PDA

View Full Version : Reconstructing Prehistoric African Population Structure



drobbah
09-21-2017, 06:51 PM
3100 yr old Tanzanian pastoralist genome has been analyzed (http://www.cell.com/cell/fulltext/S0092-8674(17)31008-5).I was hoping to see some sort of E-M35/T-M70 lineage from the Tanzanian nomad but instead the Kenyan genome came out as E1b1b1b2

"We found that the ∼3,100 BP individual (Tanzania_Luxmanda_3100BP), associated with a Savanna Pastoral Neolithic archeological tradition, could be modeled as having 38% ± 1% of her ancestry related to the nearly 10,000-year-old pre-pottery farmers of the Levant (Lazaridis et al., 2016), and we can exclude source populations related to early farmer populations in Iran and Anatolia. These results could be explained by migration into Africa from descendants of pre-pottery Levantine farmers or alternatively by a scenario in which both pre-pottery Levantine farmers and Tanzania_Luxmanda_3100BP descend from a common ancestral population that lived thousands of years earlier in Africa or the Near East. We fit the remaining approximately two-thirds of Tanzania_Luxmanda_3100BP as most closely related to the Ethiopia_4500BP (p = 0.029) or, allowing for three-way mixture, also from a source closely related to the Dinka (p = 0.18; the Levantine-related ancestry in this case was 39% ± 1%)"





The study also claimed the Somalis have around 13-16% Iran_N ancestry.Doesn't seem realistic in my honest opinion...

"While these findings show that a Levant-Neolithic-related population made a critical contribution to the ancestry of present-day eastern Africans (Lazaridis et al., 2016), present-day Cushitic speakers such as the Somali cannot be fit simply as having Tanzania_Luxmanda_3100BP ancestry. The best fitting model for the Somali includes Tanzania_Luxmanda_3100BP ancestry, Dinka-related ancestry, and 16% ± 3% Iranian-Neolithic-related ancestry (p = 0.015). This suggests that ancestry related to the Iranian Neolithic appeared in eastern Africa after earlier gene flow related to Levant Neolithic populations, a scenario that is made more plausible by the genetic evidence of admixture of Iranian-Neolithic-related ancestry throughout the Levant by the time of the Bronze Age (Lazaridis et al., 2016) and in ancient Egypt by the Iron Age (Schuenemann et al., 2017)."


http://www.cell.com/cms/attachment/2108833685/2082649523/gr1.jpg

Awale
09-21-2017, 08:24 PM
This is pretty funny for me because I emailed Pontus Skoglund about something a while back and he just recently replied and one of his two emails said:


tomorrow we have a paper on African ancient DNA that is coming out in Cell which might be of interest to you.

Great to see this study, anyway. Will have to properly read it and look at its data before posting again, though.


The study also claimed the Somalis have around 13-16% Iran_N ancestry.Doesn't seem realistic in my honest opinion...

Agreed.

blackflash16
09-22-2017, 12:15 AM
Here are the qpAdm results cited in the paper:




Table S5. Details of ancestry proportions inferred using qpAdm.

















Target population
P-value for chosen model
Number of sources
Source 1
Source 1 ancestry
Source 1 SE
Source 2
Source 2 ancestry
Source 2 SE
Source 3
Source 3 ancestry
Source 3 SE
Outgroup population in worst outlier for f4(Target population, fitted Target Population; Base (Mbuti), Outgroup)
Z-score for worst outlier
Number of SNPs (Ancient individuals are enriched for 1.2M SNPs, present-day populations genotyped for ~600k)



Kenya_400BP
0.263546414
1
Ethiopia_4500BP
100.0%
0.0%








27,532



Tanzania_Pemba_1400BP
0.070819424
1
Ethiopia_4500BP
100.0%
0.0%








22,680



Hadza1
0.044373607
1
Ethiopia_4500BP
100.0%
0.0%








74,967



Tanzania_Pemba_600BP
0.356951982
1
Mende
100.0%
0.0%








102,412



Malawi_Chewa
0.818452487
1
Mende
100.0%
0.0%








69,106



Malawi_Ngoni
0.683597822
1
Mende
100.0%
0.0%








77,158



Malawi_Tumbuka
0.596606063
1
Mende
100.0%
0.0%








70,736



Malawi_Yao
0.56514891
1
Mende
100.0%
0.0%








69,887



Yoruba
0.699650257
1
Mende
100.0%
0.0%








65,208



Esan
0.889835332
1
Mende
100.0%
0.0%








73,045



Gambian
0.010650124
1
Mende
100.0%
0.0%








74,979



Luo
0.020767054
1
Mende
100.0%
0.0%








72,548



BantuKenya
0.062937628
1
Mende
100.0%
0.0%








74,528



BantuSA_Ovambo
0.459703325
1
Mende
100.0%
0.0%








82,121



Himba
0.025321143
1
Mende
100.0%
0.0%








77,920



Wambo
0.311331451
1
Mende
100.0%
0.0%








73,539



BantuSA_Herero
0.038491594
1
Mende
100.0%
0.0%








83,930



Tanzania_Luxmanda_3100BP
0.027331465
2
Ethiopia_4500BP
62.2%
1.4%
PPNB
37.8%
1.4%



South_Africa_2000BP
2.5
124,595



Tanzania_Luxmanda_3100BP_PMD
0.164555237
2
Ethiopia_4500BP
62.8%
1.8%
PPNB
37.2%
1.8%



Anatolia_Neolithic
-57.7
60,127



South_Africa_1200BP
0.797655643
2
South_Africa_2000BP
59.7%
2.3%
Tanzania_Luxmanda_3100BP
40.3%
2.3%



Ust_Ishim
1.1
87,068



South_Africa_1200BP_PMD
0.79752436
2
South_Africa_2000BP
46.0%
7.4%
Tanzania_Luxmanda_3100BP
54.0%
7.4%



Punjabi
1.3
14,391



Tanzania_Zanzibar_1400BP
0.1349785
2
Ethiopia_4500BP
69.2%
3.0%
South_Africa_2000BP
30.8%
3.0%



PPNB
-1.7
108,179



Malawi_Fingira_6100BP
0.120308552
2
Ethiopia_4500BP
39.5%
5.5%
South_Africa_2000BP
60.5%
5.5%



Iran_Neolithic
-1.4
25,077



Malawi_Hora_8100BP
0.048618646
2
Ethiopia_4500BP
46.6%
2.8%
South_Africa_2000BP
53.4%
2.8%



Dinka
2.7
90,781



Malawi_Fingira_2500BP
0.041801755
2
Dinka
35.0%
2.8%
South_Africa_2000BP
65.0%
2.8%



Mende
2.2
94,371



Mandenka
0.698583702
2
Mende
97.2%
0.6%
PPNB
2.8%
0.6%



Dinka
-1.9
68,671



Biaka
0.263704374
2
Mbuti
72.1%
2.4%
Mende
27.9%
2.4%



Iran_Neolithic
-1.9
68,488



Hadza
0.02738864
2
Dinka
19.3%
8.5%
Ethiopia_4500BP
80.7%
8.5%



Ust_Ishim
2.7
69,413



Masai
0.014608198
2
Dinka
50.9%
2.1%
Tanzania_Luxmanda_3100BP
49.1%
2.1%



Ami
-2.4
69,646



Datog
0.128763111
2
Dinka
34.0%
3.0%
Tanzania_Luxmanda_3100BP
66.0%
3.0%



Mixe
-2.0
75,739



Kikuyu
0.018943915
2
Mende
62.8%
2.0%
Tanzania_Luxmanda_3100BP
37.2%
2.0%



Dinka
-3.0
77,057



Ju_hoan_North
0.050248981
2
South_Africa_2000BP
90.9%
1.2%
Tanzania_Luxmanda_3100BP
9.1%
1.2%



Denisova
3.3
66,308



Ju_hoan_South
0.060549412
2
South_Africa_2000BP
89.5%
1.2%
Tanzania_Luxmanda_3100BP
10.5%
1.2%



Denisova
2.9
72,546



Taa_North
0.048898318
2
South_Africa_2000BP
88.1%
1.2%
Tanzania_Luxmanda_3100BP
11.9%
1.2%



Denisova
2.7
69,475



Naro
0.092694084
2
South_Africa_2000BP
88.0%
1.2%
Tanzania_Luxmanda_3100BP
12.0%
1.2%



Denisova
2.5
70,640



Afar_WGA
0.035029029
2
Ganj_Dareh_Iran_Neolithic
21.1%
2.5%
Tanzania_Luxmanda_3100BP
78.9%
2.5%



