Mohammed Lebbadi

    ….why Morocco and Yemen as a suggestion for a case study? Because apparently the two populations are related, and the traditional architecture of the two countries is similar and attests to a close relationship long before Islam. And some historians say that the populations of both countries were predominantly Jewish. I think such a study would be fascinating for the whole world, it would reveal so many things, like the history of the Falashas (a 2001 study by the Department of Biological Sciences at Stanford University found a possible genetic similarity between 11 Ethiopian Falashas and 4 Yemenite Jews who took part in the testing), and perhaps confirm what some scholars are saying about the “Jewish people” being an invention.

  29. The Khazars were originally Turkish, like the Bulgars, so a Caucasian population wouldn’t necessarily be a good proxy. But not necessarily not, either, since the Turks coming in from the east might have intermarried with a population already present there. I’ve read what I could find about the Khazars, and when you throw out the speculation there’s not a lot known, and most of it is from old texts rather than archeology.

    Something I recently found out that I hadn’t known: speakers of Caucasian languages lived far out on the steppe up until the 18th and 19th centuries, when they were defeated and exiled by the Russians. Circassians now live scattered through the Caucasus, Turkey, and Israel.

  30. … the Ashkenazi appear close to the Adygei. (The Adygei, being a Caucasian population, serving as a proxy for the Khazars.) Although this doesn’t prove Khazar theory, it still would seem to leave the question open.

    uniparental markers imply the “european” component is more western/southern, not eastern. the adygei are a caucasian group which exhibit somewhat similar middle eastern-european mix, and so they are similar on that PCA plot. but remember that PCA can be deceptive.

  31. Razib . What is the bottom line right now about Ashkenazi origin? to what portion their origin is from the middle east ?

  32. bioIgnoramus

    Forgive my interference, Razza, but could use of “caucasian” to mean “from the Caucasus” confuse Americans used to “caucasian” meaning “white”?

  33. In 1917, my grandfather translated the Sephardic Syrian prayers books from Hebrew into Arabic. A very close comparison to the more recent Sephardic prayers books will reveal that they now contain Iraqi liturgy.

  34. @bioIgnoramus: The irony is that what has become the common U.S. (and Canadian) usage of “Caucasian” today, is that many people who are actually Caucasian (from Caucasia) would not be considered “Caucasian” in the U.S. (and Canada).

  35. Mohammed Lebbadi

    I think there’s nothing wrong with the “Jewish question,” it is a big question and we shouldn’t hide from it. Israel was built on an interpretation of who the Jews are, a people, a religion, or what? My idea, derived from readings of opposite theories, is that just like there are no “Christian people” or “Muslim people,” there are no “Jewish people” as such! Judaism is a religion, no more no less, and the fact that scientists find similarities between different Jewish groups is no surprise, because conversion to Judaism was not as massive as for the other religions. There were substantial coversions in some places, southern Arabia was one, and from there Judaism expanded into the horn of Africa and probably moved west to the Maghreb. I think throughout history there were converts and reverts, so why the big fuss, why try by all means to find a “Jewish people” when no such group exists! Or, if one tries too hard, he will find that most Yemenis and most Moroccans are of Jewish origin and then we would claim our “rights” to Israel and give it back to its original inhabitants, the Palestinians. There is a “Jewish question” and I believe it should be answered!

  36. deadpost

    So, regarding those far-flung, and “non-mainstream” Jewish groups, such as the Ethiopian, Keralite and Chinese, I assume that if they were put into the picture, they might not differ that much from their local Asians and Africans?

  37. I believe the word “Ashkenazi” originally referred to Germany rather than eastern europe. So it is surprising that their genes are relative far from Germans and close to Italians. Is the working hypothesis that there was an original founding hybridization that occurred in Italy, and then the Italian Jews were the ones who stayed while the Ashkenazi moved to Germany but did not subsequently hybridize again?

    The Adygei don’t look so close in PC1 (4.4).

    Mohammed, it’s all stolen from the Neandertals.

  38. nebbish

    You are correct about the origin of the term “Ashkenaz.” The Ashkenazim started out in France and the Rhineland and moved east due to massacres associated with the Crusades and repeated expulsions that took place during the medieval period.

    This autosomal study found little evidence of introgression from either Germans or Slavs, and based on uniparental markers as well as the pattern of clustering and the genetic distance data, admixture from southern Europeans (Romans/Italians, Greeks and probably Anatolians who were heavily Greek influenced during the Hellenistic period) during the pre-Christian era has been proposed. This fits the known presence of significant, proselytizing Jewish populations in these areas from at least 100 B.C.E. Moreover, intermarriage was severely punished by both sides after the rise of Christianity.

  39. toto

    It seems that this study definitely nails the general picture, at least as far as Ashkenazis are concerned – Middle Eastern population moves to Italy, hybridizes, then a portion of it moves to Central and Eastern Europe.

    Like previous commenters, I’d also like to know more about the relationship between modern Jews and modern (non-Jewish) Palestinians. The default hypothesis is that the Palestinians are also hybrids between the Jews who stayed and converted, and migrants from Christian and Muslim areas.

    The difficulty is that much of these migrations would come from neighbouring regions, which are likely to share much genetic heritage with the original Jewish population. I understand that the “Cohen modal haplotype” is more common in Jordan than among Cohanim (?) So if you find that Palestinians are closer to their neighbours than Jews, what does that tell you? Ancient origin, or migration?

