Genetic History of the Levant — Part A

Recently there has been a surge of interest in Levantine genetics. Here is a summary of the genetic history of the Levant in two parts.

Natufian Culture (13,000–9,000 BCE)

The Natufian culture was a hunter-gatherer society with the remarkable feature of being sedentary or semi-sedentary. The main Y-DNA haplogroups associated with Natufians include archaic forms of E1b and CT lineages.

Levant Neolithic (9,000–4,500 BCE)

The Neolithic in the Levant is divided into two phases: Pre-Pottery Neolithic (PPN) and Pottery Neolithic.

Agriculture spread to the southern Levant from more northern regions. The core area where farming was first developed was in the Urfa region (Portasar / Göbekli Tepe) near the modern Syrian–Turkish border. Genetic data show an important northward shift in autosomes in the southern Levant during this period, particularly visible in PPNB samples from Israel and Jordan.

A strong genetic indicator of this migration is the haplogroup E1b-Z1919, which has two major branches:

• E-L618, which moved to Europe via Anatolia and today is common in the Balkans.

• E-V22, which moved south into Egypt and introduced farming there. Today it is common in northern Egypt.

Another important lineage from this period is E-M123, which today is mainly found in West Asia but also appears in Central Asia. Sporadic occurrences of H2 and T1a are also documented.

Copper Age / Chalcolithic (4,500–3,300 BCE)

There are relatively few samples from this period. The limited available data, associated with the Ghassulian culture, show another autosomal shift toward northern Iraq.

This shift is consistent with archaeological evidence indicating the emergence of new burial traditions. At present, it is difficult to determine exactly which cultural phenomenon caused this change in the southern Levant during the Chalcolithic period.

A similar shift toward northern Mesopotamia had already begun earlier in the northern Levant (Tell Kurdu). A plausible cultural candidate connected with this process is the Halaf culture.

The presence of T1a1a-L208 during this period also supports the idea that new groups were involved that were not directly descended from earlier Neolithic populations.

Early Bronze Age (3,300–2,200 BCE)

The transition from the Chalcolithic to the Early Bronze Age in the Levant was a period of dramatic cultural change. Many old settlements were abandoned, while new mobile pastoralist groups spread throughout the region. These groups are usually identified with early Semitic tribes.

The autosomal change during this period is relatively subtle, showing a slightly more eastern genetic shift. However, the change in Y-DNA is much more pronounced. The frequency of E1b lineages declines significantly, while J1-Z1853 becomes the most common haplogroup from this period onward. J2b1 also increases in frequency during this time.

The exact location of the Proto-Semitic homeland remains debated. It is often placed in the Arabian Peninsula, and the slight increase in Natufian ancestry does support a southern origin. However, the appearance of additional ancestry related to historic Armenia suggests a region closer to Mesopotamia (see the map), where J1 may have already been present in earlier periods. Ultimately, J1 appears to originate further north in the Taurus–Zagros–Caucasus mountain region.

At present, ancient DNA evidence is still insufficient to precisely determine the Proto-Semitic homeland or the exact migration path of J1.

Wherever the Proto-Semitic homeland was located, Semitic populations expanded widely:

• northward to the Taurus Mountains,

• eastward to the Zagros Mountains,

• and southward across the Red Sea, reaching Ethiopia around 1000 BCE.

The earliest branch of the Semitic language family is East Semitic, represented by Eblaite and Akkadian, both now extinct. These populations settled in northern Levant and Mesopotamia, where their culture merged with that of the Sumerians and likely contributed to the end of the Uruk cultural system around 3100 BCE.

The Akkadians later established one of the earliest empires in history, uniting Mesopotamian city-states around 2330 BCE. This empire collapsed around 2150 BCE.

After the 2200 BCE climatic event and the fall of the Akkadian Empire, a new period began in the Levant, which will be discussed in Part B.

Alashkert Armenians

Alashkert Armenians

Alashkert Armenians display quite remarkable genetic characteristics. In FST analyses, they appear to be the closest to three geographically distant Armenian groups: Artsakh, Sasun, and Hamshen Armenians.

It is difficult to determine at this stage whether this pattern reflects their central or “average” position within Armenian regional genetic diversity, or whether it has deeper historical causes.

It should also be noted that a negative FST value effectively means zero, indicating the absence of any meaningful genetic distance between the compared populations.

 

Genetic History of Greece

Yediay et al. (2024) published a large number of ancient DNA samples from Bronze Age Greece, which helps clarify the origins of the Greek population. If we set aside the still poorly sampled Paleolithic period, Greece appears to have experienced three major migrations in prehistoric times.