Dinka
-3.9
72,683



Luhya
0.588961185
2
Dinka
39.8%
5.9%
Mende
60.2%
5.9%



South_Africa_2000BP
1.6
72,591



Gana
0.177044532
2
Dinka
38.4%
1.6%
South_Africa_2000BP
61.6%
1.6%



Anatolia_Neolithic
-1.5
71,036



Khwe
0.08673416
2
Dinka
62.1%
1.5%
South_Africa_2000BP
37.9%
1.5%



Ethiopia_4500BP
2.1
70,886



Xuun
0.047257056
2
Dinka
29.9%
1.6%
South_Africa_2000BP
70.1%
1.6%



Denisova
1.6
66,801



Hoan
0.183725467
2
Dinka
27.8%
1.9%
South_Africa_2000BP
72.2%
1.9%



Mixe
1.5
70,984



Taa_West
0.141966574
2
Dinka
18.5%
1.8%
South_Africa_2000BP
81.5%
1.8%



Denisova
2.0
67,814



Taa_East
0.023684896
2
Dinka
22.1%
1.9%
South_Africa_2000BP
77.9%
1.9%



Mixe
2.5
70,560



Mbuti
0.144248841
2
Mende
47.5%
2.7%
South_Africa_2000BP
52.5%
2.7%



Iran_Neolithic
2.4
71,998



Damara
0.684312915
2
Mende
92.4%
1.8%
South_Africa_2000BP
7.6%
1.8%



Denisova
1.7
70,908



BantuSA
0.749653639
2
Mende
80.6%
1.7%
South_Africa_2000BP
19.4%
1.7%



Denisova
2.2
72,668



Tswana
0.643588346
2
Mende
79.0%
1.8%
South_Africa_2000BP
21.0%
1.8%



Denisova
2.5
72,673



Kgalagadi
0.250834567
2
Mende
65.3%
1.7%
South_Africa_2000BP
34.7%
1.7%



Denisova
3.1
72,699



Somali
0.015401137
3
Dinka
21.7%
3.5%
Ganj_Dareh_Iran_Neolithic
16.2%
2.6%
Tanzania_Luxmanda_3100BP
62.1%
5.6%
Mende
-2.1
68,793



Burbur_WGA
0.114531304
3
Anatolia_Neolithic
1.3%
58.1%
Dinka
19.7%
5.0%
PPNB
79.0%
62.8%
Iran_Neolithic
2.0
73,083



Gui
0.034545878
3
Mbuti
27.1%
9.2%
South_Africa_2000BP
61.4%
8.0%
Tanzania_Luxmanda_3100BP
11.5%
1.8%
Mixe
2.1
70,837



Shua
0.093270308
3
Mende
49.5%
2.8%
South_Africa_2000BP
35.8%
1.8%
Tanzania_Luxmanda_3100BP
14.8%
1.5%
Denisova
2.2
68,906



Khomani
0.100823752
3
Dinka
13.9%
3.3%
South_Africa_2000BP
64.5%
1.9%
Tanzania_Luxmanda_3100BP
21.5%
2.0%
Ethiopia_4500BP
1.6
67,477



Sandawe
0.045962257
3
Dinka
37.2%
3.6%
Mbuti
22.4%
2.3%
Tanzania_Luxmanda_3100BP
40.4%
2.0%
Punjabi
-1.7
66,718

[/tbody]


The one for Somalis in particular:





Somali
0.015401137
3
Dinka
21.7%
3.5%
Ganj_Dareh_Iran_Neolithic
16.2%
2.6%
Tanzania_Luxmanda_3100BP
62.1%
5.6%
Mende
-2.1
68,793

Awale
09-22-2017, 11:55 AM
Yeah, I finally got to that part:


Western-Eurasian-related ancestry is pervasive in eastern Africa today (Pagani et al., 2012, Tishkoff et al., 2009), and the timing of this admixture has been estimated to be ∼3,000 BP on average (Pickrell et al., 2014). We found that the ∼3,100 BP individual (Tanzania_Luxmanda_3100BP), associated with a Savanna Pastoral Neolithic archeological tradition, could be modeled as having 38% ± 1% of her ancestry related to the nearly 10,000-year-old pre-pottery farmers of the Levant (Lazaridis et al., 2016), and we can exclude source populations related to early farmer populations in Iran and Anatolia. These results could be explained by migration into Africa from descendants of pre-pottery Levantine farmers or alternatively by a scenario in which both pre-pottery Levantine farmers and Tanzania_Luxmanda_3100BP descend from a common ancestral population that lived thousands of years earlier in Africa or the Near East. We fit the remaining approximately two-thirds of Tanzania_Luxmanda_3100BP as most closely related to the Ethiopia_4500BP (p = 0.029) or, allowing for three-way mixture, also from a source closely related to the Dinka (p = 0.18; the Levantine-related ancestry in this case was 39% ± 1%) (Table S4).

While these findings show that a Levant-Neolithic-related population made a critical contribution to the ancestry of present-day eastern Africans (Lazaridis et al., 2016), present-day Cushitic speakers such as the Somali cannot be fit simply as having Tanzania_Luxmanda_3100BP ancestry. The best fitting model for the Somali includes Tanzania_Luxmanda_3100BP ancestry, Dinka-related ancestry, and 16% ± 3% Iranian-Neolithic-related ancestry (p = 0.015). This suggests that ancestry related to the Iranian Neolithic appeared in eastern Africa after earlier gene flow related to Levant Neolithic populations, a scenario that is made more plausible by the genetic evidence of admixture of Iranian-Neolithic-related ancestry throughout the Levant by the time of the Bronze Age (Lazaridis et al., 2016) and in ancient Egypt by the Iron Age (Schuenemann et al., 2017).

Seems rather implausible. I'd like to see what we'll get with Treemix and nMonte, though. But I'm not surprised that there's something more basal about the South-Erythraeaic speaking pastoralist (http://anthromadness.blogspot.com/2015/05/south-cushitic-admixture-in-southeast.html)'s West-Eurasian ancestry. Lank and I noticed several years ago with ADMIXTURE alone that there was something minutely different about the admixture in Southeast Africans and Somalis (Somalis sometimes looked a little more "Med", for example) but 16% extra Neolithic-Zagrosian and some extra Dinka-like ancestry makes no sense. For one, our region is mostly, not entirely, a dead-zone for Western-African related ancestry so direct admixture from full-on Dinka-like people after this South-Erythraeic speaker's ancestors left the Horn makes little sense so maybe it's just a sign of this pastoralist being admixed somehow with extra Mota-like ancestry, idk. The 16% extra Neolithic-Zagrosian makes no sense, though. The only things that could probably cause that would the Proto-Ethiopiansemitic speaking community and perhaps also admixture acquired during the Islamic period but neither make sense because:

1) Somalis never language-shifted. How on Earth are we carrying so much ancient Zagrosian ancestry possibly owed to the Proto-Ethiopiansemitic speaking community, especially while other analyses and lines of evidence (like uniparentals) never picked up on it?

2) There's just been no proof in either Uniparentals or auDNA up to this point that our ancestors notably out-mixed during the Islamic Period (~700 CE and beyond).


So I wonder what caused this if it's legitimate admixture. Weird stuff, I gotta say. But really interesting. On another note, it's great to finally see what archaeology, linguistics, even craniometrics and modern DNA have surmised for a while now backed up by aDNA:


We used our modeling framework to show that the South_Africa_1200BP pastoralist individual from the western Cape is consistent with being a mixture of just two streams of ancestry relative to non-southern African populations, with 40.3% ± 2.3% ancestry related to the Tanzania_Luxmanda_3100BP individual (54% ± 7% when restricting analysis to sequences with postmortem damage) and the remainder being related to the South_Africa_2000BP hunter-gatherers (Table S5). This supports the hypothesis that the Savanna Pastoral Neolithic archaeological tradition (http://anthromadness.blogspot.com/2015/07/the-savanna-pastoral-neolithic-cushitic.html) in eastern Africa is a plausible source for the spread of herding to southern Africa. Even the Ju|'hoan San group with the least genetic affinity to eastern Africans (Ju_hoan_North), have 9% ± 1% of their ancestry most closely related to Tanzania_Luxmanda_3100BP, consistent with previous findings that the ancestries of all present-day San and Khoe were affected by agro-pastoralist migrations in the last two millennia (Pickrell et al., 2014).

