Some of you have asked me about a new paper on East Africa, Ancient DNA reveals a multistep spread of the first herders into sub-Saharan Africa.The reality is some of you know this topic better than I, so I don’t have much original to add. But, I was curious today when a preprint dropped, West Asian sources of the Eurasian component in Ethiopians: a reassessment. This is out of Luca Pagani’s group, and it was he who published 2012’s Ethiopian Genetic Diversity Reveals Linguistic Stratification and Complex Influences on the Ethiopian Gene Pool. I spent a fair amount of time at his ASHG poster talking to him about his admixture estimate. He found which West Eurasian ancestry in Ethiopians (highland) which “may represent gene flow into Africa, which we estimate to have occurred ∼3 thousand years ago (kya).”
The West Eurasian ancestry into Ethiopians surprised no one. Rather, the relatively recent estimate of admixture was surprising. The Bible and Homer refer to “Ethiopians.” These mentions date to a period between 500 and 1000 BCE (even if the events predate that). My question was simple, “who were the Ethiopians mentioned by the Greeks and the Hebrews if the genetic character of the people who we today all Ethiopians only came into being ~3,000 years ago?”
Obviously, the term “Ethiopian” can be used in a more generic sense than for the people of Ethiopia proper. The Greeks confused Indians and Ethiopians because both were dark-skinned peoples, and presumably black African people of non-Ethiopian origin were sometimes identified as Ethiopian in Antiquity.
Let me quote the abstract of the preprint:
Previous genome-scale studies of populations living today in Ethiopia have found evidence of recent gene flow from an Eurasian source, dating to the last 3,000 years. Haplotype and genotype data based analyses of modern and ancient data (aDNA) have considered Sardinia-like proxy, broadly Levantine or Neolithic Levantine populations as a range of possible sources for this gene flow. Given the ancient nature of this gene flow and the extent of population movements and replacements that affected West Asia in the last 3000 years, aDNA evidence would seem as the best proxy for determining the putative population source. We demonstrate, however, that the deeply divergent, autochthonous African component which accounts for ~50% of most contemporary Ethiopian genomes, affects the overall allele frequency spectrum to an extent that makes it hard to control for it and, at once, to discern between subtly different, yet important, Eurasian sources (such as Anatolian or Levant Neolithic ones). Here we re-assess pattern of allele sharing between the Eurasian component of Ethiopians (here called NAF for Non African) and ancient and modern proxies area after having extracted NAF from Ethiopians through ancestry deconvolution, and unveil a genomic signature compatible with population movements that affected the Mediterranean area and the Levant after the fall of the Minoan civilization.
The big issue, which they acknowledge in their preprint, is that assignment of ancestry to regions of the genome can cause issues. They came to have checked for this, but I’m not sure they did. One thing I did notice as well as the strong affinity to Middle Eastern groups with less recent cosmopolitan ancestry.
Remember when ancient Egyptians shared a lot of drift with Sardinians? I think due to less cosmopolitan ancestry there is a tendency for ancient or endogamous groups to exhibit affinities. Additionally, the Minoans in the sample were mostly Anatolian Neolithic, with a secondary Caucasus drift component. Perhaps the non-African in Ethiopians mimics this through different admixture pathway?
The vast majority of Ethiopians speak Afro-Asiatic languages. But they are divided about in half between those who speak Semitic languages, and those who speak Cushitic languages. Semitic languages are well known, from Hebrew to Akkadian to Arabic. Interesting, the Ethiopian Semitic languages are closer to Hebrew and Arabic than the ancient East Semitic languages, such as Assyrian and Akkadian. The divergence of Akkadian from these other languages has to be older than 4,500 years ago since that’s the oldest Akkadian texts we have. The closest languages to the Ethiopian Semitic languages are South Arabian dialects, which are going extinct in the face the advance of Arabic.
All this needs to be interpreted in the context of the earlier paper.
We infer a more recent average date (~2200 B.P.) for two late PIA individuals, likely associated with additional Sudan-related ancestry (table S11). Our power to detect multiple waves of admixture is limited with ancient data, but for one pair of PN individuals from Naivasha Burial Site, we are confidently able to identify two separate events, the first at ~5100 B.P. and the second at ~4000 B.P. We also infer two waves for a pair of individuals from Gishimangeda Cave, dating to ~6000 B.P. and ~4000 B.P. In light of our qpAdm results, and given the associated MALDER amplitudes (table S11), these multiple dates plausibly represent estimates of the times of (i) the formation of admixed ENP ancestry and (ii) admixture in eastern Africa between local foragers and descendants of the first mixture, leading to the three-component ancestry of PN individuals. In this context, the single (and intermediate) estimated dates for other PN pairs can be interpreted as averages of these two processes (Fig. 5).