1. Neolithic Migration (around 7000 BCE)

The first migration involved Neolithic farmers from Anatolia, ultimately originating from the Urfa region. These farmers moved into Greece around 7000 BCE, introducing agriculture. Farming appeared in Greece earlier than in any other part of Europe, and from there agricultural populations expanded northward into the rest of the continent.

2. Chalcolithic Migration (around 4000–3000 BCE)

A second migration arrived from regions associated with historic Armenia during the Chalcolithic period. This movement is associated with the Chaff-Faced Ware (CFW) cultural horizon in Anatolia. In Greece, this migration had its strongest impact in Crete and the Cycladic islands of the Aegean and may be connected with the formation of Minoan civilization (Linear A culture). Heggarty (2023), however, attempted to associate this migration with the origins of the Greek language.

This migration introduced several new haplogroups and may also be partly responsible for the so-called Mediterranean substrate present in both Greek and Armenian. A linguistic example is Armenian iwl / el (իւղ, եղ) meaning “oil,” which can be compared with Greek elaion (ἔλαιον) meaning “olive oil.”

3. Steppe Migration (around 2200 BCE)

The third migration came from the Pontic–Caspian Steppe around 2200 BCE. Similar to developments in Armenia, the transition from the Early Bronze Age to the Middle Bronze Age in Greece is associated with migrations from the steppe.

These migrants introduced haplogroups such as R1b-Z2103 and R1b-PF7562, along with smaller numbers of I2 and J2b2a. Mycenaean Greeks also carried significant levels of J2-Z6055, a lineage that had already been present in Greece since the Neolithic. Its high frequency in the Bronze Age likely reflects a founder effect.

Mycenaean Period (1700–1100 BCE)

After a relatively short Middle Bronze Age period, steppe ancestry becomes diluted during the Late Bronze Age (1700–1100 BCE), which corresponds to the Mycenaean civilization (see chart).

The Linear B script has been deciphered and represents an early form of the Greek language. Mycenaean Greeks possessed ancestry both from local Neolithic farmers and from earlier migrants arriving from more eastern regions. In genetic charts, this eastern component is sometimes labeled Anatolia_C, which is related to populations associated with Chaff-Faced Ware cultures of historic Armenia.

Yediay et al. (2024) connect the steppe migration with Indo-European speakers associated with the Yamnaya horizon, thus providing a convincing explanation for the origins of the Greek population.

Late Bronze Age Collapse and Greek Expansion

Around 1200 BCE, the Mycenaean civilization collapsed, forcing many populations to migrate. During this period, several groups moved into the eastern Mediterranean and became known as the Sea Peoples. One of these groups eventually settled in Cyprus, where they introduced the Greek language.

Even earlier, Greeks had already begun settling along the western coast of Anatolia. In Hittite texts, these groups may correspond to the Ahhiyawa. Numerous conflicts were fought over control of these coastal regions. The memory of these wars is preserved in the Greek epic poems the Iliad and the Odyssey.

Modern Greek is generally considered to derive from the Attic–Ionic dialect group.

Attempt to redo the models from recent papers.

 

Attempt to Redo the Models from Recent Papers

Modeling Armenia LBA as a simple mixture of Kura-Araxes and Catacomb fails: the p-value is too low (pic. 1).

Adding an Anatolian population such as İkiztepe_C improves the p-value. This approach was used in Skourtanioti (2024), and the model becomes feasible (p ≈ 0.2).

Using Çamlıbel Tarlası_LateC instead of İkiztepe further improves the p-value. This was done in Yediay (2024), yielding p > 0.3.

Replacing the Anatolian component with Leyla Tepe (from what is now Azerbaijan) improves the model much more (p > 0.7). However, the standard errors increase substantially because Kura-Araxes and Leyla Tepe are too close genome-wide. To reduce the standard errors, I need to adjust the settings (right populations). That will require some time, but I think it is already quite clear what is happening here.

Finally, when I add Ukraine Neolithic hunter-gatherers (Ukraine_N HG) as an additional source, the p-value rises to ~0.9, i.e., close to certainty. The standard errors also improve slightly, although they still require further refinement.

Conclusions

The populations that contributed to the formation of the Trialeti–Vanadzor culture (TVC) likely came from what is now the Krasnodar region, where they acquired a minor Ukraine_N HG component. They then moved via Dagestan, mixed with older South Caucasus Chalcolithic groups in the Kura-Araxes lowlands, and subsequently split into two branches:

• One branch moved upstream into the Kura–Debed region and formed the TVC.

• The other moved into the Urmia and Van basins and formed the Van–Urmia culture, although their genetic impact there appears to have been lower than in the TVC.

The Y-DNA associated with these groups should therefore be sought in the Krasnodar region, especially lineages such as L584, I2a2b, and PF331. Identifying Y4364 will be more difficult. It is even possible that the true homeland of Proto-Yamnaya was also located there.