Lank
09-22-2017, 02:14 PM
Let's see how well Pagani's 3000 YBP admixture date (repeated ad nauseum by academia) holds up when someone looks at this genome? :)

This study is a great first step, nice to see an early Cushite exploring Tanzania. Regarding the Somali admixture model, wildly speculating here, but perhaps the model is the problem? The Tanzanian may not be representative of the Cushitic ancestry in Somalis, forcing the addition of implausible, Nilotic and prehistoric Iran-related ancestry. Modern day South Cushites, who still retain a very high level of ancestry similar to this Tanzanian, show admixture with local hunter-gatherers, and vice versa. If the Tanzanian from 3100 ya already had this admixture, he could be a poor fit for the Cushitic ancestry of Somalis.

It would have been a nice start to have an ADMIXTURE analysis with more than just Sardinians as a proxy for Eurasian ancestry... Makes it a bit difficult to differentiate between ancient Levant and Iran/CHG-related ancestries.

Awale
09-22-2017, 02:26 PM
Let's see how well Pagani's 3000 YBP admixture date (repeated ad nauseum by academia) holds up when someone looks at this genome? :)

Inshallah. I was actually surprised they didn't immediately go "Pagani-boi was wrong, baby." in this new paper but I think what held them up is that Somalis look to have a later layer of admixture based on their models. If Somalis and this Tanzanian were identical; they might've claimed Pagani et al.'s findings are implausible since we have a 3,100ybp sample from Tanzania with all the same admixture as modern Erythraeic speaking Horn-Africans like Somalis. But, by their findings, this is still roughly the case as the belief Pagani et al. spread was that all the West-Eurasian admixture hit around 3,000ybp but now, even by their models, it's clear the majority of it is shared with this Tanzanian sample and is most likely much older than the likely ES-speaking-community related signal Pagani et al. noticed all those years ago.


This study is a great first step, nice to see an early Cushite exploring Tanzania. Regarding the Somali admixture model, wildly speculating here, but perhaps the model is the problem? The Tanzanian may not be representative of the Cushitic ancestry in Somalis, forcing the addition of implausible, Nilotic and prehistoric Iran-related ancestry. Modern day South Cushites, who still retain a very high level of ancestry similar to this Tanzanian, show admixture with local hunter-gatherers, and vice versa. If the Tanzanian from 3100 ya already had this admixture, he could be a poor fit for the Cushitic ancestry of Somalis.

It would have been a nice start to have an ADMIXTURE analysis with more than just Sardinians as a proxy for Eurasian ancestry... Makes it a bit difficult to differentiate between ancient Levant and Iran/CHG-related ancestries.

Agreed with everything written above.

blackflash16
09-22-2017, 08:12 PM
From a cave near the Indian Ocean, north of Mombasa:



I0595, Kenya_400BP

qpAdm: 100% Mota

The individual was a tall, robust young adult male. He was buried in a shallow grave in a crouched position with two hands and one foot in the small of the back and the skull disarticulated and placed by the knees. The individual was buried by sediment containing marine shell beads, small knapped stone tools, and Tana Tradition potsherds. The associated faunal remains are exclusively wild, with the exception of a single possible caprine bone. Large numbers of remains of birds, rodents, and other microfauna suggest that the cave may have only been sporadically occupied when the human remains were deposited. We infer from the material culture and fauna that the cave was occupied by foragers during the time the individual was buried, although food producers were present at nearby settlements such as Mtsengo and Mbuyuni (Helm, 2000). Crop remains of African sorghum, pearl millet and finger millet found at the site suggest these foragers had access to agricultural resources.


Two Individuals from the same cave on Pemba Island:


I0589, Tanzania_Pemba_1400BP.

qpAdm: 100% Mota

This limestone cave at the northern end of Pemba Island in the Zanzibar archipelago has been excavated in multiple campaigns, the most recent two seasons conducted by the ERC-funded Sealinks Project at Oxford University in 2012 and then the Max Planck Institute for the Science of Human History in 2016 (unpublished; see also (Chami et al., 2009)). The sequence shows clear evidence of
human occupation beginning around 1400 BP with an escargotie`re layer of giant African land snail shells, pottery, and disarticulated human remains. Above this layer, the sequence shows regular human use of the cave into the last thousand years.


I2298, Tanzania_Pemba_600BP

qpAdm: 100% Mande

A lower molar from context 301 (Sealinks Project faunal catalog no. 15624), which lies just below the surface and was dated to 639-544 BP (623 ± 20 uncalibrated BP, Wk-43308). Both specimens are associated with a highly unusual faunal assemblage, dominated by fragmented crocodile (Crocodylus cf. niloticus) remains and diverse microfauna, including Rattus rattus (Asian black rat), a
nonnative rodent that must have arrived to the area via maritime exchange routes. There are no taphonomic indicators in the faunal assemblage of hunting by humans, nor of crocodile predation on humans. During both of the occupational phases targeted in this study, there were nearby settlements occupied by farmers whose ancestors likely came from the mainland, for example at the sites of Tumbe (c. 1400-1000 BP) and Chwaka (1000-400 BP) (Fleisher and LaViolette, 2013).


From a cave on Zanzibar Island:


I0589, Tanzania_Zanzibar_1400BP

qpAdm: 69% Mota + 31% South_Africa_2000BP

A complete second phalanx of an adult (Sealinks Project faunal catalog no. 4353). It was recovered from context 1011, in association with local Tana Tradition ceramics typical of the MIA, moderately-sized limestone lithic artifacts, and diverse wild game animals, but no additional human remains (Prendergast et al., 2016). The specimen is directly dated to 1370-1303 BP (1479 ± 23 uncalibrated BP, OxA-31427), thus placing it at the beginning of the MIA phase. While Kuumbi Cave is interpreted as a forager site,
elsewhere on the island at this time, large settlements such as Unguja Ukuu emerge, occupied by farmers whose origins are likely on the mainland (Crowther et al., 2015; Juma, 2004).


East, "West" and "South" African ancestry in the Zanzibar Archipelago within the last 1500 years, along with The Hadza and Cushitic pastoralists on the Tanzanian mainland. SSA party:beerchug:

Shouldn't this thread be in the aDNA sub-forum?

blackflash16
09-23-2017, 05:16 PM
The Kenya_500BP individual has |Z|>3 indication of Levant_N admixture (relative to Mota), his assigned Y-DNA is E1b1b:


Kenya_500BP Mota Levant_N Chimp 0.0401 3.643 4219 3894 74432

blackflash16
09-23-2017, 10:50 PM
Here are some more d-stats for the Kenyan:


Kenya_500BP Mota Tanzania_Luxmanda_3000BP Chimp 0.0251 2.038 4060 3861 74507
Kenya_500BP Mota Somali Chimp 0.0175 2.040 4834 4668 89810
Kenya_500BP Mota Tanzania_Pemba_700BP Chimp 0.0018 0.139 3277 3265 65242
Kenya_500BP Mota Masai Chimp 0.0062 0.733 4707 4649 89810




Tanzania_Pemba_1400BP Mota Levant_N Chimp -0.0184 -1.398 2744 2846 52742
Tanzania_Luxmanda_3000BP Somali Mota Chimp -0.0090 -1.970 22038 22437 457675
Tanzania_Luxmanda_3000BP Somali Tanzania_Pemba_1400BP Chimp 0.0169 1.555 2532 2448 51969
Tanzania_Luxmanda_3000BP Somali Kenya_500BP Chimp 0.0093 1.044 3728 3659 74519
Tanzania_Luxmanda_3000BP Somali Hadza Chimp -0.0123 -3.509 22149 22700 457767

blackflash16
09-24-2017, 10:32 PM
I uploaded the Luxmanda sample to gedmatch (Kit #: Z626350 ) but I'm not sure I did it right. The calculator results seem to be okay:

Near East Neolithic K13:



Admix Results (sorted):

# Population Percent
1 SUB_SAHARAN 50.15
2 NATUFIAN 30.41
3 ANATOLIA_NEOLITHIC 11.04
4 IRAN_NEOLITHIC 1.51
5 EHG 1.43
6 SHG_WHG 1.41
7 SIBERIAN 1.36
8 PAPUAN 1.33
9 SE_ASIAN 1.23





Least-squares method.