If you look at the figure above, it aligns with the text. There were several movements of pastoralists into East Africa from the north. They began mixing with groups in Sudan 5 to 6 thousand years ago. Later on, they mixed with foragers to the south and east related to the Ethiopian Mota sample, and so resulting various savanna pastoralist groups. The Iraqw people of Kenya and Tanzania likely descend from these early Afro-Asiatic pastoralists. They speak a Cushitic language. I think it is likely that these early pastoralists which pushed down the Nile valley were Cushitic speakers, and therefore related to those Afro-Asiatic people who pushed west and settled the Maghreb and became Berbers.
Over the last 2,000 years, it looks like most of these Savanna pastoralist people were absorbed into expanding Nilo-Saharan speaking peoples, such as the Masai, or to some extent into Bantu peoples, such as the Kikuyu (going bypassed the results in the paper). Meanwhile, at some point, much of Ethiopia seems to have been settled by Cushitic speakers, though likely later than in the regions to the west and south.
I believe that the ~3,000-year date reported by Pagani et al. in 2012 is picking up a late admixture of Semitic peoples from Arabia with local Cushitic speakers. The language affinities with South Arabian and later historical cross-Red Sea interactions points to Yemen.
Why is there this strange affinity to Minoans and Mediterranean Jews? Rather than indications of some pulse of gene flow from this particular region, I think it’s an artifact of the reality that Arabian populations are highly mixed and cosmopolitan today. This dates to the Islamic period, but probably earlier, as north and south Arabian peoples seem to have swapped locations regularly around the arid core of the peninsula.
“Why is there this strange affinity to Minoans and Mediterranean Jews? Rather than indications of some pulse of gene flow from this particular region, I think it’s an artifact of the reality that Arabian populations are highly mixed and cosmopolitan today.”
I’m still puzzled. Why not Yemenite Jews, in that case? (Or Samaritans?) They explicitly favor a model that’s all Anatolian Neolithic + CHG, and no Levantine—which sounds totally wrong, but there must be some kind of explanation.
Why not Yemenite Jews, in that case?
using their deconvolution method the always found that the yemenite jews don’t match as well as levant populations. this was a weird early finding….
I don’t really trust this too much. Even small errors in calling hard to call elements in their de-convolution seem likely to skew the outcome, given how much sharing Levant_N and Anatolia_N have, esp, if they are shifting more diverged Basal Eurasian elements that distinguish Levant_N from Anatolia_N into the African component for their target populations.
(This is on top of where you find that small presence of deeply diverged ancestries can really knock direct f4 stats like they test in Fig 2 – Sardinians tend to score a maximum (most negative) on f4(Mbuti, Levant_N; Sardinian, Near East), despite the non-Levant_N ancestry being pretty probably fairly low for many of the Near East in question).
I don’t see any information on the proportion of deconvoluted “NAF” ancestry they estimate for their populations. Best seems to be “We demonstrate, however, that the deeply divergent, autochthonous African component which accounts for approx. 50% of most contemporary Ethiopian genomes” and that’s all you get.
given how much sharing Levant_N and Anatolia_N have, esp
yeah. you hit the nail on the head. these groups which are very similar…small artifactual issues can swamp/skew/distort our perception of relatedness i suspect.
The crazy thing is that Eastern Cushitic and Southern Cushitic share a branching relative to other Cushitic languages, probably dating back to 4,500 – 5,000 years ago – yet the ydna between these groups shifted rapidly. East Cushitic groups became E-M78 heavy, and South Cushitic groups E-Z827. Both are E-M35 lineages, but diverge close to the root of the 24,000 E-M35 mrca. But the South Cushitic samples also had E-V22, and E-V32 – and both lineages, along with E-M293, are found frequently in the Bantu and Nilotic groups in Southeast Africa.
The most clear thing that can be said now is that the Tutsi are mostly (55%) descended from these South Cushitic nomads. Their biggest ydna is E-M239, the South Cushitic lineage most the ancient Pastoral Neolithic and Elmenteitan males had.
The Arabian Ethiosemites contributed around a quarter to contemporary Ethiosemites – which is perfectly alinged with their quarter % of J1 lineages (J1-P58 + J1-P56). The early Ethiosemites mixed with highlander Central Cushitic (Agaw) rather than Somalis. The former – prior to Semitic admixture – was heavy on Omotic and Chabu-like ancestry, unlike Somalis. The very high (world’s peak?) frequency of A3b2’s Cushitic lineage in Agaws, at around 40% – belonging to a A3b2 lineage divergent from the Nilo-Saharan/Sardinian clade of A3b2 – makes sense of A3b2 making up 15-25% of Ethiosemitic males. That, and paleolithic Ethiopian ydna such as E1b1b*, E1b1c, and basal E1b1a lineages in Agaws, and present in turn in Ethiosemites, all the while these lineages being essentially non-existent in Somalis.
