Saturday, February 1, 2025

Ancient Egyptian Old Kingdom DNA (2500-3000 BCE): 10% Sub-Saharan and no Iran-Chalcolithic

I have larger, more expansive posts in the works but I had to share this as yet another piece of quick, interesting findings that seemed to have not popped up on a lot of people's radars. Credit should go to a Somalispot poster by the username of Xareen for bringing it to my attention on there. We now have a Bronze-Age Egyptian sample thanks to the thesis of one Adeline Morez from Liverpool John Moores University.[1]

In her thesis she goes over a sample dated to ~2,868-2,492 BCE from the Nuerat cemetery just south of Beni Hasan in the Minya governorate of Upper Egypt and they found that this sample is about ~10% Dinka-like in its ancestry


NUE001 also carries ~10% ancestry similar to the one found in the 4,500-year-old Ethiopian genome, derived from the eastern sub-Saharan African component
That component has gone by many names over the years in the anthropology sphere. East-African (EA), Ancestral East African (AEA), "Nilo-Saharan", or simply "Dinka-like" as it makes up the vast majority of the ancestry of Nilo-Saharan groups such as Dinkas and the Gumuz above, as well as the Mota Hunter-Gatherer of Southern Ethiopia. It further makes up the vast majority of the Sub-Saharan African ancestry in modern Horn-Africans such as Somalis, Amharas, Oromos, Tigrinyas, Aris and Wolaytas who are mostly genetically intermediate populations between Middle-Easterners and groups such as the Dinka (see here).

It appears this Egyptian, unlike the Iron-Age Egyptians from 8 years ago [2], has a bit of this ancestry much like modern Copts do:


They also seem to lack or have very little Iran-Chalcolithic-related ancestry, unlike the later Egyptians of the Iron Age, with the author in my opinion correctly speculating that this is due to the later period Egyptians having significant admixture from Asiatic (Semitic-speaking) groups such as the Hyksos:

The Nuerat sample did not carry the Caucasus Hunter-Gatherer genetic component that started to spread across West Asia ~4,000 years ago and is widely spread in present-day populations. The presence of this component in Egypt is likely associated with admixture between local Egyptian populations and Bronze Age-related populations from West Asia. This admixture pattern might result from the dominance of Lower Egypt by Canaanite (Levantine) rulers during the Second Intermediate Period (ca. 1,650-1,550 BCE).
A group of West Asiatic foreigners, possibly Canaanites, labelled as Aamu (ꜥꜣmw), including the leading man with a Nubian ibex labelled as Abisha the Hyksos (𓋾𓈎𓈉 ḥqꜣ-ḫꜣsw, Heqa-kasut for "Hyksos"). Tomb of 12th-dynasty official Khnumhotep II, at Beni Hasan (c. 1890 BC).


So, it seems many of us who assumed Ancient and Pre-historic Egypt may have not had any Dinka-like ancestry were being premature in going off those Iron-Age genomes we got several years ago, if indeed this sample is representative and I see no reason to believe it is not. It appears that when the West-Asian migrants came in they brought with them not just significant Iran-Neolithic/Caucasus-Hunter-Gatherer type ancestry but they also diluted out the previously present Sub-Saharan ancestry.

This sample, if we can get it into David Wesolowski's Eurogenes Global 25 PCA should be extremely interesting to run for many populations including my own Horn-Africans as I strongly suspect it will prove quite representative of our non-Arabian MENA ancestors who consistently appear very similar to Natufians—with a slight Iberomaurusian shift—in G25 based runs:
I reckon this individual will essentially be that with ~10% Dinka-like ancestry. We'll see, and here's to seeing more ancient genomes from Egypt from around this period and earlier. 