Comments on the Chart Before Reviewing Yediay et al. (2024)

Comments on the Chart Before Reviewing Yediay et al. (2024)

Before reviewing the recent Yediay et al. (2024) preprint, a few comments about this chart are necessary.

Armenia Middle and Late Bronze Age samples clearly show ancestry from three main sources.

1. Kura–Araxes ancestry

The first component comes from Armenia EBA, associated with the Kura–Araxes culture. This is expected and does not require further explanation.

2. “Anatolia_C” ancestry

The second component is labeled Anatolia_C. At first glance this may seem surprising, but it actually has a straightforward explanation.

The Anatolia_C component largely derives from Neolithic and Chalcolithic populations of historic Armenia, associated with the Chaff-Faced Ware (CFW) cultural horizon. These groups lived before the rise of the Kura–Araxes culture and were also present in what is now Azerbaijan, where they are known archaeologically as the Leyla Tepe culture.

When the Caucasus hunter-gatherer–shifted Kura–Araxes culture expanded from its homeland in the South Caucasus, it did not completely replace these earlier populations. In many areas, they continued to live alongside the Kura–Araxes communities.

When steppe ancestry arrived from the north about 4,500 years ago, these groups most likely entered the region through what is now Azerbaijan, where they encountered the remaining populations associated with the Leyla Tepe culture. After mixing with them and acquiring this Anatolia_C-like ancestry, they moved toward the upper Kura–Debed river region, where they mixed with populations related to Armenia EBA.

Another possible source of Anatolia_C-like ancestry may have been populations living in the southern parts of the Araxes plain.

This pattern was also noted in Skourtanioti et al. (2024). However, the authors interpreted it incorrectly, proposing two separate migrations occurring during the same period—one from Anatolia and another from the steppe. Genetic bloggers had already pointed out earlier that this explanation is unlikely, and Davidski even opened a dedicated discussion thread on the issue.

3. Apparent CWC ancestry

The third component is the so-called CWC ancestry. This is almost certainly not a real signal.

The Corded Ware culture (CWC) is strongly associated with R1a, yet no R1a lineages have been found in Middle and Late Bronze Age samples from the South Caucasus (including Armenia and Georgia).

The reason CWC appears in the model is likely that the steppe groups who migrated into the South Caucasus carried a small amount of WHG-related ancestry. This additional WHG / Ukraine Neolithic hunter-gatherer affinity was already present east of the Azov region even before the formation of the Yamnaya culture.

This additional component could also explain the appearance of I2a2b in ancient Armenia.

It is important to note that CWC is genetically similar to Yamnaya, but it contains roughly:

• about 10% additional UNHG-related ancestry

• about 20% Euro-Anatolian farmer ancestry

Even a very small 1% WHG introgression into a Yamnaya-like population can create a statistical signal resembling 10% CWC ancestry in modeling. Given that MLBA samples already contain excess Anatolian ancestry, it is not surprising that the calculator interpreted this mixture as CWC rather than Yamnaya.

Thus, the CWC component in this case is an artifact of the modeling, although it reflects real underlying genetic processes.

Urartian period samples

A similar situation can be observed in Urartu-period samples, where the Anatolia_C component is higher. This indicates a stronger Neolithic-derived ancestry, while the steppe component is lower.

There is one exception—an outlier individual, who appears to have been a migrant from Etiuni.

A Greek Sample in the Armenian Genetic Cluster

A Greek Sample in the Armenian Genetic Cluster

One of the Greek samples in Hovhannisyan et al. (2024) falls within the modern Armenian cluster on the PCA. I was unable to find detailed information about this individual, except that the Greek DNA used in the paper was taken from Lazaridis et al. (2014, Nature).

Based on its position on the PCA, the sample does not appear to belong to Cappadocian Greeks. It also does not resemble Trabzon Greeks. The most likely possibility is that the individual belongs to the Urum population from southwestern Georgia.

The Urums settled in the Tsalka region of Georgia after the Russian–Turkish wars of the 19th century. We have one Urum individual in our group whose DNA appears similar to that sample. Of course, this remains only a hypothesis and requires further verification.

It should also be noted that Urums should not be confused with the Urumu tribe of the Iron Age. The word Urum derives from Rome / the Roman Empire, whereas the name Urumu has an uncertain origin, although it may possibly be related to Aramu.

In the second PCA, the main modern Greek cluster can be observed, clearly separate from modern Armenians. Compared with Mycenaean-period Greeks, the modern Greek cluster appears somewhat shifted toward the north.

Another Greek sample appears to be almost certainly of Anatolian origin, as it plots approximately between the Armenian and Greek clusters.