Anglo-Saxon aDNA; Gretzinger, Francis Crick and beyond

[Capitalis]

I ran this analysis for my own interest but will share here.




* The good-coverage Medieval Eastry, Kent sample has clear North and Sub-Saharan African ancestry. I couldn't identify another ancient West Eurasian sample with such high Sub-Saharan African ancestry.

* I couldn't find any evidence of African ancestry in the three low- to middling-coverage Medieval Groningen samples labelled by the Reich lab as "Africa 1/2 outliers" .

* The low-coverage Medieval Kilteasheen, Co. Roscommon sample, labelled by the Reich lab as a "Steppe outlier", has evidence of Sub-Saharan African ancestry, but shouldn't really be in the G25 dataset.

:-)

[JonikW]

I ran this analysis for my own interest but will share here.




* The good-coverage Medieval Eastry, Kent sample has clear North and Sub-Saharan African ancestry. I couldn't identify another ancient West Eurasian sample with such high Sub-Saharan African ancestry.

* I couldn't find any evidence of African ancestry in the three low- to middling-coverage Medieval Groningen samples labelled by the Reich lab as "Africa 1/2 outliers" .

* The low-coverage Medieval Kilteasheen, Co. Roscommon sample, labelled by the Reich lab as a "Steppe outlier", has evidence of Sub-Saharan African ancestry, but shouldn't really be in the G25 dataset.

:-)

I'm looking forward to the separate paper we were promised on the Eastry "Updown Girl." I see she has her own Wikipedia page now. This must be one of the most interesting discoveries yet from Anglo-Saxon England.

[Capsian20]

I ran this analysis for my own interest but will share here.




* The good-coverage Medieval Eastry, Kent sample has clear North and Sub-Saharan African ancestry. I couldn't identify another ancient West Eurasian sample with such high Sub-Saharan African ancestry.

* I couldn't find any evidence of African ancestry in the three low- to middling-coverage Medieval Groningen samples labelled by the Reich lab as "Africa 1/2 outliers" .

* The low-coverage Medieval Kilteasheen, Co. Roscommon sample, labelled by the Reich lab as a "Steppe outlier", has evidence of Sub-Saharan African ancestry, but shouldn't really be in the G25 dataset.

:-)

I'm looking forward to the separate paper we were promised on the Eastry "Updown Girl." I see she has her own Wikipedia page now. This must be one of the most interesting discoveries yet from Anglo-Saxon England.

This interesting i was seen there SSA Haplogroup Y-DNA and mtDNA from UK  ( Y-DNA E-M2 and mtDNA L (Negative M and N))
https://www.yfull.com/tree/E-Y310685/

[Megalophias]

That Eastry Updown girl is interesting. I couldn't find much about her, except a news article saying she was 33% West African (related to Yoruba or Esan) and it was on her father's side, maybe her grandfather or great-grandfather. I'm pretty sure she didn't inherit a third of her DNA from one great-grandfather, but anyway.

68% English, 32% Yoruba - distance 5.55%

Crappy fit. Much better with North African:
49% Norwegian, 26% Tiznit Berber, 24% Yoruba - 2.78%

No need for it to be Nigerian, other West African is fine. Probably they said Esan or Yoruba because those are common reference populations:
50% Norwegian, 24% Tiznit Berber, 25% Bambara Mali - 2.61%

But actually better with a Central African source, may have some East African ancestry in there:
50% Norwegian, 24% Tiznit Berber, 26% Laka - 2.13%

I doubt the proportions here are a coincidence, I'd say her mother was a North European and her father was a North African with half Sub-Saharan ancestry. Looking at just her X chromosome might tell something.

[Capitalis]

That Eastry Updown girl is interesting. I couldn't find much about her, except a news article saying she was 33% West African (related to Yoruba or Esan) and it was on her father's side, maybe her grandfather or great-grandfather. I'm pretty sure she didn't inherit a third of her DNA from one great-grandfather, but anyway.

68% English, 32% Yoruba - distance 5.55%

Crappy fit. Much better with North African:
49% Norwegian, 26% Tiznit Berber, 24% Yoruba - 2.78%

No need for it to be Nigerian, other West African is fine. Probably they said Esan or Yoruba because those are common reference populations:
50% Norwegian, 24% Tiznit Berber, 25% Bambara Mali - 2.61%

But actually better with a Central African source, may have some East African ancestry in there:
50% Norwegian, 24% Tiznit Berber, 26% Laka - 2.13%

I doubt the proportions here are a coincidence, I'd say her mother was a North European and her father was a North African with half Sub-Saharan ancestry. Looking at just her X chromosome might tell something.

This is outside of my area of "expertise", so I won't post any more models here; I just chose the most outlying ancient African samples to exclude any West Eurasian signal in modern Africans impacting the confirmation of Sub-Saharan ancestry.