Concerning the ancestors of these Cushitic groups tested, both ancient and modern – the early Cushitic population lived in the Eastern Desert of Egypt and Sudan were mesolithic hunter-gatherers, who after diverging from their Semitic/Berber relatives when the latter group hived off north towards the Delta, interacted more heavily with the Chadic/Egyptian clade in the Eastern Desert and especially in the adjacent Nile region. This is probably why we see typically Cushitic E-V32 in Chadics, T1-L208 in all these groups, although T-L208 may be a late Afroasiatic phenomenon. The early Cushitic group became pastoralists early on with the other late Afroasiatic groups, only post-breakup. The early/proto-Egyptians were very similar anthropometrically to these Cushitic groups, especially seeing their strong similarities with contemporary Cushitic groups. This, and the same dominant ydna of both groups, such as E-V12, T-L208, and E-V22 in addition to the same cultural practices makes it very likely that the early/ancestral Egyptians and early Cushitic pastoralists of southern Egypt/NE Sudan were genetically very similar.
I think what Razib and others here might find most interesting is that a single language family had 3 waves of different subgroups migrate into a single area and displace the earlier linguistic subgroup with which it shares a common language family. The relevant scenario took place in the Horn in regards to the Afroasiatic language family.
First we have the Omotic wave. The Omotic subgroup is firmly considered to be the most basal Afroasiatic subgroup, and they broke off from the other branch of Afroasiatic (Late Afroasiatic) 13,000 years ago. The proto-Omotic and early Omotics were hunter-gatherers who displaced and likely mixed with the earlier preceding and pre-Afroasiatic paleolitic hunter-gatherers of the Horn, correlated with intrusive Egyptian/Sudanese hunter-gatherer cultural layer in northern Ethiopia/Eritrea before 7,000 years ago.
Then around 7-5,000 years ago, septs of Cushitic pastoralists start migrating into the Horn and, in turn, displacing the earlier Omotic hunter-gatherers/incipient farmers from northern Ethiopia/Eritrea and driving them further south into the current SNNPR and central Ethiopia. By around 4,000 years ago, the Ethiopian Highlands was agropastoralist Central Cushitic (Agew) territory.
Then around 3,500/3,000 years ago, West Semitic Ethiosemites migrate from Arabia into northern Ethiopia/Eritrea, and heavily admix and assimilate the Central Cushitic Agew agropastoralist host. This wave probably ushured in the Ethiopian Iron Age, and a new agricultural toolkit.
I would say the Ethiopians of the Greek/Romans were Nilo-Saharans, not the Ethiosemitic groups. It essentially referred to the Kushites – the Eastern Sudanic conquerors of Egypt and Levant, of 25th dynasty fame – who the Ancient Egyptians depicted as looking exactly like contemporary Eastern Sudanics, with red braids, prognathic mouths, and very black skin – the latter trait being the only one they share with Cushitic groups.
The Ethiosemites in this era of glorious Kush were a backwater of Iron Age tribal kingdoms in the Ethiopian Highlands and adjacent Gash basin lowlands. The early Ethiosemites were farmers who kept cattle, camels, sheeps, and goats and lived in clan-based tribal kingdoms. This era of early Ethiosemitic state formation is very misty and unknown, but what we do know is that there was some degree of conflict and consolidation as one tribal nobility gained ascendance over their neighbors, and at some point gained critical mass at the expense of the other tribal kings, and formed the precursor to the later famous unified Ethiosemitic empire, the Aksumites.
The Aksumites were never referred to as Ethiopians by the pre-Roman Greeks outside accident, but rather they were known to the Greeks and Romans as the Aksumae. The Somalis and Beja, nomadic Cushitic who inhabited from near the Sinai to central Somali coastline were referred to as the Barbaroi by the Greeks, and were also distinguished from the Greco-Roman era Ethiopians. In fact, the Aksumae (Ethiosemitic Aksumites) and possibly even the northern Barbaroi (pastoral Cushitic Beja) were at war and bitter enemies of the Ethiopians (Eastern Sudanic Kushites). The Beja likely caused havoc to the Kushites as warlike barbarian nomads at their northern and eastern frontiers, and the Aksumae were their imperial rivals – and it was Imperial Aksumite armies under Emperor Ezana who left their Highlands hearth and destroyed and essentially annihilated the Kushites and their kingdom, and razed it into the literal ground. It was then – for the first time – the Aksumites hijacked the Ethiopian name for themselves. But most people continued to refer to them as either Aksumites, or Habasha. It was the self-name of Habesha that Abyssinia was derived, whereas Ethiopia being used to refer to Ethiosemites was a post-Roman phenomenon.
