Reconciling the claimed West-African ancestry in certain Great Lakes clans

[drobbah]

Ask K33/Kale/Memonics what they think. Amhas and other Maghrebis will plot past any non-ANA harbouring individuals of the same Yoruba/Dinka score. Way past actually.

From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.

edit: the Shum Laka paper actually had Ghost ANA as 54% and the original Taforalt paper at 45. C'mon bro.




46% Non-african related Taforalt. Lol.

"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

[gihanga.rwanda]

From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.

"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

Taforalt was modeled as approximately 65% Natufian and 35% modern SSA (Yoruba + Hadza) in the original paper by Loosdrecht et al. 2018, but this was never a great fit for Taforalt, which is 10s of thousands years older than Natufian or any of these extant populations. Lazaridis et al. 2018 proposed an alternative model using more ancient aDNA from Western Eurasia (WE) and was able to model Taforalt as approximately 55% Dzudzuana and 45% Ancient North African (ANA); the Natufians were modeled as 73% Dzudzuana and 27% Taforalt, and thus about 12% ANA, in the same paper. Lipson et al. 2020 was more focused on ancient population substructure in Africa rather than WE but build off the model proposed by Lazaridis et al. 2018. In this paper, Taforalt can be modeled as 54% ANA and 47% WE, compared to the Agaw, who can be modeled as 51% East African agro-pastoralist, 8% Mota (which contains ghost modern ancestry), and 41% WE.

These are all competing models and we won’t be able to draw any strong conclusions until we have more aDNA from North and NE Africa, but the Reich lab appears to be leaning towards a version or combination of the latter two models.

Side note. Natufians having North African ancestry is very likely IMHO and would answer a lot of lingering questions including why Natufian tends to absorb or depress some purported African ancestry in many extant populations.

[ThaYamamoto]

From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.




"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

See what gihanga said. Your source is an earlier paper, the Shum Laka study from Reich espoused a greater proportion of ANA/SSA in Taforalt. And that's what they going with for the time being. You can already see what I linked, figures from the earlier study aren't gonna change the fact they [Reich lab] are leaning toward the 54% margin. I don't see what's controversial about this, K33's calculator had Taforalt at 53% ANA - his work is exceptional and commendable. The great thing with the Shum Laka paper was they could attempt to emulate Basal West+Central African ancestries, further increasing ANA in Taforalt which has been increasing steadily with each study.

[drobbah]

If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration tthe Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off

[Mnemonics]

If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration the Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off

I would be cautious about reading too much into the position of a specific North African. Many North Africans have high amounts of Anatolia_N admixture which can interfere with a direct comparison with Levantine rich populations. Anatolian populations had far less Basal/African affinity than Natufian/Levantine rich populations and will therefore plot further north by default. You would have to compare individuals with similar amounts of Anatolian and Levantine ancestry to judge the impact that actual ANA ancestry has on their position.


Minor note, whatever the Eurasian element in the Dinka sample from the Reichlabs dataset seems like it is not primarily from a Taforalt like population, it has some broadly Eurasian affinities that make it hard to link it to a solely West Eurasian origin.

[drobbah]

I would be cautious about reading too much into the position of a specific North African. Many North Africans have high amounts of Anatolia_N admixture which can interfere with a direct comparison with Levantine rich populations. Anatolian populations had far less Basal/African affinity than Natufian/Levantine rich populations and will therefore plot further north by default. You would have to compare individuals with similar amounts of Anatolian and Levantine ancestry to judge the impact that actual ANA ancestry has on their position.


Minor note, whatever the Eurasian element in the Dinka sample from the Reichlabs dataset seems like it is not primarily from a Taforalt like population, it has some broadly Eurasian affinities that make it hard to link it to a solely West Eurasian origin.

If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

[Mnemonics]

If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

I'm uncertain as to the exact percentage of West Eurasian ancestry in Taforalt, my qpF4ratios support the idea that they are around 55% percent Eurasian but I have had a hard time modeling them in qpadm as more than 40% Dinka.

Nevertheless, the southern position of the Eritreans can be explained as a result of the fact that both the Natufians and Neolithic Iranians are very high in Basal/African ancestry and are likely much more south shifted than the Hunter-Gatherer population that contributed to the Iberomaurusians which seems to be more rich in Anatolia_HG like ancestry.

This is evident in the same f4 ratios that show Taforalt as 55% Eurasian which indicate that the Natufians and Ganj_Dareh samples are in the 82-88 range, while Anatolia_N and Pınarbaşı are more in the 92-97 range.

[gihanga.rwanda]

If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

I think that we can consider several factors to try to answer your question.

1. Taforalt can be modeled as ~50% ANA and Dzudzuana, which the Reich lab views as a appropriate stand in for Western Eurasian-related ancestry. This model implies that Natufian has North African ancestry related to ANA or maybe one of the other “Neo-African” components described in the Shum Laka paper. If that’s the case, we’re currently underestimating the amount of African ancestry in the Middle East and adjacent regions. If Natufian can be modeled as 86% Dzudzuana and 14% ANA, Eritreans may be closer to 50% African.