However, her African ancestry seems complex, as when adding sample I5950 aka "Mota" (2576-2465 calBCE, Gamo Highlands, Ethiopia) to the model, her Iberomaurusian ancestry remains similar but her Central African ancestry halves with the remainder assigned to Mota. Someone with more knowledge of African DNA could probably narrow this down.

She is dated broadly to 400-1100 CE in the Reich lab AADR file, but I do find it odd that she is missing from the Gretzinger et al. supplementary Excel file and is not highlighted in the supplementary Word file. Did they have doubts regarding her authenticity at some stage?

---

At the risk of upsetting everyone here again, I wasn't impressed with the Gretzinger et al. paper and think there were some real quality control issues in the way they ran their analysis. For example, on pp. 80-81 of the supplementary Word file, we find the following,

We therefore used unsupervised ADMIXTURE to identify potential Asian, Middle Eastern, or African ancestry within our ancient samples. At K = 7, we identify two components mainly located in sub-Saharan Africa (a southern and western African), one north African/Middle Eastern component, one northern European component, one component maximised in populations from the Caucasus and South Asia, as well as two components located in eastern Asia, one maximised in northern Eurasians and one in East Asians (Supp. Fig. 5.2a).

[...]

The only exception is the outlier individual I11570 from Worth Matravers. This individual exhibits 22.4% ancestry from the western African component maximised in present-day Yoruba, Mende, and Esan. Similar amounts are measured in present-day north Africans like Algerians and Mozabites, however, those populations also carry high amounts of the North Africa/Middle Eastern component, which is minimal in I11570. It is therefore more likely that I11570 is the product of recent admixture between a northern European and a west African source. While small amounts of sub-Saharan African ancestry might be the result of low-coverage and/or contamination, there is no reason to assume this for I11570.

According to the Reich lab AADR file, sample I11570 has coverage of 0.030x, which is atrocious. This directly contradicts the last sentence of the quote above. At Anthrogenica, I pointed out that the PCA plot on p.80 showing multiple outliers, including modern Russian-like outliers, is the result of Gretzinger et al. plotting extremely low coverage samples which are known to show spurious ancestry (see the Kilteasheen "outlier" above).

As I attempted to show in the thread here regarding the Early Medieval samples from the Tyrol, you cannot trust these supposedly professional research papers to analyse the samples coherently, you must check their work for yourself before accepting their conclusions.

Rant over. :-)

[Capitalis]

However, her African ancestry seems complex, as when adding sample I5950 aka "Mota" (2576-2465 calBCE, Gamo Highlands, Ethiopia) to the model, her Iberomaurusian ancestry remains similar but her Central African ancestry halves with the remainder assigned to Mota. Someone with more knowledge of African DNA could probably narrow this down.

She is dated broadly to 400-1100 CE in the Reich lab AADR file [...]

Assuming she is "real" and has complex African ancestry, could she be (the descendant of) an early victim of the Medieval African slave trade? (see the map linking North, West and East Africa by slave trade routes).

[JonikW]

That Eastry Updown girl is interesting. I couldn't find much about her, except a news article saying she was 33% West African (related to Yoruba or Esan) and it was on her father's side, maybe her grandfather or great-grandfather. I'm pretty sure she didn't inherit a third of her DNA from one great-grandfather, but anyway.

68% English, 32% Yoruba - distance 5.55%

Crappy fit. Much better with North African:
49% Norwegian, 26% Tiznit Berber, 24% Yoruba - 2.78%

No need for it to be Nigerian, other West African is fine. Probably they said Esan or Yoruba because those are common reference populations:
50% Norwegian, 24% Tiznit Berber, 25% Bambara Mali - 2.61%

But actually better with a Central African source, may have some East African ancestry in there:
50% Norwegian, 24% Tiznit Berber, 26% Laka - 2.13%

I doubt the proportions here are a coincidence, I'd say her mother was a North European and her father was a North African with half Sub-Saharan ancestry. Looking at just her X chromosome might tell something.

This is outside of my area of "expertise", so I won't post any more models here; I just chose the most outlying ancient African samples to exclude any West Eurasian signal in modern Africans impacting the confirmation of Sub-Saharan ancestry.

However, her African ancestry seems complex, as when adding sample I5950 aka "Mota" (2576-2465 calBCE, Gamo Highlands, Ethiopia) to the model, her Iberomaurusian ancestry remains similar but her Central African ancestry halves with the remainder assigned to Mota. Someone with more knowledge of African DNA could probably narrow this down.

She is dated broadly to 400-1100 CE in the Reich lab AADR file, but I do find it odd that she is missing from the Gretzinger et al. supplementary Excel file and is not highlighted in the supplementary Word file. Did they have doubts regarding her authenticity at some stage?

