What Tunisian DNA Actually Shows

16.4K

Every few months a colourful map of “Tunisian genetic admixture” goes round social media: pie charts over Bizerte, Sfax, Douz, each sliced into Berber, Arab, European, Anatolian, Caucasian, Sub-Saharan. It is handsome, it confirms something everyone already believes about Tunisia being a crossroads, and it is not science. It carries no author, no sample size, no method, and its category names come from consumer ancestry kits rather than from any published paper.

The real research is more interesting than the map, and in places it says the opposite of what people assume. Over the past fifteen years, genome-wide studies from the Institute of Evolutionary Biology in Barcelona, the Faculty of Sciences of Tunis, and groups in Porto and elsewhere have sampled named Tunisian villages — Chenini, Douiret, Sened, Matmata, Jradou, Wesletia, Kairouan, Testour — and counted what is actually in the genomes. Here is what they found.

Four Components, Not Six

Population geneticists working on North Africa consistently resolve the region’s ancestry into four components, not the six on the viral map:

• The Maghrebi component — the autochthonous North African ancestry, sometimes called the North African or Amazigh component. It is the oldest layer, and the one that makes North Africa genetically distinct from both Europe and sub-Saharan Africa.
• The Middle Eastern component — arriving chiefly with the Arab-Muslim expansion from the seventh century onward.
• The European component — Mediterranean gene flow, ancient and recent.
• The sub-Saharan component — overwhelmingly the result of the trans-Saharan trade, and much more recent than most readers expect.

“Anatolian” and “Caucasian” are not components in this literature. They are labels a commercial algorithm assigns to fragments that the academic models place inside the Middle Eastern and European signals. That does not mean no Ottoman or Caucasian ancestry exists in Tunisia — it does, and there is a specific and well-documented reason for it, which we come to below — only that it is not a separate ancestral population in the way the map implies.

The Maghrebi component is the headline finding. Brenna Henn and colleagues, working with roughly 730,000 markers across seven North African populations, identified it as a genuinely indigenous ancestry that diverged from Middle Eastern populations somewhere between 38,000 and 18,000 years ago — a “back-to-Africa” migration predating the Holocene. It forms a gradient that is strongest in the west and thins toward Egypt. And its most extreme expression anywhere in North Africa is Tunisian: the Amazigh of Chenini, in the far south, carry very nearly 100% of it, with no substantial evidence of admixture from sub-Saharan, European or Near Eastern sources.

That is why Chenini turns up repeatedly in the literature as the reference population for North African ancestry. When a geneticist anywhere in the world needs a proxy for “indigenous North African”, they often reach for a Tunisian village of a few hundred people in Tataouine governorate.

The Studies, and What Each One Found

Study	What it sampled	Key finding for Tunisia
Henn et al., PLoS Genetics, 2012	~730,000 SNPs; 125 individuals across 7 North African populations, including 18 from Tunisia	Identified the autochthonous Maghrebi component; Tunisian Amazigh (Chenini) carry close to 100% of it, the highest recorded anywhere
Fadhlaoui-Zid et al., American Journal of Physical Anthropology, 2011	Y-chromosome markers in five Tunisian groups: three Amazigh isolates, one Andalusian, one cosmopolitan Arab	E-M81 at 71% overall, and fixed at 100% in Chenini–Douiret and Jradou; Near Eastern J-M267 above 30% in some groups; no major sub-Saharan or European paternal input
Arauna et al., Molecular Biology and Evolution, 2017	~900,000 SNPs; 5 Amazigh and 6 Arab groups across the Maghreb	No strong genetic difference between “Berbers” and “Arabs”; admixture peak around the 7th century CE matching Arabization; sub-Saharan gene flow traced from the 1st century BCE onward
Elkamel et al., Scientific Reports, 2021	Y-chromosomes from Kairouan, Wesletia and four Andalusian towns (Testour, Slouguia, Qalaat El Andalous, El Alia)	Tunisian paternal pool: 71.67% North African, 18.35% Middle Eastern, 5.2% sub-Saharan, 3.45% European, 1.33% Asian. Kairouan’s “Arabs” are 73.47% E-M81 — arabized Imazighen. Wesletia is the opposite: 71.8% Middle Eastern, the highest ever recorded in North Africa
Lucas-Sánchez, Fadhlaoui-Zid & Comas, Human Genetics, 2022	New genome-wide Tunisian and Moroccan samples	Two distinct trans-Saharan admixture events: one around the 13th–14th century with a Senegambian-like source, another around the 17th century involving Tunisians specifically and an East African source related to present-day South Sudanese and Kenyan populations

A note on one paper readers may encounter: a 2020 genome-wide study of southern Tunisia by Anagnostou and colleagues, which examined Douz, the R’Baya and several Amazigh villages, was subsequently retracted. Its findings should not be cited, and we do not rely on them here.

“Berber” and “Arab” Are Not Two Genetic Populations

This is the finding that most contradicts the viral map, and it is the one the literature is most united on.

The map colours Tunisia as a blend of a Berber population and an Arab population, as though two distinct peoples had mixed in varying proportions from town to town. The genome-wide evidence does not support that picture. Arauna and colleagues, comparing five Amazigh and six Arab groups across the Maghreb with around 900,000 markers, found no systematic genetic difference between them and no correlation between genetic clustering and ethnic label. The Barcelona group that led the work summarised it bluntly: Berbers and Arabs are not as genetically different as was thought.

What the data show instead is that Amazigh and Arab are cultural and linguistic categories layered over a largely shared genetic substrate. The Arabization of the seventh century was overwhelmingly a cultural and linguistic process. A real but limited Middle Eastern genetic input accompanied it, and the great majority of Tunisians who today identify as Arab are the descendants of Imazighen who adopted Arabic.

The clearest single demonstration comes from Kairouan — the holy city, founded by the Arab conquest itself, whose population is unambiguously Arab in identity. Its men are 73.47% E-M81, the autochthonous North African paternal lineage: genetically, they cluster with Amazigh groups. Fifty kilometres away, the village of Wesletia carries 71.8% Middle Eastern paternal lineages, the highest figure recorded anywhere in North Africa. Two Arab populations in the same region, with opposite genetic profiles. The label tells you nothing; the village tells you a great deal.

The corollary is that the real axis of Tunisian genetic variation is not ethnicity but isolation. Villages that married inward for centuries — Chenini, Douiret, Jradou — are homogeneous and divergent, with long runs of homozygosity and small effective population sizes. Towns on trade routes and coastlines are mixed. Tunisian Amazigh groups differ more from each other than the average “Berber” differs from the average “Arab”; Sened, 120 kilometres from the other southern villages, is the most diverse of them. Our guide to Tunisian Arabic describes the same layering in the language, where Amazigh substrate, Arabic grammar and French, Italian, Spanish and Turkish loanwords sit in a single sentence.

Fathers and Mothers Tell Different Stories

One of the most striking results in Tunisian genetics is an asymmetry between the paternal and maternal lines, and it is invisible on any admixture map.

The Y chromosome, passed father to son, is overwhelmingly indigenous. E-M81 — the North African lineage that emerged in the northwest and is essentially absent outside it — runs at around 71% across Tunisian populations and reaches fixation, 100%, in Chenini–Douiret and Jradou. Fadhlaoui-Zid’s survey found no major sub-Saharan or European paternal contribution anywhere in the samples.

Mitochondrial DNA, passed mother to child, tells a different story. The maternal pool is a mosaic: North African U6 lineages alongside European H, Middle Eastern lineages, and a substantial sub-Saharan L component that reaches around 27% in the Amazigh of Sened. Where the fathers are almost entirely local, a meaningful share of the mothers are not.

The explanation is uncomfortable and well documented. Gene flow into Tunisia was disproportionately female and disproportionately coerced — the trans-Saharan slave trade moved far more women into North African households than men, and those women’s lineages persist in the mitochondrial record while leaving no trace on the Y chromosome. The pattern is consistent across the Maghreb. Any honest account of Tunisian ancestry has to include it, and a map of coloured pie charts cannot.

The Sub-Saharan Component: Two Migrations, Two Dates

The original article attributed sub-Saharan ancestry vaguely to “trans-Saharan trade routes”. The dating work is much more precise than that, and the answer is that it is recent.

Across North Africa, most sub-Saharan gene flow occurred within the last 700 years. Lucas-Sánchez, Fadhlaoui-Zid and Comas, working with new Tunisian genome-wide samples, resolved it into two separate events with different sources:

• The 13th–14th centuries, involving a western source genetically similar to present-day Senegambian populations — the classic trans-Saharan caravan route from the Niger bend.
• The 17th century, a specifically Tunisian event involving an eastern source related to present-day South Sudanese and Kenyan populations — a different trade network entirely, running up through the Fezzan.

The quantities vary enormously between individuals. Modelling of the Tunisian sample as a whole infers a minor sub-Saharan-like source contributing around 23%, but that average conceals a wide spread: some Tunisians carry almost none, while a minority carry more than half. Across North Africa as a whole the published range runs from roughly 1% to 55%.

This variation is geographic and historical rather than ethnic. Southern oasis towns on the caravan routes carry more; isolated mountain villages that sat off the routes carry strikingly little, which is precisely why Chenini looks the way it does.

Europe, Anatolia, and the Caucasus

The European component is real, modest, and older than the French protectorate. It shows more strongly in the coastal north than inland, and it has at least three distinct sources: ancient Mediterranean contact, the Andalusian refugees expelled from Spain who founded towns such as Testour, Slouguia and Qalaat El Andalous in the seventeenth century, and the Italian, Maltese and French communities of the nineteenth and twentieth. On the paternal side it is small — around 3.45% of Tunisian Y-chromosomes — which again points to gene flow that was not symmetrical between the sexes.

There is also a much older European trace. Ancient DNA from an individual buried at Djebba in northern Tunisia around 8,000 years ago carries ancestry from European hunter-gatherers, probably reflecting movement across the Strait of Sicily in the early Holocene. The Mediterranean has been a road, not a wall, for a very long time.