The Beja were great enemies of the Romans, and their tribes allied with rebel Roman leaders. The victory of Emperor Trajan over their tribes led to Beja tribesmen being paraded in Rome to a intrigued Roman population. The Beja were also enemies of the later wave of Eastern Sudanics who arrived in NE Sudan and southern Egypt during the fall of the Roman Empire, the Noba – known to the Romans as the Nobatae – who are the namesake of the Nubians and Nubia, and established the Nubian states of Nobatia, Makuria, and Alodia. Their descendants are the Nubians of today.
Samaritans may have drifted too strongly, and Yemenis may have experienced a Cushitic backflow in the interim, making them more distant from the putative source of admixture and more similar to its putative target?
But anyway the “Minoan” affinities they describe aren’t easy to reconcile with the linguistic evidence…
[OR-179] The genetic landscape of Ethiopia: diversity, intermixing and the association with culture – Garrett Hellenthal
4:45 PM–5:00 PM Jul 24, 2019
G. Hellenthal 1,*, S. Lopez 1, A. Tarekegn 2, N. Bradman 3, M. Thomas 1, E. Bekele 2
“In addition, we describe the ancestral history of different Ethiopian groups, revealing a southwest-northeast cline across Ethiopia defined by ancestry related to Central and West African groups versus Egyptian groups. We date distinct admixture events that show correlations with major language classifications and geography, including events dated to ~1600-2800 years ago involving sources carrying DNA similar to present-day Egyptians versus events starting ~1,450 years ago involving sources carrying DNA related to western Sub-Saharan Africans.”
———
“Egyptian”, even if proxied by Copts, also sounds slightly off (if better than “Minoan”), but I think the point is that no Mediterranean or Near Eastern population is obviously the right metrical fit.
Ethiopia (or something close to it) is referred to as “Kush” in Biblical Hebrew, and the the catchphrase “from India to Ethiopia” is used to describe the vastness of the ancient Persian empire.
It seems that by the Rabbinic era, the adjective “Kushi” (“Kushite”) was used to refer to all dark-skinned Africans and as a descriptor for products that came from Africa. The term is considered pejorative in modern Hebrew (like the English “n-word”) and not in common use — the Greek-derived “Etiopi” is used instead. Jews of Ethiopian origin make up about 2% of the Israeli population.
It all depends, history is rather murky in these periods… They could be refugees fleeing from a Dinarid invasion too, as some local historians propose
@Vytautus
What do you think the early proto-Egyptians and Cushites looked like genetically? Something like Iberomaurusians?
What has become of the Persian Gulf refugium hypothesis?
http://dienekes.blogspot.com/2010/12/persian-gulf-oasis-hypothesis.html
https://www.nature.com/articles/srep25472
(just askin!)
If I am not mistaken, the people the Greeks and Romans called “Ethiopians” (a term meaning “burnt face” referring to dark-skinned peoples) were most often peoples from the north and central Sudan to the south of Egypt (i.e. Nubians and related and nearby groups—likely typically Nilo-Saharans, or a mixture of local Nilo-Saharan and northern Cushitic cultures) including those from the kingdom of Kush and its successor Meroe, rather than those of what is now called Ethiopia (the Kingdom of Axum) of which the Greeks and Romans had less knowledge (and which may they may have called “Abyssinia” at least by Roman times in the rare cases they refer to it).
^Yes, the Abyssinians were irrelevant in that age to the Greeks and Romans. They were only mentioned in passing, and were not called Ethiopians but Aksumae – clearly derived from Aksum.
The Ethiopians of Greek and Roman fame were Nilo-Saharans, likely speaking Eastern Sudanic languages related to Dinka-Nuer and Nubian, and arrived at sometime in the later Old Kingdom/Middle Kingdom to displace and/or assimilate the earlier Afroasiatic populations of “Nubia”.
The likely endonym of this ancient Nilo-Saharan ethnic group was something similar to Kush – the origin of the Hebrew slur, Cush (son of Ham), and Cushitic (language family) originate. Very odd that a people of a whole different language family gave their name to a group that was unrelated to them, especially to groups which were considered enemies of the Kushi.
Which is why it sounds so funny and ironic to say – the Kushites displaced earlier Cushitic peoples.
The Kushi in all likelyhood despised their “Cushitic” neighbors, and would be horrified to learn the progeny of their enemies co-opted not only their Greco-Roman designation and thus leech unrightfully from their ancient fame – which they later destroyed and brought to an end as is the case with modern “Ethiopians”, but their Blemmeye enemies and their relatives progeny would be called by their very own endonym.
FWIW, using some reasonable references, the method below (which also seems to distinguish between Anatolia_N and Levant_N surprisingly well most of the time, per what’s known about the ancestry of various groups and what the papers also estimate with other methods) does also give slight preference to a Levant_N/Iran_N mix as the Eurasian part of the ancestry compared to an Anatolia_N/CHG-Iran_N one.
https://eurogenes.blogspot.com/2019/07/getting-most-out-of-global25_12.html