Using 1 population approximation:
1 Somali @ 19.217751
2 Masai @ 22.123108
3 Moroccan @ 32.525169
4 Hadza @ 33.111984
5 Saharawi @ 33.831158
6 Algerian @ 36.990288
7 Mota @ 37.305138
8 Libyan @ 41.139465
9 Yemeni @ 42.639397
10 Egyptian @ 44.120411
11 BedouinA @ 49.894337
12 Jordanian @ 53.766560
13 Palestinian @ 53.870270
14 Jew_Libyan @ 55.550819
15 Gambian @ 55.836201
16 Syrian @ 55.932667
17 Saudi @ 55.962032
18 Jew_Yemenite @ 56.303688
19 Jew_Tunisian @ 56.830215
20 Lebanese @ 57.033176

Using 2 populations approximation:
1 50% Gambian +50% Levant_BA @ 8.558372


Using 3 populations approximation:
1 50% Hadza +25% Levant_N +25% Somali @ 5.534059


Using 4 populations approximation:
++++++++++++++++++++++
1 Hadza + Levant_N + Mota + Somali @ 5.357584
2 Hadza + Hadza + Levant_N + Somali @ 5.534059
3 Levant_N + Mota + Mota + Somali @ 5.573419
4 Gambian + Levant_N + Masai + Somali @ 5.648787
5 Hadza + Levant_N + Masai + Masai @ 5.673450
6 Esan + Levant_BA + Masai + Saharawi @ 5.909684
7 Levant_BA + Masai + Saharawi + Yoruba @ 5.909684
8 Esan + Hadza + Levant_BA + Saharawi @ 5.969609
9 Hadza + Levant_BA + Saharawi + Yoruba @ 5.969609
10 Gambian + Hadza + Levant_BA + Saharawi @ 5.981442
11 Levant_N + Masai + Masai + Mota @ 6.030968
12 Levant_N + Masai + Masai + Masai @ 6.082205
13 Esan + Levant_N + Masai + Somali @ 6.100034
14 Levant_N + Masai + Somali + Yoruba @ 6.100034
15 Esan + Hadza + Levant_N + Saharawi @ 6.359837
16 Hadza + Levant_N + Saharawi + Yoruba @ 6.359837
17 Gambian + Levant_BA + Masai + Saharawi @ 6.407068
18 Esan + Hadza + Levant_N + Libyan @ 6.505563
19 Hadza + Levant_N + Libyan + Yoruba @ 6.505563
20 Gambian + Hadza + Levant_N + Saharawi @ 6.516825

Done.

Elapsed time 0.2023 seconds.



But the one-to-one comparison doesn't at 50/1:



Largest segment = 2.3 cM
Total of segments > 1 cM = 35.2 cM
27 matching segments


I'm guessing there's more that has to be done to the files before uploading for one-to-one comparisons (rather than simply converting to 23andMe format).

drobbah
09-24-2017, 11:20 PM
I uploaded the Luxmanda sample to gedmatch (Kit #: Z626350 ) but I'm not sure I did it right. The Calculator results seem to be okay:

Near East Neolithic K13:





But the one-to-one comparison doesn't at 50/1:

I ran him through Eurogenes K13 and he seems more East Med and less Red Sea than us Somalis but still quite similar to us.




Admix Results (sorted):

# Population Percent
1 Northeast_African 55.81
2 Red_Sea 18.89
3 East_Med 14.02
4 Sub-Saharan 7.41
5 West_Med 1.89
6 East_Asian 1.17


Finished reading population data. 204 populations found.
13 components mode.

--------------------------------

Least-squares method.

Using 1 population approximation:
1 Ethiopian_Oromo @ 2.808020
2 Ethiopian_Wolayta @ 4.379772
3 Somali @ 6.751717
4 Ethiopian_Amhara @ 14.380640
5 Ethiopian_Tigray @ 16.168303


Using 2 populations approximation:
1 50% Ethiopian_Oromo +50% Ethiopian_Wolayta @ 2.583546


Using 3 populations approximation:
1 50% Ethiopian_Oromo +25% Ethiopian_Oromo +25% Ethiopian_Wolayta @ 2.343076


Using 4 populations approximation:
+++++++++++++++++++++++++++++++
1 Ethiopian_Gumuz + Ethiopian_Tigray + Ethiopian_Tigray + Ethiopian_Wolayta @ 2.166793
2 Ethiopian_Gumuz + Ethiopian_Tigray + Ethiopian_Tigray + Somali @ 2.224057
3 Ethiopian_Oromo + Ethiopian_Oromo + Ethiopian_Oromo + Ethiopian_Wolayta @ 2.343076
4 Ethiopian_Amhara + Ethiopian_Gumuz + Ethiopian_Tigray + Ethiopian_Wolayta @ 2.468127

beyoku
09-25-2017, 10:13 AM
What is to be said about his 40% Non African yet his PCA clustering among Nilotes?

SWAHILLI_PRINCE16
09-25-2017, 02:06 PM
Deleted

SWAHILLI_PRINCE16
09-25-2017, 02:08 PM
Alhamdulillah i was looking for something like that :D , based on thie qpAdm table what are Tanzanian Pembas and Zanzibaris mixed with as its not clear to me

Megalophias
09-25-2017, 03:35 PM
What is to be said about his 40% Non African yet his PCA clustering among Nilotes?

Where else should she cluster? There's no Horners or Southern Cushites on the PCA and no Eurasian-African axis.

"We used smartpca (Patterson et*al., 2006) to compute principal components using all transversion and transitions SNPs, and the present-day populations shown in Figure*1 and Figure*S2. We projected the ancient individuals the option lsqproject: YES, on eigenvectors computed using the present-day populations on the Affymetrix Human Origins array. To deal with the confounder factor of recent admixture with western Eurasian-related populations on the PCA, we removed all northern Africans, eastern African Cushitic speakers, Nama, and 8 Khomani individuals that had 5% or more cluster membership in the shared with Europeans in an ADMIXTURE analysis."

Lank
09-25-2017, 03:56 PM
I ran the Tanzanian through GEDmatch to check if she does indeed have local East African HG admixture, as modern South Cushites do. Modern East African HGs tend to show significant Khoisan-like ancestry in ADMIXTURE, and the antiquity of this affinity has been confirmed by the ancient southern-eastern cline discovered in this paper. So San-like affinities can presumably be used as a proxy for this admixture.

PuntDNAL K13:



# Population Percent
1 East_Africa 43.87
2 SW_Asia 35
3 South_Africa 10.86
4 SW_Europe 3.72
5 West_Africa 1.93
6 West_Asia 1.73
7 NE_Asia 1.2
8 Oceania 0.91
9 Americas 0.4
10 NE_Europe 0.39

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 89.4% Somali + 10.6% San @ 3.06
2 87.5% Somali + 12.5% Ju_Hoan @ 3.46
3 91.1% Oromo + 8.9% San @ 3.62
4 89.6% Oromo + 10.4% Ju_Hoan @ 4.37
5 85.6% Somali + 14.4% Mbuti_Pygmy @ 5.8
6 88.5% Oromo + 11.5% Mbuti_Pygmy @ 6.61
7 66.6% Wolayta + 33.4% Tigray @ 7.11
8 62.7% Wolayta + 37.3% Amhara @ 7.16
9 89.5% Wolayta + 10.5% Bedouin @ 7.2
10 74% Aricultivator + 26% Bedouin @ 7.35



puntDNAL K15 Oracle results (part of the San-affiliated stuff is probably absorbed by Omo River here):



# Population Percent
1 Horn_Of_Africa 35.32
2 Wht_Nile_River 25.21
3 Omo_River 10.7
4 SW_Asian 10.49
5 S_African 7.87
6 Mediterranean 6.1
7 W_African 2.52
8 E_Asian 1.29
9 Amerindian 0.5



Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 93.5% Oromo + 6.5% San @ 6.07
2 92.1% Oromo + 7.9% Gumuz @ 8.48
3 55.4% Afar + 44.6% Wolayta @ 8.5
4 82.5% Oromo + 17.5% Maasai @ 8.59
5 84.8% Oromo + 15.2% Sandawe @ 8.59
6 92.5% Oromo + 7.5% Anuak @ 8.6
7 92.8% Oromo + 7.2% South_Sudanese @ 8.66
8 60.9% Amhara + 39.1% Wolayta @ 8.78
9 94.2% Oromo + 5.8% Pygmy @ 8.82
10 57.3% Eritrean + 42.7% Wolayta @ 8.84

Awale's K15 by comparison:



Horn_Of_Africa: 51.32
Wht_Nile_River: 27.69
SW_Asian: 6.47
W_African: 4.42
Omo_River: 3.86
Mediterranean: 3.67
NE_European: 1.56


This confirms some local HG admixture in this early Tanzanian pastoralist, albeit with predominantly Cushitic affinities. No doubt the figure of HG admixture would rise (to perhaps ~20%?) if some East African HG references without recent admixture were used. This may explain the weird Somali mixture model.

beyoku
09-25-2017, 06:10 PM
Where else should she cluster? There's no Horners or Southern Cushites on the PCA and no Eurasian-African axis.

