No, Ashkenazi Jews Are Not Genetically European

One does not have to look far to see misinformed claims that Jews are “white Europeans” and “genetically European” widely circulating. Of course, these incorrect assertions are driven by an ideological agenda.

To use the words of Rabbi Jonathan Sacks, antisemitism is “a virus that mutates”. That is, antisemitism is constantly adapting to shifting zeitgeists. Antisemites in the past attacked Jews for not being Europeans. Yet, in today’s political climate, antisemites are attacking Jews, particularly Ashkenazi Jews (i.e. Jews with exile history in Central and Eastern Europe) for the exact opposite reason, and DNA has become their weapon of choice. So the legacy continues.

Antisemitic claims on Twitter. Possibly relevant “John Hopkins” genetic study/author disputed here.

Their claim, essentially, is that Ashkenazi Jews are largely — if not entirely — unrelated to the Jews of antiquity, instead deriving their ancestry from local Europeans who, at various points, converted to Judaism. The intent of this claim is to mark Ashkenazi Jews as “privileged oppressors”, or more specifically, to delegitimize Zionism (and, thus, Israel’s re-establishment) as a “white European settler-colonial project” by erasing Ashkenazi origins in the Levant.

Their arguments may seem convincing to those unfamiliar with genetics but, like other antisemitic “arguments”, they rely on distortions of data, omissions, and outright lies. Equally, they rely on people not being aware of the science behind genetics.

Thankfully, it is not difficult to disprove such claims, as this article aims to do. Below you will find abundant evidence that Ashkenazim are in fact genetically of Middle Eastern origin, and very substantially so. Included in the appendix are some additional explanations of the science involved.

Key Takeaways (if you want the TLDR version)

• Ashkenazi Jews, while having absorbed some ancestry from Europe due to a lengthy exile, have still retained the majority of their Israelite ancestry.
• The paternal lineages are overwhelmingly Middle Eastern, being defined by haplogroups J and E1b being predominant, but also the minor presences of R1b, G, and R1a.
• When we compared the paternal lineages of Ashkenazim with European host populations and other Middle Eastern populations, the Ashkenazi profile was virtually indistinguishable from the Middle Eastern ones yet completely different from the European ones.
• Many paternal lineages in Ashkenazim that might look European on the surface are actually Middle Eastern.
• The maternal lineages in Ashkenazi Jews, while showing a general affinity to European lineages, still show some retention from the Middle East.
• Other genetic studies (autosomal admixture and proximity) show the Ashkenazi gene pool is roughly 60% Middle Eastern overall, and a genetic closeness to other Jewish populations is clearly observable, along with an adjacency and partial overlap with non-Jewish Levantine populations.
• Commercial DNA testing such as at AncestryDNA or 23andme can mislead the public into thinking Ashkenazi Jews are genetically European due to the wording and scheming of commercial tests. Thankfully, the companies provide additional information to correct misunderstandings, but it is best they adjusted their presentation.
• The theory that Jews are significantly (or entirely) descended from the Khazars has been long-debunked.
• The nature of Ashkenazi Jewish genetic mixture is in many ways comparable to other populations, such as Roma, Mestizos, Metis, or African Americans.
• Multiple examples presented show how antisemites weaponize public unawareness of DNA to lie and mislead the public about the origins of Ashkenazi Jews, and use double standards to invalidate their identity as a Middle Eastern People.
• Regardless of the exact proportions of European vs Middle Eastern DNA they may possess, Ashkenazi Jews are a diaspora population with beginnings in the Middle East, and their identity as a Middle Eastern group is as valid as any other.

Full article:

Ashkenazi Y-DNA (Direct Paternal Lineages)

What you may notice about the “Ashkenazi Jewish = European” claims, is that despite being possibly the most useful tool of population genetics, Y-DNA is almost never cited as evidence.

Naturally, this is because no matter what study you consult, Ashkenazi Y-DNA clearly reflects an overwhelming paternal origin in the Middle East.

The Ashkenazi Y-chromosomal gene pool includes the following paternal haplogroups: J (~38%), E (~20%), G (~10%), R1a/R1b (~7.5% and ~10% respectively), and a few trace haplogroups at 5% or under (eg. T, Q, I, and R2).