    So, regarding those far-flung, and “non-mainstream” Jewish groups, such as the Ethiopian, Keralite and Chinese, I assume that if they were put into the picture, they might not differ that much from their local Asians and Africans?

    If these people simply converted en masse to judaism, “Khazar style”, without any significant influx of people, then yes. But if they were founded by Jewish settlers, then you could still detect a genetic signal – presumably on the Y chromosome.

    For the Ethiopians, some people have already looked. Verdict: negative. OTOH, there seems to be a clear signal among the Southern African Lemba.

  40. @TGGP: The paper is behind a pay wall, so I can’t just read the paper to figure it out myself, but…. What’s the difference between the 2 PC plots?

    Is the x-axis the same eigenvector on both? (And the y-axis different eigenvectors in each?) (Possibly both x-axises being the principal component?)

    Also, what are the eigenvalues of each of these eigenvectors? It’s difficult to judge what is and isn’t negligible without the eigenvalues.

    I.e., without knowing what the eigenvalues are for each of the eigenvectors (which are the axises in those PC plots) it’s difficult to know if all the variation we see in those graphs actually “matters” or if some of it can be considered “negligible”. (Well, assuming I’m reading those PC plots correctly.)

  41. Jordan

    I wish I knew enough math and genetics to understand this work better. As it is, I can just get the gist of what it is describing.

    FWIW, my understanding is that while there are a few lines of evidence to suggest that some of the Khazar nobility / royalty adopted Judaism or at least Jewish practices, there was never any decent evidence for a “mass conversion” of the Khazar people as a whole.

    The Italian / southern European connection would be fascinating to explore in more depth. I have read claims that a significant number of people from Calabria have Jewish ancestry (either through the Sephardic exodus or earlier migrations). It’d be very interesting to have a finer-grained look at Italy and perhaps southern France as well to see if there are particularly strong connections with any Jewish sub-groups.

  42. Ben

    Great study, and great analysis. I think, though, that in some ways the study overly conflates Jewish populations based on nationality; “Iraqi Jews,” for example, are really two very distinct communities: Arabic speakers in the south and center, and Aramaic speakers in the north. It weakens the study that there is a single category; it would be useful to know what group(s) the people in the study were drawn from, as this is more informative than mere national origin.

  43. bioIgnoramus

    Forgive the OTity, Razib, but I thought that this tale of medieval european/outsider contrasts might take your interest.

  44. benj

    Very interesting, thank you.

    A few comments.
    1. The myth of the Jewish Berber Kahina is just that – a myth. Never happened.
    2. Regarding mass conversions during the Roman Empire – this is not fully established. There were conversions but to what extent nobody knows at all. We have no figures. In fact the most probable theory is that most of these converts were never more than “Judaizing” and were then the first to convert to Christianity.
    3. The Pharisee were the vast majority of Judean Jews and the mainstream, not just a sect among others. The situation was maybe different in Alexandria, we don’t know for sure, and it seems that “Hellenic Judaism” was much stronger there. But in the end, rabbinical Judaism has been the mainstream form of Judaism for 2300 years.

  45. Leslie Schwartz

    21. Razib Khan Says:
    June 6th, 2010 at 4:22 pm

    “Presuming any of these persons or populations existed

    key issue. i don’t think we can go much further than the mid-first millennium BCE, so abraham, etc., is somewhat moot. additionally, the scholarship that i’ve seen suggests that abram’s sumerian origins may either be a translation error (a different ur than that of sumeria), or, an attempt to establish a mythological deep tie to mesopotamia during the first exile”

    OK, of course I have to defer to your scientific knowledge of genetics, but I would like to learn more about this.

    I understand that even the Neanderthal genome is being worked out now, down to the 3 billion base pairs. So that must mean they are working with samples originating at least 30,000 years ago. It might be that while sets of the entire base pair structure can be worked out, perhaps the haplotypes which are used to work out population sub groups, are no longer available typically in samples from prior to 500 BCE as asserted above? If so, I find this surprising.

    On the other hand, I have seen some studies of Sumerian burials where some genetic studies were attempted, and they must range back to 4000 to 6000 BCE.

    I have also never heard before that anyone questioned the Sumerian origins of “Abram”.

    Abram, could have well been a mythical person, and we do not have his remains to test genetically, but I would also like to know what good scholarship exists to dispute the Sumerian origins of Abram, again, presuming he is not entirely a fictional character.

    There also must be well identified ancient burial places in Israel, which are associated with specific regional areas associated with different tribal settlements, according to the Old Testament, and again, if enough resources were applied over time it might be possible to see if those populations show a single family relationship, over all the tribal areas that can be identified.

    My guess again would be that there would be more diversity found among ancient burials as well, and that not all ancient Hebrews do descend from a single family, a good concept for building social cohesiveness, but genetically likely, – maybe not.

  46. Bulan

    Khazaria will rise again!

  47. B

    I find it interesting that the Jews of the Maghreb and Iran were supposedly independent of the elites; Prof. Netanyahu, in his work on the Spanish Inquisition, sets as his central tenet that the Jews of Spain survived by the exact opposite strategy. Any idea on the reason for this difference?


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar