The Maghreb has the oldest directly sequenced population in Africa outside the Nile, and it is a population that still lives there. The foragers buried at Taforalt in eastern Morocco fifteen thousand years ago carried an ancestry that survives, diluted but unmistakable, in Amazigh and Arabic-speaking Moroccans, Algerians, Tunisians and Libyans today.
Ancestry is not identity, and this post keeps the two apart. A genome can show which ancient populations a person's ancestors resembled; it cannot say whether they spoke Tamazight or Arabic, what they believed, or what they called themselves. There is no "Berber gene", no purer or less pure Maghrebi, and no ranking in any number below.
The short answer: Maghrebi genomes rest on a deep local layer, the Iberomaurusian ancestry of Taforalt, itself a blend of Near Eastern and African streams. Farmers from Iberia and the Levant added European and Levantine Neolithic ancestry after 5500 BC; Phoenician, Roman and Arab-era contacts added eastern Mediterranean and Arabian ancestry; and trans-Saharan contact added sub-Saharan African ancestry along a gradient rising from the coast toward the desert. The Canary Islands' indigenous genomes preserve the pre-Islamic version of this mixture.
Taforalt: the Iberomaurusian foundation#
In 2018 van de Loosdrecht and colleagues published genomes from Grotte des Pigeons at Taforalt, dated to about 15,000 years ago and associated with the Iberomaurusian stone tool industry. The Taforalt people were not related to the European hunter-gatherers across the strait; their closest ancient relatives were the Natufians of the Levant, and the authors modelled them as roughly two thirds Natufian-related ancestry and one third an African ancestry with no good ancient match. Their Y chromosomes belonged to haplogroup E-M78 and their mitochondrial lineages to U6 and M1, both still characteristic of North Africa.
This is the deep layer of the Maghreb, old enough that "Near Eastern" and "African" must be read as fifteen-thousand-year-old streams rather than later migrations. The Green Sahara genomes from Takarkori belong to a related lineage.
The Neolithic: farmers from Iberia and the Levant#
Fregel and colleagues (2018) sequenced Early Neolithic individuals from Ifri n'Amr or Moussa, about 5000 BC, and found them essentially continuous with Taforalt: the first Moroccan farmers were, genetically, the local foragers. Their Late Neolithic individuals from Kelif el Boroud, around 3000 BC, carried a large share of ancestry related to Iberian Early Neolithic farmers, the Anatolian farmer stream arriving by sea.
Simões and colleagues (2023) filled in the sequence. At Kaf Taht el-Ghar near Tetouan, the earliest farmers of about 5500 BC were mostly of Iberian Neolithic ancestry. Later, individuals from Skhirat-Rouazi showed a distinct Levantine-related ancestry arriving along the coast with pastoralism. By the Late Neolithic the three streams had merged, and that mixture of local Iberomaurusian, Iberian farmer and Levantine ancestry is the profile the historical period inherited.
A dated timeline for ancestry in the Maghreb#
- About 15,000 years ago: the Iberomaurusian foragers of Taforalt.
- About 5500 to 5000 BC: Iberian farmer migrants at Kaf Taht el-Ghar; local continuity at Ifri n'Amr or Moussa.
- About 4800 BC: Levantine-related pastoralists at Skhirat-Rouazi.
- About 3000 BC: the Late Neolithic mixture at Kelif el Boroud.
- 800 to 150 BC: Phoenician and Carthaginian settlement on the coast.
- 150 BC to 430 AD: Roman Africa and Mauretania; Numidian and Mauretanian kingdoms.
- First centuries AD to the 1400s: the indigenous Canary Islanders, sealed off before Islam.
- Seventh to eleventh centuries: the Arab conquests and the Banu Hilal migrations.
- Eighth century onward: trans-Saharan trade brings sub-Saharan African ancestry, unevenly.
The Guanche window on pre-Islamic Berber ancestry#
The Canary Islands were settled from North Africa around the start of the first millennium AD and isolated until the fifteenth century, so their indigenous inhabitants are the best available sample of a Berber population before the Arab era. Fregel and colleagues (2019) showed that their mitochondrial lineages were North African, with U6 prominent beside European and Near Eastern lineages that had already reached the Maghreb in the Neolithic. Serrano and colleagues (2023) published genome-wide data from dozens of individuals across the archipelago and modelled them as a North African population closest to the Late Neolithic Moroccans, with a minor sub-Saharan contribution already present. They show that the Neolithic mixture was still the profile of Berber-speaking communities a millennium later, and that a sub-Saharan component existed at low levels before the Islamic trans-Saharan trade.
Punic, Roman and Arab-era contacts#
The Carthaginian coast has now been sampled directly, and the finding, discussed in our post on Punic DNA, is that Punic genomes drew mostly on Sicilian, Aegean and North African ancestry rather than on the Levant. Roman Africa brought eastern Mediterranean and Italian ancestry into the coastal cities, while inland populations retained the local profile.
The Arab conquests of the seventh century and the Banu Hilal migrations of the eleventh brought Arabian-related ancestry that is visible in modern genomes. Genome-wide analyses (Henn et al. 2012; Arauna et al. 2017) find a Middle Eastern component across the Maghreb that is somewhat higher in Arabic-speaking than in Amazigh-speaking communities, but the difference is modest and geography predicts ancestry better than language. Arabisation was mostly linguistic.
Sub-Saharan African ancestry and its gradient#
Every Maghrebi population carries some later sub-Saharan African ancestry. Henn and colleagues (2012) found it ranging from a few percent in parts of Tunisia and northern Morocco to a fifth or more in southern Moroccan and Saharan communities, and Arauna and colleagues (2017) dated most of it to the last thousand years, the era of the trans-Saharan trade routes. The gradient runs coast to desert and applies to Arabic and Amazigh speakers alike: it follows where a community lives, not what it speaks.
What the Y-DNA and mitochondrial DNA add#
The paternal lineage of North Africa is haplogroup E-M81, nearly confined to the Maghreb, which reaches very high frequencies among Amazigh-speaking communities, on the order of two thirds to four fifths of men in some Moroccan and Algerian Berber groups and roughly half of men in the general populations of Morocco and Algeria (Arredi et al. 2004; Bekada et al. 2013). Haplogroup J1, associated with Arabia, is the second lineage and is more frequent in Arabic-speaking and eastern populations; E-M78, the Taforalt lineage, and R1b-V88 are minor.
Maternal lineages are more mixed: the Taforalt-era U6 and M1 persist at moderate frequencies, European and Near Eastern lineages such as H, HV, J, T and K arrived with the Neolithic and later, and sub-Saharan L lineages account for roughly a tenth to a quarter of mitochondrial DNA depending on the region. E-M81 is a marker of shared paternal history, not a test of identity.
Limitations#
The ancient record of the Maghreb has large gaps. Nothing genome-wide has been published from the Numidian kingdoms, little from Algeria and Libya at any depth, the Neolithic transect is Moroccan, and the Arab era on the mainland is unsampled. The African third of Taforalt's ancestry still has no adequate ancient reference, so every model with a sub-Saharan source is partly measuring that ancient stream and partly the medieval one. And no result here bears on who is Amazigh: ancestry and identity are separate facts.
Frequently asked questions about Berber and Maghrebi DNA#
Are Berbers and Arabs in North Africa genetically different?#
Only slightly. Genome-wide studies find that Amazigh and Arabic-speaking Maghrebis share the same deep North African ancestry, with Arabic speakers carrying a modestly larger Middle Eastern component on average. Geography predicts ancestry better than language does, so neighbours of either language are usually closer to each other than to distant communities.
What is the oldest ancestry in the Maghreb?#
The Iberomaurusian ancestry of Taforalt, sequenced from individuals about fifteen thousand years old. It was itself a mixture of a Natufian-related Near Eastern stream and a deeply diverged African stream, and it survives as the local foundation in every modern Maghrebi genome, diluted by later additions.
How much sub-Saharan African ancestry do Maghrebis have?#
It varies with geography more than with language. Published estimates range from a few percent in parts of Tunisia and northern Morocco to a fifth or more in southern Moroccan and Saharan communities. Most of it entered within the last thousand years through trans-Saharan trade, on top of the much older African component already present at Taforalt.
Are the Guanches related to modern Berbers?#
Yes. Genome-wide data from indigenous Canary Islanders show a North African population closest to Late Neolithic Moroccans, settled from the mainland around the start of the first millennium AD. Because they were isolated before the Arab conquests, they are the best available reference for pre-Islamic Berber ancestry, and they carry the same E-M81 paternal and U6 maternal lineages.
How to model Berber and Maghrebi ancestry with qpAdm#
A qpAdm model tests whether a genome can be written as a mixture of chosen ancient sources and reports the fit with a p-value and standard errors. The Maghreb is unusually well served in the Ancestrify catalog at the deep end: Iberomaurusian (15000 - 8000 BC) is the Taforalt foundation itself, North African Farmer (5200 - 4000 BC) covers the Moroccan Neolithic, and Anatolian Neolithic Farmer (8500 - 6000 BC), Levantine Neolithic Farmer (8800 - 6500 BC) and Natufian (12000 - 8000 BC) supply the streams that arrived by sea and along the coast. For the historical period, Numidian Berber (0 - 500 AD) and Indigenous Berber (300 - 1400 AD) give a local pre-Islamic baseline, Carthaginian (800 - 150 BC) and Phoenician (850 - 50 BC) cover the Punic coast, Eastern Mediterranean (0 - 600 AD) the Roman era, and Arabian Peninsula (0 - 600 AD) the Arab-era layer.
Two streams are represented only broadly. The sub-Saharan component is a single Sub Saharan African (300 BC - 400 AD) source, which cannot distinguish West African from Nilotic ancestry and which absorbs part of Taforalt's ancient African third whenever the Iberomaurusian source is absent. And there is no medieval mainland Maghrebi source, so the Arab-era mixture is inferred rather than observed. Read the model record for which sources the data require. A Global25 analysis is a good first pass, placing a Maghrebi genome along the coast-to-desert gradient before any formal test.
Sources and further reading#
- van de Loosdrecht, M. et al. (2018). Pleistocene North African genomes link Near Eastern and sub-Saharan African human populations. Science 360, 548-552.
- Fregel, R. et al. (2018). Ancient genomes from North Africa evidence prehistoric migrations to the Maghreb from both the Levant and Europe. PNAS 115, 6774-6779.
- Simões, L. G. et al. (2023). Northwest African Neolithic initiated by migrants from Iberia and Levant. Nature 618, 550-556.
- Serrano, J. G. et al. (2023). The genomic history of the indigenous people of the Canary Islands. Nature Communications 14, 4641.
- Henn, B. M. et al. (2012). Genomic ancestry of North Africans supports back-to-Africa migrations. PLOS Genetics 8, e1002397.
- 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, 318-329.
Editorial note: the hero artwork in this article was generated with AI as conceptual illustration. It does not reproduce a scientific figure, an ancient individual or a measured migration route.