At face value, the Y-DNA haplogroups of Ashkenazim are pretty much exactly what one would expect from a native population of the Levant. When we compare the Y-DNA haplogroup frequencies of Ashkenazi Jews to that of ethnic Europeans — especially groups like Poles and Germans, which hosted large Ashkenazi populations — there is an obvious and gaping disparity.

As is clear from the charts, the Ashkenazi and European Y-DNA profiles are wildly incongruent. The claim that Ashkenazim are the “same as Poles and Germans” is laughably ironic, as looking at Polish and German Y-DNA, they are actually the furthest genetically from Ashkenazi Y-DNA when we compare it with other MENA (Middle East & North Africa) populations.

Meanwhile, the Ashkenazi profile clearly resembles that of other Jewish/MENA populations:

It is clear that haplogroups J and E1b most clearly define these MENA populations, and to a lesser extent R1b, G, and R1a. Conversely, the European populations are predominated by R1a or R1b with substantial I1 & I2 input, yet relatively little J and E1b.

This is not to say there is zero overlap with, or admixture from European populations on the paternal line of Ashkenazi Jews. However, the overlap here, for example in R1a and R1b (which also exist in the other Middle Eastern populations), is actually in large part due to ancient relatedness between Europeans and Middle Easterners, rather than post-exile admixture from Europeans, as one might assume.

Some may try to lie with statistics and claim that R1a and R1b in Ashkenazim must be smoking gun proof of Ashkenazi European-ness (such as what happened with viral tweets about Benjamin Netanyahu’s son Yair being R1a), but that is not necessarily the case. In fact, in most cases, it is actually evidence of non-European ancestry when it comes to Ashkenazim. R1a and R1b, while predominant in Europe, are not exclusive to Europe. To wit, they have existed in the Middle East since ancient times and, in the case of Ashkenazim, were mostly inherited there and not in Europe.

For example, analyses of Ashkenazi Levitesshow that virtually 100% of their R1a sublineages belonged to the Asiatic R1a-Z93 clade (or more specifically R1a-M582/R1a-CTS6), not the clades that most European R1a’s belong to (sub-lineages of R1a-Z282). Other research suggests R1b in Ashkenazim is also in many cases inherited from pre-exile ancestors in the Middle East, rather than as European admixture. Oftentimes, there is no way to assume the origin of a haplogroup without deep clade testing (see appendix for more details).

Something else to note is that European admixture is evident in some Arab populations. For example, one study found substantial evidence of Western European Y-DNA lineages in Lebanese Christians, as a result of admixture from Crusaders during the Middle Ages. Yet curiously, no such claims of being “genetically European” are ever leveled at Lebanese Christians, only at Jews.

Ashkenazi mtDNA (Direct Maternal Lineages)

According to one study, a preponderance of mtDNA lineages (direct maternal lines) found in Ashkenazim are conclusively of European origin, with the remainder being either Middle Eastern or ambiguous.

Costa et al., 2013 U8 & K lineages in Ashkenazi Jews, determined as European, Middle Eastern, or undetermined

While there may be truth to this, and some lies do contain a kernel of truth, it is important not to assume that this vindicates the lie that Ashkenazim are simply “genetically Europeans”. This conclusion gets widely misrepresented as Ashkenazi Jews being mostly, or entirely European overall, such as in the title of this widely posted article, and of course, antisemites take advantage of it while making deliberate omissions regarding the rest of Ashkenazi DNA, which is much more clearly Middle Eastern.

One could theorize that the reason Middle Eastern source ancestry in Ashkenazim is skewed more towards the paternal lines, is that there were more Jewish slaves taken to Rome who were male than female. This would cause local European women to assimilate into the gene pool, and in terms of absolute numbers and the fact males can produce more children than females, would favor male lineages — or, in other words, for Middle Eastern ancestry to be inherited in greater proportions along the paternal lines.