The “Anatolian” and “Caucasian” slices on the viral map need correcting, and a reader got there before we did.

The Ottoman period did leave a genetic mark, but not through mass settlement — Ottoman rule in Tunisia was thin and administrative. The mark comes from a specific institution: the mamluks. From the seventeenth century until the 1880s, the beys of Tunis staffed their household and their state with enslaved men bought as boys in Circassia, Georgia and the Balkans, converted, trained and promoted into the highest offices — Mustafa Khaznadar and Khayr al-Din among them. Perhaps 250 to 300 mamluks were in Tunisia at any time. Women from the same regions entered the Bardo as concubines and wives, and several beys were the sons of Circassian mothers.

So the Caucasian signal is not, as the original article suggested, the residue of “nomadic populations from the Caucasus” or of ancient Mediterranean contact. It is the Husaynid court. It is concentrated in Tunis and the administrative centres, it is small, and it has a precise and documented history — one our readers knew better than our text did.

What This Does Not Mean

Four caveats, because ancestry percentages are among the most misread numbers in science.

They are population averages, not personal identities. A figure for Sfax tells you about a sample of men from Sfax. It tells you nothing about any particular Sfaxien, and the variation between individuals in one town routinely exceeds the difference between towns.

The components are models, not peoples. “Maghrebi” is a statistical cluster that geneticists name after the region where it is most common. It is not a tribe, it does not correspond to anyone who ever called themselves anything, and at a different value of K the software would draw the boundaries differently. This is the deepest flaw in the viral map: it presents modelling choices as biological facts.

Sample sizes are small and sampling is uneven. Henn’s landmark study used 18 Tunisian individuals. Several of the most-cited “Tunisian Berber” figures come from one or two villages of a few hundred people, chosen precisely because they are unrepresentative isolates. Coastal cities and the Tunis conurbation — where most Tunisians actually live — are comparatively under-sampled.

And ancestry is not identity. The genetics make this point more forcefully than any argument could: Kairouan’s Arabs are genetically Amazigh, and it would be absurd to tell them they are something other than what they are. Tunisian identity has been built over three thousand years out of Phoenician, Amazigh, Roman, Arab, Andalusian, Ottoman, Jewish, African and European contributions, and the country’s own self-understanding — described in our guide to the cultural aspects worth knowing — has rarely required picking one. The DNA, read honestly, says the same thing.

The Bottom Line

Tunisians are, in the main, the descendants of the people who have been in North Africa since before recorded history. On top of that foundation sit a substantial Middle Eastern layer from the Arab expansion, a modest and mostly coastal European layer, a recent sub-Saharan layer that arrived through the caravan trade in two distinct waves, and a small, precisely-sourced Ottoman-Caucasian trace from the mamluk households of the Bardo.

The layers are unevenly distributed, and the thing that predicts them is not whether a town calls itself Arab or Amazigh but how isolated it has been. A mountain village in Tataouine and a port in Bizerte have had different histories, and their genomes record it.

That is a more interesting story than the map tells, and it has the advantage of being true. For how that history shows up in daily life rather than in a genome, see our guides to Tunisian Arabic and life in Tunisia.

Sources

• Henn, B. M. et al. (2012). Genomic Ancestry of North Africans Supports Back-to-Africa Migrations. PLoS Genetics 8(1): e1002397.
• Fadhlaoui-Zid, K. et al. (2011). Genetic structure of Tunisian ethnic groups revealed by paternal lineages. American Journal of Physical Anthropology.
• Arauna, L. R. et al. (2017). Recent Historical Migrations Have Shaped the Gene Pool of Arabs and Berbers in North Africa. Molecular Biology and Evolution 34(2): 318–329.
• Elkamel, S. et al. (2021). Insights into the Middle Eastern paternal genetic pool in Tunisia: high prevalence of T-M70 haplogroup in an Arab population. Scientific Reports 11.
• Lucas-Sánchez, M., Fadhlaoui-Zid, K. & Comas, D. (2022). The genomic analysis of current-day North African populations reveals the existence of trans-Saharan migrations with different origins and dates. Human Genetics.
• Oualdi, M. (2011). Esclaves et maîtres. Les mamelouks au service des beys de Tunis. Publications de la Sorbonne — on the Caucasian and Balkan mamluks of the Husaynid court.

---

From the Carthage Magazine Bookshelf

The genome records three thousand years of arrivals; these three are about what those arrivals built.

• All About Tunisia — the definitive English-language traveler’s guide. 572 pages, 27 chapters, all nine regions, every UNESCO inscription, five thematic trails — and the practical answers (visa, currency, transport, etiquette) most travelers wish they’d had on the plane. $24.99 · PDF & EPUB.
• Speak Like a Local — 200+ Tunisian Arabic phrases with native audio recorded in Tunis. The phrases for the taxi, the souk, the café, and the dinner table. $14.99 · PDF, EPUB, MP3.
• The Authentic Tunisian Cookbook — sixty traditional recipes from the heart of North Africa. For when you get home and find yourself missing the food. $9.99 · PDF & EPUB.

All three available as a bundle for $39.99 — guide, language, and food, delivered together.