References

1. Morez A. Reconstructing past human genetic variation with ancient DNA: case studies from ancient Egypt and medieval Europe [doctoral thesis]. Liverpool: Liverpool John Moores University; 2023. Available from: 

2. Schuenemann VJ, Peltzer A, Welte B, van Pelt WP, Molak M, Wang CC, et al. Ancient Egyptian mummy genomes suggest an increase of Sub-Saharan African ancestry in post-Roman periods. Nat Commun. 2017 May 30;8:15694. doi: 10.1038/ncomms15694. Available from: https://www.nature.com/articles/ncomms15694

3. Wikimedia Commons. Drawing of the procession of the Aamu group tomb of Khnumhotep II at Beni Hassan [Internet]. Available from: https://commons.wikimedia.org/wiki/File:Drawing_of_the_procession_of_the_Aamu_group_tomb_of_Khnumhotep_II_at_Beni_Hassan.jpg.


Simplified for total laymen:

About eight years ago, Iron-Age Egyptian genomes were analyzed, showing no detectable "Black" African ancestry, unlike modern Egyptians—both Muslim and Christian—who have such ancestry at usually a range of 10-20%. This led to the assumption that such ancestry was a later addition and largely absent in Ancient Egypt.

Media headlines sensationalized the findings, with some even falsely and absurdly claiming that Ancient Egyptians were "closer to Europeans" than to modern Egyptians, their direct descendants. This is misleading because mostly non-Eurasian admixed Africans are genetically quite distinct from Eurasians. Even if two populations share 90% of their ancestry, a small (~10%) amount of Sub-Saharan ("Black") ancestry can inflate genetic distance, making them appear more distant from a closely related group than other Eurasians are, even when those Eurasians (such as Europeans) share no ancestry from within the last 5,000 to 10,000 years or more with the group that lacks Sub-Saharan ancestry unlike the group that's 10% Sub-Saharan. To be fair, the researchers didn't necessarily sensationalize the findings like that—the media did.

However, this new data discussed above regarding an Old Kingdom sample suggests that migrations from the Levant (Palestine-Israel region)—possibly during the Middle Kingdom—introduced new Eurasian ancestry, temporarily diluting the earlier Sub-Saharan African component. 

So, Egypt seems to for now have had some minor but not negligible Black-African ancestry (~10%), lost it for a time due to migrations from Asia, then got it back through a mixture of intermixing with Nubians (likely what happened with Copts) and the Trans-Saharan Slave Trade (mostly what seems to have happened with the Muslims), the latter case of which brought some completely new Black-African elements such as West-African ancestry. The rest is nerdy stuff about the other strains of ancestry from the Middle-East the average person probably won't be that interested in but I'll be discussing in future posts so feel free to stick around if you're intrigued.

30 comments:

  1. Those Neolithic Sudanese samples from Ghaba could've been informative. We also need some actual "Ancient East African" samples or something closer to it than Mota. Modelling with Ancient Egyptians with Mota like in this paper is so clearly flawed.

    ReplyDelete
  2. How do we know how much AEA ancestry dinkas have? We don’t have any fully AEA samples atm so if the dinka proxy has any eurasian admixture we wouldn’t know since we lack any AEA samples to evaluate it.

    ReplyDelete
    Replies
    1. To be fair, in her thesis the component peaks in the Gumuz and Mota. We've had some runs on how much Eurasian ancestry the Gumuz have and it's quite low at about 1-2% from one of the best attempts I've ever seen at quantifying the exact amount of Eurasian affinities in certain African populations that's held true with later similarly good formal stat runs:

      https://arxiv.org/abs/1307.8014

      And Mota doesn't have any Eurasian admixture. Just odd Eurasian affinities probably due to being East-African as ancient East-Africans might very well have an uptick in Eurasian affinities due to Eurasians seeming like a very old EA off-shoot going off of mtDNA. So that's pretty much an overwhelmingly SSA component, I would say.

      Delete
    2. So in terms of autosomal mixture if you take a Old Kingdom Egyptian but add some more Dinka like ancestry with a pinch of Omotic and a little Yemeni admixture you get a Cushite. Would that be somewhat correct to say?

      Delete
    3. It's looking that way, yes. But we'll see if and when this samples and others from its period and earlier are available. But that is what I suspect and will go into in some future posts. Horner Cushites and Ethiosemites basically being:

      Predynastic to Early Dynastic Egyptian + Dinka-like + Mota + Iron-Age Yemeni

      Delete
  3. What sample was used to evaluate the AEA ancestry in dinkas? That’s what im still wondering.