Looking at Y-DNA and mtDNA together, the genetic input in Ashkenazi Jews is clearly bifurcated — the Middle Eastern and European components in Ashkenazim are inherited disproportionately along paternal and maternal lines, respectively. For other populations that underwent admixture events resulting from similar conditions (eg. exile, displacement, and colonization), it is not unusual that the source ancestry and the admixture also derive heterogeneously along either the paternal or maternal lines (depending on the circumstances of the mixture events), such is the case for the Indigenous Red River Metis people, the Mestizos of the Latin American world, and African Americans.

Autosomal DNA

We have all seen the ethnicity estimate pie charts on TV commercials. That is autosomal DNA. Unlike Y-DNA and mtDNA, autosomal DNA is inherited from all ancestral lines and is better at reflecting overall recent ancestry, but is sometimes used in population genetics as well.

In terms of admixture, Ashkenazim are most often described as “equal parts” Middle Eastern and European. However, a more precise ratio is about 60/40, the former being Middle Eastern (mostly Levantine) and the latter being European, according to a 2020 study analyzing a set of Bronze Age Canaanite remains found in Megiddo. These kinds of figures, along with the Y-DNA haplogroup frequencies, would be impossible if Ashkenazi Jews were just “genetically European”, or according to long-discredited theories, simply “Khazar descendants”.

It is inevitable that any colonized population will acquire some degree of admixture while in forced displacement from their homeland for great lengths of time. Considering Ashkenazim underwent 1000–2000 years of exile, these proportions are actually staggeringly high. Other non-European populations that underwent displacement, colonization, or both, appear to have retained far less of their source ancestry, or otherwise lost substantial amounts of it to admixture in much shorter frames of time. For example, African Americans have on average 24% European ancestry, and Latinos were found to have 65%. Roma were found to have less than 20% South Asian source ancestry and are believed to have been displaced from their homeland much later than Jews were. There are also some Indigenous nations that do not employ a blood quantum and thus have plenty of citizens, and in some cases leaders, with much less source ancestry yet. Of course, the same people using admixture to whitewash Ashkenazi Jews and alienate them from their homeland would probably never venture to make the same claims about these aforementioned groups. It is simply a double standard against Jews.

Commercial Ethnicity Estimates

With the exception of LivingDNA, most commercial autosomal tests classify (eg. map out on charts) Jewish DNA by recent diaspora location, not by their source ancestry. That is, a customer who is 100% Ashkenazi will not see a breakdown to the effect of “50% Levant, 10% Mesopotamian, 30% South European, etc”. Instead, they will be registered automatically as “100% European”, with a “100% Ashkenazi” sub-slot beneath that, as is the case with 23andme. Or, they will see simply “Jewish” with Central/Eastern Europe highlighted on the map (although AncestryDNA has since included the entire South and East Mediterranean as periphery in their map).

When people see this, they think of it as a slam dunk for the “Ashkenazim are European” argument. What they fail to realize is that when you classify a group as European, and then that group takes the test and it comes back as European, it is the circular result of classifying it as European in the first place. The same thing would happen for any group, not just Jews. If DNA tests considered, say, Turkish DNA to be European, Turks who test would come back as ~100% European too. This is the subjectivity involved with what is called population scheming, ultimately, how populations are grouped and mapped out for the test and subsequently analyzed (see appendix for more details).

In fairness to the DNA companies, it could be argued that this schema is more genealogically useful, since it looks within a relevant timeframe for discovering ancestors and relationships through DNA matching. If you were adopted and looking for your birth family, and the test said all your ancestors were in Israel and ignored diaspora, there would be limited utility for finding your family that lives/lived in the diaspora. Ultimately speaking, these ethnicity estimates are designed for finding your relatives and giving clues into your recent migration path; it is not meant as some objective “race test”, and certainly not meant to be used politically.

Moreover, the DNA companies generally include an explanation of their populations. For the examples of 23andme and AncestryDNA, they do clearly state that Ashkenazim have roots in the Levant (see below). The difference is that they treat the Middle Eastern ancestry as smoothed into the base of the Ashkenazi gene pool (i.e. as the mixture occurring further back in time), not unlike how all English people have Germanic ancestry from continental Europe (i.e. the Anglo-Saxons), as well as Celtic, Norse, etc., yet the test will show them as 100% British, rather than as German or Dutch or anything else. So, for Ashkenazim, the companies choose to report matching to their Ashkenazi data sets as matching to the geographic diaspora location, not the source location.