---

At the risk of upsetting everyone here again, I wasn't impressed with the Gretzinger et al. paper and think there were some real quality control issues in the way they ran their analysis. For example, on pp. 80-81 of the supplementary Word file, we find the following,

We therefore used unsupervised ADMIXTURE to identify potential Asian, Middle Eastern, or African ancestry within our ancient samples. At K = 7, we identify two components mainly located in sub-Saharan Africa (a southern and western African), one north African/Middle Eastern component, one northern European component, one component maximised in populations from the Caucasus and South Asia, as well as two components located in eastern Asia, one maximised in northern Eurasians and one in East Asians (Supp. Fig. 5.2a).

[...]

The only exception is the outlier individual I11570 from Worth Matravers. This individual exhibits 22.4% ancestry from the western African component maximised in present-day Yoruba, Mende, and Esan. Similar amounts are measured in present-day north Africans like Algerians and Mozabites, however, those populations also carry high amounts of the North Africa/Middle Eastern component, which is minimal in I11570. It is therefore more likely that I11570 is the product of recent admixture between a northern European and a west African source. While small amounts of sub-Saharan African ancestry might be the result of low-coverage and/or contamination, there is no reason to assume this for I11570.

According to the Reich lab AADR file, sample I11570 has coverage of 0.030x, which is atrocious. This directly contradicts the last sentence of the quote above. At Anthrogenica, I pointed out that the PCA plot on p.80 showing multiple outliers, including modern Russian-like outliers, is the result of Gretzinger et al. plotting extremely low coverage samples which are known to show spurious ancestry (see the Kilteasheen "outlier" above).

As I attempted to show in the thread here regarding the Early Medieval samples from the Tyrol, you cannot trust these supposedly professional research papers to analyse the samples coherently, you must check their work for yourself before accepting their conclusions.

Rant over. :-)

There were no doubts about Updown Girl at the time and she only wasn't published in full because they said she's of sufficient interest to deserve a paper all of her own. I imagine we can expect this soonish. They said at the time she was closest to Esan and Yoruba populations on her African side.

ADD: her grave goods, which included a finely decorated pot, were typical of her early seventh century site. They assumed her African ancestry was on her father's side because of her European mtDNA.

[Capitalis]

There were no doubts about Updown Girl at the time and she only wasn't published in full because they said she's of sufficient interest to deserve a paper all of her own. I imagine we can expect this soonish. They said at the time she was closest to Esan and Yoruba populations on her African side.

ADD: her grave goods, which included a finely decorated pot, were typical of her early seventh century site. They assumed her African ancestry was on her father's side because of her European mtDNA.

So they hid her from view. :-)

[JonikW]

There were no doubts about Updown Girl at the time and she only wasn't published in full because they said she's of sufficient interest to deserve a paper all of her own. I imagine we can expect this soonish. They said at the time she was closest to Esan and Yoruba populations on her African side.

ADD: her grave goods, which included a finely decorated pot, were typical of her early seventh century site. They assumed her African ancestry was on her father's side because of her European mtDNA.

So they hid her from view. :-)

They're keeping us waiting for sure :-) Here's her grave.

[Capitalis]

They said at the time she was closest to Esan and Yoruba populations on her African side.

As Megalophias showed above, this is unlikely to be correct. This is how those modern West African populations look compared to Updown Girl in my basic ancient model.

[Capitalis]

To highlight how significantly outlying Updown Girl is, here are all of the G25 England samples from the Iron Age to the Medieval / Viking Age with signs of ancestry outside of the four typical components of modern English people. I consider coverage of at least 0.1x as reasonable, so these are all okay.

[Megalophias]

EAS003 has decent coverage and is a 3rd degree relative of EAS001 and EAS002 from the same cemetery according to the Reich .anno file. Table S1 in the original paper gives a radiocarbon date of 605-654 cal AD for EAS001, but I don't know whether this was directly on the remains or on something associated with the grave; the supplement doesn't mention it, and the .anno file only gives a broad context date. A chapter on Eastry from 2011 says this cemetery is dated to the 7th century. I don't know if there is any isotope data or anything to say whether she came from the continent.

In any case she looks legit, but her father must have an interesting story. A foreign merchant, diplomat, or soldier maybe? If the East African element is real, his Sub-Saharan ancestor(s) (his father?) may have come across the Central Sahara, via the Garamantian route maybe? Could be an escaped or freed slave, or a free traveller, who knows. Too bad we have so few samples to compare.

[Rodoorn]

Hypothesis:

Anglo-Saxon aDNA=

McColl et al 2024:

"The Danish Isles ancestry that was widespread on Zealand from 2200 BP disappears from ~1600 BP."

C.q. appears as Anglo-Saxons around the North Sea from ~1600 BP.

"Truth or dare"