    ReplyDelete
    Replies
    1. There's no sample. It's basically just a hypothesized ancestral population (a statistical construct) that pops up in ADMIXTURE and some PCA based outputs. It seems to capture the ancestry in groups like Dinkas that isn't West-African (WA) or Eurasian (EUR).

      Until we have an actual population made up of this component it will always be somewhat imperfect as some few WA and EUR alleles will wind up in it. But it probably was a real population because, whatever it is, the vast majority of the ancestry in the Gumuz and Mota (who lack WA admixture, especially the latter) and the Dinka is made up of it so a real, "pure" group of this kind likely existed at some point.

      Delete
    2. Interesting. Do you think there’ll be a fully AEA sample available in the near future? Also if you had to break down the non-AEA ancestry of Somalis into natufian, anatolian, chalcolithic iranian and iberomaurusian what would be the percentage of each?

      Delete
    3. Can't say. We'll have to see whenever we can get samples from regions like Sudan, South Sudan and Chad. That would be where we'd mostly find them. As for those percentages I'll go into them more in future posts. Stay tuned!

      Delete
    4. Alright thanks

      Delete
    5. Was the Kingdom of D'mt around the 8th c B.C. what brought Iron Age Yemeni admixture to Eritrea and Ethiopia? Do Somalis have any of this admixture as well cause from what I hear we dont. If this is the case then I would assume that Somalis were to the east of Eritrea/Ethiopia around 8th c B.C. in northern Somaliweyn due to us not having much of that admixture and everyone in that region including fellow LEC Cushites like Afars and Saho having the admixture. Do you agree?

      Delete
    6. Yes, I wouldn't say it was specifically just D'mt but yes. Around 2,800 to 3,000 years ago the Proto-Ethiosemitic (PES) community landed in Northern Eritrea and spread into the Northern Highlands and seems to have brought the Ethiosemitic languages, Arabian ancestry, possibly Iron metallurgy and some new Y-DNA and mtDNA lineages, most prominently J-P58 subclades.

      As for Somalis, I do not agree. We do have the later Arabian admixture as you can see in the G25 based model above. And in case you argue that not all of the Somalis samples show it; they do. PCAs are similar to ADMIXTURE and can be skewed by drift quite a bit—especially when you have more samples and are not running on the averages for the populations—and Somalis tend to be quite something in that regard, often being the focal point for an "Ethio-Somali" component in ADMIXTURE runs, if you use the right samples and keep running Somali samples against one another you'll notice they all do in fact have it, it's just that some start appearing more "pure" in a weird manner approaching something "Ethio-Somali"-like in ADMIXTURE. Syncs well with how Somali SSA to MENA scores are pretty much homogenous, especially among the samples used above (variance of just 2%) and how our mtDNA and Y-DNA markers don't show any signs of there being some varied sourcing of Arabian admixture; I would say it's homogenous across the ethnic group unless someone has recent known outside admixture.

      As for where it comes from, it looks Intra-Horn mediated. As in if you run a model like "Amhara->Somali" or "Oromo->Somali" it completely throws out any Jawf-Yemeni admixture so we seemingly did intermix with other Horners at some point. How and when remains to be seen. Guess we'll know someday via ancient DNA.

      Beyond that Somalis even appear to have our own unique influences from ancient Yemen in that our dominant Y-DNA T subclade T-BY182320 has a basal carrier from the Asir region of Saudiya which is historically a mere extension of Yemen and used to be a part of it:

      https://www.yfull.com/tree/T-BY182320/

      Might have come to the region with the domestication of camels; asiatic admixture in our caprines and bovines; and OSA loans into our language among other things.

      https://www.academia.edu/5529034/2013_Strata_in_Semitic_loanwords_in_Northern_Somali

      Delete
  4. Somalis are best modelled using Jordan_EBA samples rather than Yemeni on qpAdm afaia. Perhaps E-Z813 is directly from Bronze Age Sinai?