Nevertheless, while this presentation is not totally wrong, it remains problematic in how it leads people to understand the origins of Ashkenazim. People are less likely to actually find, let alone read the disclaimers or understand the intent of the DNA companies. As such, this presentation becomes exploitable for those with a political agenda seeking to erroneously whitewash Ashkenazim. If nothing else, to the layperson, it puts an inaccurate impression in their mind that Ashkenazi Jews are not a Levantine diaspora population, but merely genetic Europeans of Jewish faith. It is impossible to count the number of times that antisemites (and even other Jews) have tried to “correct” the statement that Ashkenazi Jewish origins are Levantine on the basis that “23andme says they’re European!”.

People’s ignorance of the science and their unwillingness to actually read all of the information is partly responsible. Still, it is incumbent upon these DNA companies to rethink how they present Ashkenazim, as regardless of their intentions or reasoning, this presentation in effect amounts to a political statement on who and what Jews are, as it is evidently interpreted and misinterpreted as such. It needs to be made clear, visually, from the maps and the charts that Ashkenazi roots are indeed Middle Eastern, while also remaining effective in guiding the user to discovering their ancestors and family in the diaspora.

PCA Charts (Principal Component Analysis)

There are two ways to organize data for PCA charts. On a sample-based PCA, you will find an array of individual samples from each population scattered across a small area. These charts are useful for showcasing genetic diversity and overlap between different populations. Ashkenazi Jews, for instance, typically straddle the space between the Levant and the Mediterranean regions surrounding it, with some Ashkenazi samples directly overlying non-Jewish Levantine populations and others hewing closer to Southern Europeans or Caucasians.

Sample-based PCA, Behar 2013 et al. Ashkenazim cluster with other Jewish populations, and adjacent to other Middle Eastern populations, partially overlapping

Cluster-based PCAs, by contrast, forgo individual samples and instead use a single “cluster” or dot representing an aggregate of individuals. Where this singular dot lands may be determined in part by outliers within the population, which can skew the average. That is, if one or two samples within a North African population (hypothetically) were to cluster nearer to Italians than other North Africans, it would drag the overall cluster north towards Europe and away from Africa. For similar reasons, the dot representing Ashkenazim will usually land near Sephardic Jews, Cretan Greeks, and Southern Italians.

PCA using clustering, Eurogenes K15. Populations are shown as a single aggregate rather than as samples

Typically, anti-Zionists will point to cluster-based PCAs to say that Ashkenazim are “actually from Italy”, with no connection to the Levant, which is nonsense. Generally, they will avoid sample-based PCAs, as they almost all show Ashkenazim overlapping with other Levantine populations. Having any kind of European admixture is obviously going to drag the Ashkenazi average somewhat closer to Europe, but it has absolutely no bearing on the actual components of Ashkenazi ancestry nor their origins.

To be clear, this does not mean the actual data (the blue/European color-coding for Jewish populations notwithstanding) in cluster-based charts is incorrect. They are designed to give a visual distance between averages of samples, rather than the overlap between individual samples or the range of expressions within a population. However, averages are notoriously poor for illustrating geographical ancestry, only similarity. Few people understand that similarity (“clustering” at or near a population due to a statistical average) does not necessarily indicate ancestry from that population. To illustrate, one could be, say, heterogeneously half Scottish and half Romanian and plot somewhere around Germany, because it is the average between the two populations, given the genetic differences between Scots and Romanians. It does not mean that such an individual is “genetically German” or had ancestors from Germany.

Cluster-based PCAs can have a variety of purposes. However, again, people misunderstand what the charts are actually saying, and misuse their erroneous interpretations of them to push a political agenda.

Proximity Lists

You’ll often see proximity lists like these (from Vahaduo, GEDmatch oracle, etc) being shared around by anti-Zionists and Arab nationalists on Twitter and Reddit. Almost as often, they will go as far as to conveniently crop Ashkenazim out of these lists entirely, so that it appears they are not connected to ancient Israelites/Canaanites at all.

Vahaduo proximity list

As with cluster-based PCAs, these lists are based on population aggregates and genetic similarity (proximity). They do not necessarily represent genetic ancestry. To wit, it is possible — if not commonplace — to have no ancestry whatsoever from the referent sample (e.g. Bronze Age Levantine found in Sidon) and still chart fairly close, particularly if the sample being tested is of mixed ancestry or belongs to a mixed population.

For instance, someone who is half-Cherokee and half-Nigerian will chart further away from “Cherokee” than will an Indigenous Siberian or Indigenous Amazonian individual with no Cherokee ancestry whatsoever. To give a more pertinent example: someone who is 75% Arabian and 25% Levantine will chart much closer to “Bronze Age Levantine” than someone who is 75% Levantine and 25% European.

Why? Because Arabians and ancient Levantines are geographically neighboring populations. They are already much genetically closer to each other than Europeans (particularly Europeans further away from Southern Europe) are to either. Thus, even if a sample is 75% Levantine, that 25% European will strongly pull their plot away from the Levant.

On such lists, Ashkenazi groups (i.e. aggregates, not individual samples) will chart closer to Sicilians than non-Jewish Levantines. Does this mean Ashkenazi Jews originate in Italy? No. All it means is they are a similar genetic average when their base components are mixed together — Ashkenazim are Levantines with a strong South European component on the maternal line (not to mention, the Greek Philistines who settled in the Levant in antiquity), and Sicilians are a largely Roman and Greek-descended population with a significant MENA pull (Phoenicians did settle Western Sicily, plus Southern Europeans have retained much more European Neolithic ancestry than other Europeans). That Ashkenazi aggregates may have stronger affinity (that is, similarity) to Sicilians than, say, Lebanese Christians is meaningless in regards to the origin of Ashkenazim as Jews.

“But don’t Ashkenazi Jews look European”?

We will let you make that determination: https://www.ashkenazijews.net/

Conclusion

No, Ashkenazi Jews are not “genetically European”.

No matter the exact proportions of DNA Ashkenazim may possess, it remains factually indisputable that they are a diaspora with ethnic origins in the Levant. Period. This is demonstrable by any sober examination of the genetic evidence. Thus, they have a fully legitimate claim to being scions of the Middle East, in particular the Levant.

It is also true that Ashkenazi Jews have acquired European admixture, like any number of other populations. However, no population — least of all one which has been colonized and involuntarily displaced — needs to be pure in order to belong. In fact, there is hardly such a thing as a “pure people’’ and it is exceedingly unlikely there ever was. Mixedness, blood quantum, and genetic purity do not determine peoplehood, ethnicity, and collective identity. Indigeneity, through ethnogenesis in a particular space and time, determines the collective identity of a people. No amount of admixture, time, cultural adaptation, phenotypical difference, or antisemitism changes that.

Appendix

The Science of Y-DNA and mtDNA

Unlike regular chromosomes (autosomes), which are inherited in 22 pairs (one from each parent), Y-DNA (Y-chromosome) and mtDNA (mitochondrial DNA) are inherited from only one parent. Y-DNA is only passed on by the father, but only genetic males inherit it. Conversely, mtDNA is only passed on by the mother, but all children inherit it.

These inheritance patterns, called direct lines, make the Y-chromosome an unbroken patrilineal (male) chain dating back thousands of years: your great-great-great-great-etc grandfathers. The same holds true for mtDNA, but with the matrilineal (female) line.

Moreover, unlike the autosomes, Y-DNA and mtDNA do not mix (recombine) every generation. Instead, they are passed on as exact copies from parent to child, save for random mutations that inevitably occur. When mutations occur, they get passed down from the ancestor who possessed the mutation, including all the previous mutations, becoming groups of mutations that are traceable across generations.

These groups of mutations are called haplogroups, which are assigned their own identifying numbers (e.g. R-Z93, E1b, etc). The frequency of haplogroups in a population can signify a lot about their ancestral origins and relatedness to other groups, both ancient and modern, when tracked across long stretches of time and geography.