"We used smartpca (Patterson et*al., 2006) to compute principal components using all transversion and transitions SNPs, and the present-day populations shown in Figure*1 and Figure*S2. We projected the ancient individuals the option lsqproject: YES, on eigenvectors computed using the present-day populations on the Affymetrix Human Origins array. To deal with the confounder factor of recent admixture with western Eurasian-related populations on the PCA, we removed all northern Africans, eastern African Cushitic speakers, Nama, and 8 Khomani individuals that had 5% or more cluster membership in the shared with Europeans in an ADMIXTURE analysis."

I assumed an edge position, not in between Hadza and Nilotes.

Megalophias
09-25-2017, 07:49 PM
I assumed an edge position, not in between Hadza and Nilotes.
Well, there's a ton of Hadza on it, so they can dominate one pole of the variation. So my guess would be the ancient South/East African cline is pointing to the "not West African but not actually Hadza" position in the east-west axis. I could be wrong, someone who knows more about PCAs can set me straight here, but I think that makes sense.

Would be nice to see some PCAs with different sample sets to compare.

Awale
09-25-2017, 09:23 PM
David sent the Sheikh the PCA coordinates for these samples like 20 hours ago (haven't been checking my email lately, lol):

,PC1,PC2,PC3,PC4,PC5,PC6,PC7,PC8,PC9,PC10
Kenya_500BP:I0595,-0.079,0.0048,0.0013,0.01,0.0219,0.0113,-0.0142,-0.0036,-0.0211,-0.0115
Malawi_Fingira_2500BP:I4426,-0.1085,0.003,0.001,-0.0227,0.077,0.0042,-0.0107,-0.0033,-0.0139,-0.0117
Malawi_Hora_Holocene:I2966,-0.0992,0.0036,-0.0013,-0.0204,0.074,0.0057,-0.0084,-0.0021,-0.0069,-0.0132
South_Africa_1300BP:I9134.SG,-0.0942,0.008,0.0012,-0.0139,0.0737,0.0092,-0.0138,-0.0007,-0.0157,-0.0049
South_Africa_2000BP:I9028.SG,-0.1204,0.0035,0.0007,-0.0399,0.113,0.0101,-0.0104,0.0025,-0.0013,-0.0047
South_Africa_2000BP:I9133.SG,-0.1201,0.0031,0.0006,-0.0386,0.1128,0.0107,-0.0128,0.0021,-0.0004,-0.0041
Tanzania_Luxmanda_3000BP:I3726,-0.0542,0.013,-0.0013,0.0311,0.0016,0.0115,-0.0273,0.0013,-0.0344,-0.015
Tanzania_Pemba_1400BP:I1048,-0.0845,0.0068,-0.0021,-0.0043,0.0405,0.0045,-0.0144,0.001,-0.0241,-0.0298
Tanzania_Pemba_700BP:I2298,-0.1137,0.0061,0.0067,-0.003,-0.0238,-0.0051,0.0087,0.0033,-0.0027,-0.0349
Tanzania_Zanzibar_1400BP:I0589,-0.0953,0.0027,-0.0002,-0.0068,0.0477,0.0024,-0.0107,0.0003,-0.0201,-0.0157

Going to start nMonte-ing real soon and share the results. Thanks a lot, David.

Awale
09-25-2017, 10:32 PM
Seems like this is the best fitting model:


[1] "distance%=0.7378 / distance=0.007378"

Tanzania_Luxmanda_3000BP:I3726

Somali 83.50
South_Africa_1300BP:I9134.SG 11.45
BedouinB 5.05

Not nearly a perfect fit but quite decent. What has me the most puzzled is the Bedouin-B which I thought to add after seeing how she keeps looking Oromo in some oracles. If this is legit, think we have our confirmation, perhaps, that there really was an admixture event from further up north and into the highlands before the arrival of the Proto-ES speaking community? Kinda makes you wonder how long Somalis' ancestors lived in Greater-Somalia to have supposedly avoided such admixture. Anyway, I added all of the following samples and populations into the input file for the above model, for those wondering:

Somali,-0.05288,0.01508,-0.00027,0.0339,-0.01062,0.00914,-0.026,0.00113,-0.03164,-0.01953
Mota,-0.0875,0.0024,-0.0015,0.0028,0.0181,-0.0036,-0.0126,0,-0.0316,-0.0257
Ju_hoan_North,-0.1192,0.00418,0.002,-0.03104,0.09868,0.0098,-0.01192,0.00276,-0.0075,-0.00792
Ethiopian_Ari_blacksmith,-0.0742167,0.00795,0.0003333,0.0167167,0.0090167,0. 0032,-0.0181667,0.0014667,-0.0274,-0.0199667
BedouinB,0.00626,0.02914,-0.00346,0.0624,0.00714,0.01396,-0.05358,0.00496,-0.00756,-0.00778
South_Africa_1300BP:I9134.SG,-0.0942,0.008,0.0012,-0.0139,0.0737,0.0092,-0.0138,-0.0007,-0.0157,-0.0049
South_Africa_2000BP:I9028.SG,-0.1204,0.0035,0.0007,-0.0399,0.113,0.0101,-0.0104,0.0025,-0.0013,-0.0047
South_Africa_2000BP:I9133.SG,-0.1201,0.0031,0.0006,-0.0386,0.1128,0.0107,-0.0128,0.0021,-0.0004,-0.0041
BantuKenya,-0.1086,0.00535,0.0052,0.00385,-0.02915,-0.00145,0.0021,5.00E-05,-0.01405,-0.04665
Bantu_SA_Herero,-0.1146,0.0062,0.0062,-0.00245,-0.01895,-0.00295,0.0075,-0.00125,0.00275,-0.0252
Bantu_SA_Ovambo,-0.1153,0.0055,0.0066,-0.0052,-0.0238,-0.0039,0.0073,0.0013,-0.0028,-0.0277
Bantu_SA_Pedi,-0.1173,0.0048,0.0052,-0.0085,-0.004,-0.0011,0.0043,0.0019,-0.0038,-0.0237
Bantu_SA_S_Sotho,-0.1168,0.0054,0.004,-0.0074,-0.0036,-0.0019,0.0033,0.0017,0.0036,-0.0281
Bantu_SA_Tswana,-0.1167,0.00525,0.00495,-0.0093,-0.00015,-9.00E-04,0.00205,0.00355,0.00065,-0.02375
Bantu_SA_Zulu,-0.1156,0.0058,0.0057,-0.0055,-0.011,-0.0033,0.0022,-4.00E-04,-1.00E-04,-0.0197

I ran Oromos' average through the same run and got:


[1] "distance%=0.2222 / distance=0.002222"

Ethiopian_Oromo

Somali 89.15
South_Africa_1300BP:I9134.SG 4.10
BedouinB 3.65
Mota 2.00
Ju_hoan_North 1.10


Now this is how they look without all the HG populations (xMota):


[1] "distance%=0.298 / distance=0.00298"

Ethiopian_Oromo

Somali 78.25
Mota 14.20
BedouinB 7.55
.

[1] "distance%=0.3255 / distance=0.003255"

Ethiopian_Oromo

Somali 74.8
Ethiopian_Ari_blacksmith 19.4
Yemenite_Jew 5.7

So I threw out every possible source of HG ancestry other than Mota and Ari-Blacksmiths when dealing with Luxmandan and got:


[1] "distance%=0.9393 / distance=0.009393"

Tanzania_Luxmanda_3000BP:I3726

Somali 57.2
Ethiopian_Ari_blacksmith 23.8
BedouinB 11.1
Mota 7.9

A worse fit but not a model that makes that much sense to me. But the level of Ari/Mota-like ancestry got me curious so I revisited an old model I ran for the Masai_Kinyawa samples' average:


[1] "distance%=0.0918 / distance=0.000918"

Masai_Kinyawa

Somali 53.65
Dinka 43.40
Ju_hoan_North 2.95

And look at what we have here:


[1] "distance%=0.0887 / distance=0.000887"

Masai_Kinyawa

Somali 46.5
Dinka 42.6
Ethiopian_Ari_blacksmith 9.1
South_Africa_2000BP:I9028.SG 1.0
BedouinB 0.8

They had Ari/Mota-related ancestry all along? :\ But, to be fair, this model looks like ovefitting but it is still pretty interesting because now these sorts of results aren't isolated to the Luxmanda sample but we seem to find similar results in SE African people (the Masai) we know have admixture from South-Erythraeic speaking pastoralists. Wasn't expecting this, I'll say that much for now.