    That being said I just checked E-Z813 and a new basal clade was found in a Sudani. Maybe not.

    ReplyDelete
    Replies
    1. qpAdm is not inherently superior to PCA coordinates and Monte Carlo simulations, as no statistical method is "magic" or universally better. It all depends on the dataset, references, and context. In this case, I'm inclined to lean with the nMonte and G25 results as they historically make much more sense.

      Nevertheless, "Jordan_EBA" would basically be a representation of early Semites from the Southern Levant which is mostly what the PES people would've been anyway. Doesn't really change anything...

      Delete
  5. This is fascinating stuff! Thank you so much for your post.

    ReplyDelete
  6. Do you think we might also have some ANA ancestry that’s hidden within the AEA results?

    ReplyDelete
  7. I personally doubt that the south Arabian ancestry in Somalis was introduced via mixing with another Horner group. Somalis according to the G25 model is 13.4% arabian while the arabian ancestry in most of the other Horners included is close to doubt the amount found in Somalis. So that would mean that nearly half of the ancestry in Somalis would have to come from a foreign Horner group with a quarter arabian ancestry which I find very hard to believe. Most or all of it coming directly from arabia seems to make the most sense. G25 seems to just be cloaking arabian ancestry behind horner groups with more recent and higher arabian ancestry when used as a reference population.

    ReplyDelete
    Replies
    1. While I don't think what you're proposing is impossible, I find it implausible simply because these affinities in Somalis appear identical to what is in other Horners. It's not simply that a "Amhara to Somali" model would immediately usurp a "Yemeni to Somali" model, but also that we match the same specific Yemeni group that Amharas and other Horners do.

      Specifically Yemenis from the al-Jawf Governorate which, funny enough, is in the rough area of Yemen you'd expect the PES people to have been from (west and north). Now if our ancient Yemeni looked Mehri or Hadhrami while all other Horners leaned Jawf I'd be inclined to side with you but that's not the case. It looks like the same stuff.

      Delete
    2. Could it be possible that the same jawf Yemenis ancestral to the PES migrated eastward into Somali territory and admixed with us? I see that as being more likely

      Delete
    3. What you are suggesting to me seems perhaps possible in that Somali Y-DNA T seems to be unique from what is in other Horners and was likely spread by OSA (maybe even PES?) speaking pastoralists from Yemen who may have been the people who brought camels; asiatic admixture in our goats, sheep and cattle; and perhaps even iron metallurgy to Somalis around 2,000-2,500 years ago but I also think the Horn intermixture, even at extremes of 50% or more, has credence and overtime you might see my point with future posts where I'll go into the history of Cushites and Ethiosemites as we so far know it.

      Delete
  8. Hey awale abdi do you think the FST distance between the medieval Sudanese christians and the old kingdom samples are they closer to say modern iranians then to other North East Africans (like the medieval Sudanese christian samples), I say this because someone of iranian descent told me that The OK samples were closer to all west eurasians then they were to the medieval Sudanese christians samples? Who seemed to be more on the iberomaurusian cline

    ReplyDelete
  9. This iranian guy went even as far to claim The OK samples had no "SSA" ancestry & that ANA wasn't even a prominent ancestry in East africans like dinka etc..(he said this after i pointed out their was ANA in Natufian) and BTW I'm not even an Afrocentric lol

    ReplyDelete
  10. Also didn't the 2017 paper find dinka like ancestry that was slightly less then modern Copts ?

    ReplyDelete
  11. Interestingly, although ADMIXTURE detects that NUE001 carried a sub saharan component (Fig 4.3), qpAdm contradicts the presence of this ancestry, what to make of this awale ?does NUE001 really carry sub saharan mota like ancestry ?

    ReplyDelete
  12. Also i saw voudu run or whatever its called and an old kingdom egyptian sample i think NUE001)

    ReplyDelete
    Replies
    1. Scored 6.4 % iran_Neolithic _farmer_garj is this accurate if they apparently lach CHG or Iranian chalcolithic

      Delete