It is imperative not to think of haplogroups, especially macro ones, as intrinsic to only one region. For example, haplogroups R1a and R1b are predominant amongst most European populations (R1a in Eastern Europe and R1b in Western Europe). However, they are also found in dense frequencies in certain non-European populations. For instance, in Northern India, R1a is found amongst Brahmins at frequencies similar if not greater than amongst Eastern Europeans. So, if we had a person who was R1a with no further information, one would need to test them for the deeper clades to know which ethnic group they actually inherited it from. If you go really far back to the days of the ancient proto-Indo-Europeans, yes, many European groups (particularly Balto-Slavic peoples) do share common ancestry with Northern Indians whose languages belong to the same family (Indo-European). Likewise, R1a and R1b are also found in many Middle Eastern and adjacent populations, especially amongst those who speak Indo-European languages (such as Iranians, Kurds, and Armenians), or those who had contact with such groups (such as Jews, Arabs, Druze, and others, albeit at lower frequencies).

Broad R1a frequencies Maulucioni, Wikimedia Commons

With the exception of the Family Tree DNA (their Big Y & mtDNA full sequence tests), most commercial DNA services do not offer testing for deep subclades, only basic or moderate ones. Not knowing the deep clade of say, R1b, is akin to linguistics: “We know the language these people are speaking is descended from an early form of Latin, but we don’t know what this language is, which dialect, or when/where it became a thing”. So, a broad haplogroup like “R1b” is a very low resolution view that spans very large geography very far back in time. The specific deep subclades narrow things down to where and when roughly each mutation group along the line occurred. It is not necessarily the frequency of broad clades in a region that determines origin.

As for mtDNA, mutations happen far less frequently in the mitochondrial DNA than mutations of the Y-DNA. Therefore, it is much harder to track changes for mtDNA over time and, by extension, the differences between populations using mtDNA. Consequently, for some Ashkenazi mtDNA clades (e.g. some K subclades), geneticists are unsure if they originated in Europe or in the Middle East — since they coexisted in both locations at the time the Jews were exiled in Europe and had not yet diverged.

The Science of Autosomal DNA

Autosomal DNA does not use Y-DNA or mtDNA at all. It works on an entirely different principle.

Autosomal DNA refers to the 22 pairs of regular chromosomes. For each of these pairs, one individual chromosome is inherited from each parent, which then undergoes recombination. Recombination means that each chromosome passed onto the child is a mixture of each parent’s two chromosomes. The X chromosome is transmitted slightly differently with its own inheritance pattern, but is usually included in ethnicity/admixture estimates and matching.

In lay terms, autosomal data analyzes the chromosomes inherited from both parents (and makes up the bulk of total DNA), as opposed to focusing on lineages inherited from a single parent. Direct line testing gives more of an objective view into one’s deep ancestry based on archaeological and living samples, whereas autosomal relies primarily on living populations and is more suited for interpreting recent ancestry or the overall “closeness” of populations.

How Commercial Ethnicity Estimates Work

These tests rely on population scheming, which means how populations are categorized. The DNA companies essentially set out and collect DNA samples from various populations around the world and organize them into collective data sets representing each of those populations. Then, they group these populations together into geographic regions and subregions. When you take the DNA test, your DNA is compared, using varying mathematical and algorithmic calculations, to the reference data sets to determine which of these geographic regions your DNA can be thereby divided into.

What is important to understand about this process is the inherent subjectivity involved. If you have taken more than one test, you will notice that each company separates and classifies each population and region differently. For example, what counts as “Eastern European” may or may not include the Balkans or the Baltics. It might be split between West and East Slavic, or some populations straddling the border between regions, for example Romania, may be classified as Eastern European, or as Balkan. Lithuania may be considered Baltic, along with Estonia, Latvia, and Finland. Or it might be shoveled into the Eastern European category. Some populations might be classified in more than one regional category. Point is, how populations are defined is up to the companies. There is no objective way to classify what constitutes any given region. Not to mention, many neighboring (or even distant) populations can be closely related but separated by regions on the tests, and as such your ancestry can be misattributed and given a false impression of where your ancestors are from (a common problem with say, Irish and Scottish).

In sum, the way the DNA companies categorize the populations has an enormous effect on the actual results you receive. Therefore it is important to understand how they are classifying your ancestral populations so you can be wary of the accuracy level and what the results mean to you and others interpreting them.