Lank
09-26-2017, 08:00 AM
Nice to see the mixture models being tested in nMonte. We know there is some sort of relationship between the Mota-related Ethiopian highland ancestry, and hunter-gatherers farther south, so it's not surprising that both groups can be used in the models. Although I think it's notable that it prefers the Khoisan populations for the Tanzanian pastoralist.

As you have noted in the past, there is some diffuse "San" affinity that is associated with Mota/Omotic-related ancestry. The authors of this paper even manage to fit some Hadza, as well as the Kenya 400BP and Tanzania Pemba 1400BP samples, as 100% Mota-related. These models look rather crude though, and Mota-related populations have vastly different uniparental markers from East African HGs sampled by Hirbo (e.g. the latter have lot of mtDNA L0f, also common in modern South Cushites, and interestingly found in 1 ancient Malawian). I would bet that with some new aDNA from East African HGs south of Ethiopia, or perhaps even the modern East African HGs sampled by Hirbo, they would be a much better fit for the HG ancestry in South Cushites, as well as the "Mota-related" ancient Kenyan/Tanzanian samples.

Awale
10-09-2017, 10:22 PM
As you have noted in the past, there is some diffuse "San" affinity that is associated with Mota/Omotic-related ancestry. The authors of this paper even manage to fit some Hadza, as well as the Kenya 400BP and Tanzania Pemba 1400BP samples, as 100% Mota-related. These models look rather crude though, and Mota-related populations have vastly different uniparental markers from East African HGs sampled by Hirbo (e.g. the latter have lot of mtDNA L0f, also common in modern South Cushites, and interestingly found in 1 ancient Malawian). I would bet that with some new aDNA from East African HGs south of Ethiopia, or perhaps even the modern East African HGs sampled by Hirbo, they would be a much better fit for the HG ancestry in South Cushites, as well as the "Mota-related" ancient Kenyan/Tanzanian samples.

I totally agree. Modelling the South-Erythraean as part South-Africa 1,300ybp fits better and makes more sense with other data than Omotic-speaker related ancestry. Also, the Maasai_Kinyawa actually fit quite well as Luxmanda + Dinka:

[1] "distance%=0.3802 / distance=0.003802"

Masai_Kinyawa

Tanzania_Luxmanda_3000BP 50.7
Dinka 49.3

Interestingly, however, when I added Somalis in, they shook up the results:

[1] "distance%=0.1731 / distance=0.001731"

Masai_Kinyawa

Dinka 47.1
Tanzania_Luxmanda_3000BP 26.9
Somali 25.9

Wonder if this points to these Masais having admixture from later East-Erythraeic speakers. Idk... Great comments as always, by the way.

beyoku
10-09-2017, 11:15 PM
I totally agree. Modelling the South-Erythraean as part South-Africa 1,300ybp fits better and makes more sense with other data than Omotic-speaker related ancestry. Also, the Maasai_Kinyawa actually fit quite well as Luxmanda + Dinka:

[1] "distance%=0.3802 / distance=0.003802"

Masai_Kinyawa

Tanzania_Luxmanda_3000BP 50.7
Dinka 49.3

Interestingly, however, when I added Somalis in, they shook up the results:

[1] "distance%=0.1731 / distance=0.001731"

Masai_Kinyawa

Dinka 47.1
Tanzania_Luxmanda_3000BP 26.9
Somali 25.9

Wonder if this points to these Masais having admixture from later East-Erythraeic speakers. Idk... Great comments as always, by the way.

Truth be told I envision Massai as having 3 successive waves of northern admixture. The last 2 being post Ethnogenisis. V32, V22, M293, and T2 IMO didn’t all travel in tandem into the Rift Valley and Great Lakes / Nile Basin.

Awale
10-27-2017, 06:06 AM
Well, my fellows, it seems like we have some new developments. I don't know what's changed about nMonte's results all of a sudden (if this is indeed a change) but my conversation with Agamemnon below should get everyone up to speed:


https://i.imgur.com/vQJ5bD7.png

.
.

https://i.imgur.com/S0KYh4w.png

.
.

https://i.imgur.com/LqfSPAG.png

.
.

https://i.imgur.com/OO7HRuM.png

.
.

https://i.imgur.com/RJtRJqx.png

.
.

https://i.imgur.com/XJubcZr.png

.
.

https://i.imgur.com/CgqRa5O.png

But yeah, it seems to me like every current day Erythræic and Ethiosemitic speaking group has seemingly post Bronze-Age, pre-historic Zagrosian ancestry carrying admixture (with Somalis having the least):


[1] "distance%=0.5453 / distance=0.005453"

Somali

Dinka 54.35
Natufian 34.40
Yemenite-Jew 9.80
Ari-Blacksmith 1.45

--

[1] "distance%=0.6277 / distance=0.006277"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 39.80
Natufian 31.85
Yemenite-Jew 22.40
Ari-Blacksmith 5.95

--

[1] "distance%=0.7816 / distance=0.007816"

Amhara

Dinka 39.20
Natufian 34.70
Yemenite-Jew 20.15
Ari-Blacksmith 5.95

--

[1] "distance%=0.958 / distance=0.00958"

Tigray-Tigrinya

Dinka 33.8
Natufian 32.6
Yemenite-Jew 20.1
Ari-Blacksmith 13.5

Whilst, like in the paper, and unlike before, the Luxmanda sample lacks this affinity and mostly just comes off as "Natufian-like + Dinka-like + South-African HG":

[1] "distance%=0.8087 / distance=0.008087"

Tanzania-Luxmanda-3000BP

Dinka 44.70
Natufian 42.45
South-Africa-1300BP 12.50
Yemenite-Jew 0.35

There are some oddities here other than me not recalling being able to get these results in the past (http://www.anthrogenica.com/showthread.php?11784-New-High-Coverage-Northeast-African-genomes&p=279988#post279988) like how the Agaw/"Afar" samples seem about as later West-Asian (http://anthromadness.blogspot.com/2016/01/the-later-west-eurasian-element-in-agaws.html) admixed as Habeshas when I recall them usually showing less (though they show somewhat less than Tigrinyas and Amharas when you use Negev Bedouin Bs instead of YJs). The higher Ari-like ancestry in Tigrinyas over Amharas is odd too but that might be skewed by an outlier among the Amharas and, as Lank in particular will remember, nMonte's had a hard time getting really decent Ari-like estimates for any group beyond Oromos, as far as I've seen. There's also the odd fact that even Aris are now showing some of this later admixture:

[1] "distance%=0.5979 / distance=0.005979"

Ari-Blacksmith

Dinka 52.2
South-Africa-1300BP 25.6
Natufian 13.4
Yemenite-Jew 8.8

-

[1] "distance%=0.896 / distance=0.00896"

Ari-Cultivator

Dinka 53.65
South-Africa-1300BP 23.60
Natufian 13.80
Yemenite-Jew 8.95

Has me surprised (but I guess it makes vague sense of all the Y-DNA J they have and even some of their mtDNA lineages (http://anthromadness.blogspot.com/2016/06/omotic-speakers-uniparental-data.html)?). I'm still thinking that something odd's amok but if this is legit; Somalis indeed do have some later West-Asian ancestry but, like I said to Agamemnon, these estimates are more acceptable to me than the ones gained via qpAdm. There is another difference as well in that nMonte clearly favors Natufians over Levant_N as the deeper, and older, layer of West-Asian ancestry in the region.

And, finally, it does make sense that the Luxmanda sample would lack such ancestry and clearly points to how it's a later, post ~3kya addition into the horn.

Chad Rohlfsen
10-28-2017, 12:10 AM
I have a thread that may be of interest in Autosomal DNA.. I am looking for feedback on a project.

drobbah
10-28-2017, 03:52 AM
Is it possible this Yemenite-Jew like admixture (in Somalis,Afars etc) arrived at the same time camels where introduced to the Horn considering majority of experts claim that camels were domesticated in SE Arabia?