2. I am also not confident that Dinka is a one to one stand in for the African ancestry in Cushitic and related groups such as Ethio-Semitic speakers. Dinka is already more West African-shifted than other extant and purported ancient Nilo-Saharan speaking groups like the Mursi etc. It also wouldn’t surprise me if Nilotic-related groups acquired more basal African ancestry after their ancestors split with the African ancestors of Cushitic-related groups. I think this is evident if you compare the yDNA/mtDNA profiles of Nilotic and Cushitic-related groups; Dinka and related groups have a lot of lineages such as yDNA B-M60 that are absent or rare in the HoA; Porto-Nilotic speakers might have picked up these lineages from indigenous hunter-gatherers when they arrived in the vicinity South Sudan escaping the drying Sahara.

3. That being said, Cushitic and related groups probably have minor basal African ancestry that is absent in Taforalt. This ancestry could’ve been present in their “Sudanese” ancestors or they might have acquired it by assimilating Mota-related groups. Any basal African ancestry, no matter how small, is going to pull you away from the Eurasian pole, which would further differentiate Taforalt from Cushitic-related groups such as Eritreans.

This is all speculative of course but based on the latest research out of the Reich lab.

[ThaYamamoto]

If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration tthe Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off

Mnemonics and gihanga etc have summed it up p much better than I can but I get what you're saying man, but I originally posted the PCAs specifically in the context of where I plot - roughly ascertaining the ancestry of individuals clustering close to me then became the imperative. Of course nothing is exact until the relevant samples are found and sequenced; however thanks to basal simulations you can roughly work out ancestral streams quite easily. In K33's thread, you can see Morocco North score the respective BasalWest+Central+East+PrePastoral+UstIshim that coincides exactly with the amount of Yoruba/Dinka scored in a standard 'deep' model plus 17-18% ANA which shifts them toward SSA (I never said ANA=SSA, but it definitely plots in this manner) thus Morocco North plots well ahead of the Egyptian average.




As I don't got any IBM/ANA, it was important to me to be able to figure out the max SSA-pull in these individuals/averages. What I do have is high basal/Iran Mesolithic which brings me on to my next point.

Its crucial to differentiate between G25 PCA (which is what it essentially is) and a rudimentary script like nMonte. nMonte is notorious for depressing SSA ancestral streams in high basal individuals i.e. Natufian heavy pops. I'm also very inclined to believe this same phenomenon occurs with high Iran mesolothic/neolithics i.e. Makranis and Iranian Bandaris. Awale mentions this in an older post relating to Egyptians here


This is why PCAs can be more informative in many ways and over-reliance on nMonte can be a mistake imho. I believe I plot where a min.max 18% SSA (thanks AncestryDNA) non-IBM harboring, high Iran Mesolithic individual should when taking into account Maghrebi's SSA+ANA, Egyptians high basal resulting in SSA depression etc, and discounting nMonte maximum likelihood faults.

[Mnemonics]

I think that we can consider several factors to try to answer your question.

1. Taforalt can be modeled as ~50% ANA and Dzudzuana, which the Reich lab views as a appropriate stand in for Western Eurasian-related ancestry. This model implies that Natufian has North African ancestry related to ANA or maybe one of the other “Neo-African” components described in the Shum Laka paper. If that’s the case, we’re currently underestimating the amount of African ancestry in the Middle East and adjacent regions. If Natufian can be modeled as 86% Dzudzuana and 14% ANA, Eritreans may be closer to 50% African.

2. I am also not confident that Dinka is a one to one stand in for the African ancestry in Cushitic and related groups such as Ethio-Semitic speakers. Dinka is already more West African-shifted than other extant and purported ancient Nilo-Saharan speaking groups like the Mursi etc. It also wouldn’t surprise me if Nilotic-related groups acquired more basal African ancestry after their ancestors split with the African ancestors of Cushitic-related groups. I think this is evident if you compare the yDNA/mtDNA profiles of Nilotic and Cushitic-related groups; Dinka and related groups have a lot of lineages such as yDNA B-M60 that are absent or rare in the HoA; Porto-Nilotic speakers might have picked up these lineages from indigenous hunter-gatherers when they arrived in the vicinity South Sudan escaping the drying Sahara.

3. That being said, Cushitic and related groups probably have minor basal African ancestry that is absent in Taforalt. This ancestry could’ve been present in their “Sudanese” ancestors or they might have acquired it by assimilating Mota-related groups. Any basal African ancestry, no matter how small, is going to pull you away from the Eurasian pole, which would further differentiate Taforalt from Cushitic-related groups such as Eritreans.

This is all speculative of course but based on the latest research out of the Reich lab.

You maybe on to something with Dinka having something more basal, there is some affinities shared by the Dinka and and Luo samples with ZAF_2000BP that seem to be less present in some West Africans.

However, I think this same basal African ancestry was also present in the African ancestors of Cushitic/Ethiosemite speakers as the same trends show up in the f4 ratios. When I try putting deeper South African groups into the F4 ratio the gap between West and East Africans in their shared ancestry with Eurasians shrinks relative to the direct comparisons, which could imply that AEA rich populations could have picked up something South African like, this also could be a result of increased affinity due to pastoralist and hunter-gatherer admixture from East Africa into South Africa, but I am uncertain about how I could adequately test the direction of the gene flow without a more pristinely AEA ancient sample.