Or do you think it was similar to what happened in the Ethiopian highlands with Semitic speakers entering the Djibouti/Somaliland area but they weren't as successful as they did in the highlands of Ethiopia?
https://static-content.springer.com/image/art%3A10.1007%2Fs10437-015-9184-9/MediaObjects/10437_2015_9184_Fig3_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs10437-015-9184-9/MediaObjects/10437_2015_9184_Fig8_HTML.gif

Awale
10-28-2017, 03:21 PM
Made sense of why the Luxmanda sample was showing Negev-Bedouin-B before as well:

[1] "distance%=0.5565 / distance=0.005565"

Tanzania-Luxmanda-3000bp

Somali 79.4
South-Africa-1300bp 12.3
Natufian 8.3
Negev-Bedouin-B 0.0

-

[1] "distance%=0.7378 / distance=0.007378"

Tanzania-Luxmanda-3000bp

Somali 83.50
South-Africa-1300bp 11.45
Negev-Bedouin-B 5.05

I noticed a while ago that, somewhat differently from some other Arabian-Southern_Levantine groups, Negev Bedouin Bs seem very Natufian-shifted:


[1] "distance%=2.1982 / distance=0.021982"

Negev-Bedouin-B

Natufian 72.8
Iran_ChL 27.2
Levant_N 0.0
Dinka 0.0
Yoruba 0.0

-

[1] "distance%=0.8847 / distance=0.008847"

Yemenite_Jew

Natufian 49.5
Iran_ChL 40.4
Levant_N 10.1
Dinka 0.0
Yoruba 0.0



Notice the very poor fit for the Negevians, though. I'm reminded of North-Africans, actually. I think they may have something perhaps even more basal than Natufian in them or just some pre-historic element we don't have yet in the aDNA record that we'll need to improve their fits. But nevertheless, you can still see how Natufian-like much of their ancestry looks when compared to Yemenite Jews. In fact I bet this is why they often provide better fits for Horn-Africans in comparison to Yemenite Jews (given that we have a lot of Natufian-like ancestry as well based on nMonte).

So the reason the South-Erythraean was turning up as part Bedouin-B seems to be because they were standing in for the extra Natufian-like ancestry she has in comparison to Somalis. This is probably also why these Bedouin samples often provide better fits for Horners than Yemenite Jews do given that we too seem rather Natufian-shifted in these models. But, for the record, they're not actually as good an explanation for the levels of Zagros-Chl these groups are showing as Yemenite Jews are:

[1] "distance%=0.8847 / distance=0.008847"

Yemenite-Jew

Natufian 49.5
Iran_ChL 40.4
Levant_N 10.1
Dinka 0.0
Yoruba 0.0

According to the above, Yemenite Jews are about ~40% Zagros-Chal so for Somalis to be about ~10% Yemenite-Jew; they'd need to be about ~4% Zagros-Chl and that roughly is the case:

[1] "distance%=0.5261 / distance=0.005261"

Somali

Dinka 54.1
Natufian 40.2
Iran_ChL 4.4
Mota 1.3

In contrast, they'd need to be about >or=15% Negev-Bedouin-B to explain 4-5% Zagros-Chl but >or=15% is not actually what they end up with:

[1] "distance%=2.1982 / distance=0.021982"

Negev-Bedouin-B

Natufian 72.8
Iran_ChL 27.2
Levant_N 0.0
Dinka 0.0
Yoruba 0.0

[1] "distance%=0.5549 / distance=0.005549"

Somali

Dinka 53.20
Natufian 32.40
Negev-Bedouin-B 11.05
Mota 3.35

It's the same for other groups like Tigrinyas. ~20% Yemenite-Jew makes sense with their 8-9% Zagros-Chalcolithic:

[1] "distance%=0.9658 / distance=0.009658"

Tigray-Tigrinya

Natufian 46.65
Dinka 35.30
Mota 9.35
Iran_ChL 8.70

Doesn't necessarily mean the population had to be exactly like Yemenite Jews but something very similar is likely. Although, there is one weird issue in that none of these Horn-African groups are showing any Levant_N... Sure, there's some in the Zagros-related ancestry:

[1] "distance%=0.5399 / distance=0.005399"

Iran_Chl

Iran_N 37.1
Levant_N 31.6
CHG 31.3

But Yemenite Jews are seemingly north of 5% Levant_N even when you account for their Zagros-related ancestry so it's odd that both Somalis and Tigrinyas don't show any Levant_N until you remove Iran_Chl:

[1] "distance%=1.3007 / distance=0.013007"

Tigray-Tigrinya

Natufian 47.1
Dinka 33.1
Mota 14.8
Levant_N 4.9

But ehhh, I don't know how much you can really read into that.


Image 1

Hmm, I wouldn't connect this admixture to those particular medieval to sometimes late classical settlements in the interior of areas like Awdal. Those are unlikely to have ever been inhabited by actual South-Arabians and seem more like they were likely, to some degree, inhabited by Southern-Ethiosemitic speakers (i.e. speaking languages related to modern Harari, most likely) who seemingly did leave a linguistic impression on Somalis (loan words (http://www.academia.edu/5529034/2013_Strata_in_Semitic_loanwords_in_Northern_Somal i)) who probably assimilated them as time went by. I say they're not a likely source of this admixture because I've seen at least two Hararis' autosomal results (https://anthromadness.blogspot.com/2017/08/being-more-accurate-about-substrates-in.html) and they probably have significant Ari/Mota-related ancestry like Habeshas do, too much to have passed ~10% Yemenite-Jewish-like ancestry onto Somalis and then have Somalis turn up as 1-3% Ari/Mota-like at best.

Besides, those settlements only seem to have existed in stride along the northwest of the Peninsula, as far as I know; but, given Somalis' relative homogeneity (and even the fact that these samples I've used are from the northeast), this admixture, if it's real, is probably something much older.


Or do you think it was similar to what happened in the Ethiopian highlands with Semitic speakers entering the Djibouti/Somaliland area but they weren't as successful as they did in the highlands of Ethiopia?

This makes somewhat more sense at least. But yeah, I'm still real weirded and intrigued by all of this.

NetNomad
10-31-2017, 09:51 AM
Tanzania underwent many changes in ethnic makeup in such a short period of time. From HG to S-Cushitic to Nilotic to Bantu. Interesting.

Angoliga
10-31-2017, 12:02 PM
Tanzania underwent many changes in ethnic makeup in such a short period of time. From HG to S-Cushitic to Nilotic to Bantu. Interesting.

Good observation.

A slight correction in chronology though. Wouldn't the major waves of migration be HG to S-Cushitic to Bantu then to Nilotic?

NetNomad
10-31-2017, 12:49 PM
Good observation.

A slight correction in chronology though. Wouldn't the major waves of migration be HG to S-Cushitic to Bantu then to Nilotic?

I'm not too sure when Nilotes got to TZ. Just a hunch.

Awale
11-02-2017, 12:18 PM
So now that we have a Copt sample, I decided to try and compare Copts and Yemenite Jews in the Horner models:


Egyptian-Copts:


[1] "distance%=0.5741 / distance=0.005741"

Somali

Dinka 52.8
Natufian 35.2
Egyptian_Copt 9.2
Ari-Blacksmith 2.8

--

[1] "distance%=0.7129 / distance=0.007129"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 36.40
Natufian 33.15
Egyptian_Copt 21.80
Ari-Blacksmith 8.65

--

[1] "distance%=0.9187 / distance=0.009187"

Amhara

Natufian 37.45
Dinka 35.85
Egyptian_Copt 17.35
Ari-Blacksmith 9.35

--

[1] "distance%=1.0609 / distance=0.010609"

Tigray-Tigrinya

Natufian 35.0
Dinka 30.4
Egyptian_Copt 17.8
Ari-Blacksmith 16.7

--

[1] "distance%=0.5699 / distance=0.005699"

Somali

Dinka 52.75
Natufian 35.60
Egyptian_Copt 9.25
Mota 2.40

-

[1] "distance%=0.7066 / distance=0.007066"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 37.50
Natufian 34.35
Egyptian_Copt 22.20
Mota 5.95

--

[1] "distance%=0.916 / distance=0.00916"

Amhara

Natufian 38.75
Dinka 37.20
Egyptian_Copt 17.80
Mota 6.25

--

[1] "distance%=1.0457 / distance=0.010457"

Tigray-Tigrinya

Natufian 37.4
Dinka 32.5
Egyptian_Copt 18.5
Mota 11.6

--

[1] "distance%=0.5198 / distance=0.005198"

Ari-Blacksmith

Mota 63.4
Dinka 17.4
Natufian 13.8
Egyptian_Copt 5.5

--

[1] "distance%=0.4229 / distance=0.004229"

Ari-Cultivator

Mota 60.8
Dinka 19.3
Natufian 13.3
Egyptian_Copt 6.6

--

[1] "distance%=0.635 / distance=0.00635"

Ari-Blacksmith

Dinka 51.65
South_Africa_1300BP 25.75
Natufian 14.75
Egyptian_Copt 7.85

--

[1] "distance%=0.8989 / distance=0.008989"

Ari-Cultivator

Dinka 52.9
South_Africa_1300BP 23.8
Natufian 14.2
Egyptian_Copt 9.2


Yemenite-Jews:


[1] "distance%=0.5453 / distance=0.005453"

Somali

Dinka 54.35
Natufian 34.40
Yemenite_Jew 9.80
Ari-Blacksmith 1.45

--

[1] "distance%=0.6277 / distance=0.006277"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 39.80
Natufian 31.85
Yemenite_Jew 22.40
Ari-Blacksmith 5.95

--

[1] "distance%=0.7816 / distance=0.007816"

Amhara

Dinka 39.20
Natufian 34.70
Yemenite_Jew 20.15
Ari-Blacksmith 5.95

--

[1] "distance%=0.958 / distance=0.00958"

Tigray-Tigrinya

Dinka 33.8
Natufian 32.6
Yemenite_Jew 20.1
Ari-Blacksmith 13.5

--


[1] "distance%=0.5402 / distance=0.005402"

Somali

Dinka 53.65
Natufian 34.65
Yemenite_Jew 9.65
Mota 2.05

--

[1] "distance%=0.6102 / distance=0.006102"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 39.6
Natufian 32.6
Yemenite_Jew 22.6
Mota 5.2

--

[1] "distance%=0.7691 / distance=0.007691"

Amhara

Dinka 39.15
Natufian 35.50
Yemenite_Jew 20.30
Mota 5.05

--

[1] "distance%=0.9257 / distance=0.009257"

Tigray-Tigrinya

Dinka 34.5
Natufian 34.4
Yemenite_Jew 20.7
Mota 10.4

--

[1] "distance%=0.468 / distance=0.00468"

Ari-Blacksmith

Mota 62.75
Dinka 18.10
Natufian 11.90
Yemenite_Jew 7.25

--

[1] "distance%=0.4007 / distance=0.004007"

Ari-Cultivator

Mota 60.5
Dinka 19.9
Natufian 12.6
Yemenite_Jew 7.0

--

[1] "distance%=0.5979 / distance=0.005979"

Ari-Blacksmith

Dinka 52.2
South_Africa_1300BP 25.6
Natufian 13.4
Yemenite_Jew 8.8

--

[1] "distance%=0.896 / distance=0.00896"

Ari-Cultivator

Dinka 53.65
South_Africa_1300BP 23.60
Natufian 13.80
Yemenite_Jew 8.95


Egyptian-Copts & Yemenite-Jews both:


[1] "distance%=0.5453 / distance=0.005453"

Somali

Dinka 54.35
Natufian 34.40
Yemenite_Jew 9.80
Ari-Blacksmith 1.45
Egyptian_Copt 0.00

--

[1] "distance%=0.6271 / distance=0.006271"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 39.60
Natufian 31.75
Yemenite_Jew 20.90
Ari-Blacksmith 6.05
Egyptian_Copt 1.70

--

[1] "distance%=0.7816 / distance=0.007816"

Amhara

Dinka 39.20
Natufian 34.70
Yemenite_Jew 20.15
Ari-Blacksmith 5.95
Egyptian_Copt 0.00

--

1] "distance%=0.958 / distance=0.00958"

Tigray-Tigrinya

Dinka 33.8
Natufian 32.6
Yemenite_Jew 20.1
Ari-Blacksmith 13.5
Egyptian_Copt 0.0

--

[1] "distance%=0.5402 / distance=0.005402"

Somali

Dinka 53.65
Natufian 34.65
Yemenite_Jew 9.65
Mota 2.05
Egyptian_Copt 0.00

--

[1] "distance%=0.61 / distance=0.0061"

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 39.55
Natufian 32.60
Yemenite_Jew 21.60
Mota 5.20
Egyptian_Copt 1.05

--

[1] "distance%=0.7691 / distance=0.007691"

Amhara

Dinka 39.15
Natufian 35.50
Yemenite_Jew 20.30
Mota 5.05
Egyptian_Copt 0.00

--

[1] "distance%=0.9257 / distance=0.009257"

Tigray-Tigrinya

Dinka 34.5
Natufian 34.4
Yemenite_Jew 20.7
Mota 10.4
Egyptian_Copt 0.0

--

[1] "distance%=0.468 / distance=0.00468"

Ari-Blacksmith

Mota 62.75
Dinka 18.10
Natufian 11.90
Yemenite_Jew 7.25
Egyptian_Copt 0.00

--

[1] "distance%=0.4007 / distance=0.004007"

Ari-Cultivator

Mota 60.5
Dinka 19.9
Natufian 12.6
Yemenite_Jew 7.0
Egyptian_Copt 0.0

--

[1] "distance%=0.5979 / distance=0.005979"

Ari-Blacksmith

Dinka 52.2
South_Africa_1300BP 25.6
Natufian 13.4
Yemenite_Jew 8.8
Egyptian_Copt 0.0

--

[1] "distance%=0.8941 / distance=0.008941"

Ari-Cultivator

Dinka 53.3
South_Africa_1300BP 23.6
Natufian 13.6
Yemenite_Jew 5.6
Egyptian_Copt 3.8

So yeah, I'd prefer if we had more samples than Need4Speed and, of course, ancient samples from Egypt, South Arabia and even the Horn but, with what we have for now, it looks like Yemenite-Jews produce better fitting models and the model mostly favors them when both them and Need4Speed are thrown in (only the "Afar" and Ari-Cultivator samples showed any "Egyptian-Copt" when Yemenite-Jews were present alongside them and the latter only did so in the South-Africa-1300ybp model).

So idk but it, for now, looks like the later admixture in the Horn might be entirely of South-Arabian origins. Wouldn't be conclusive about this until we have the samples I mentioned, however.

Awale
11-05-2017, 08:41 AM
The whole East-South cline they noted in the Skoglund paper had me real intrigued, to be honest. Basically implies that, in the not-so-distant-past, there were more Mota-like people south of Sudan in East-Africa. As in, Hunter-Gatherer populations whose genomic make-up was some sort of crossing between "Eastern African" & "Southern African" and I ran a few nMonte runs on the Malawi samples to see if that was the case:

distance%=0.7331 / distance=0.007331

Malawi_Fingira_2500BP

South_Africa_2000BP 51.8
South_Africa_1300BP 29.6
Dinka 18.6

--

distance%=1.0779 / distance=0.010779

Malawi_Hora_Holocene

South_Africa_1300BP 50.0
South_Africa_2000BP 35.6
Dinka 14.3

I mean, it's terribly anachronistic to model 2,500ybp and 8,100ybp samples as mixtures between modern Dinkas and 1,300-2,000ybp Southern-Africans but it still seems pretty interesting and the fits aren't terribly bad, quite frankly. Though I'm surprised by how much they favor Southern-Africans given how intermediate they look in their clustering here:


http://www.cell.com/cms/attachment/2108833685/2082649523/gr1.jpg

Though, to be fair, Mota who clusters way more northwards than them looks this Southern-African-shifted:

distance%=1.5989 / distance=0.015989

Mota

Dinka 59.5
South_Africa_1300BP 40.5

(the poor fit is due to the weird, and broadly Eurasian, affinities Mota tends to show and because I didn't include any Eurasians in the run)

The remaining ancestry in the modern Hadza is quite similar as well once you account for their Horn-African (Erythraean) admixture:


distance%=1.1048 / distance=0.011048

Hadza

Dinka 46.5
South_Africa_1300BP 27.3
Tanzania_Luxmanda_3000BP 18.4
South_Africa_2000BP 7.7

--

distance%=0.9512 / distance=0.009512

Hadza

Dinka 39.5
Tanzania_Luxmanda_3000BP 24.0
Malawi_Fingira_2500BP 22.9
Malawi_Hora_Holocene 13.5

Kinda contradicts the paper, no:


We find that ancestry closely related to the ancient southern Africans was present much farther north and east in the past than is apparent today. This ancient southern African ancestry comprises up to 91% of the ancestry of Khoe-San groups today (Table S5), and also 31% ± 3% of the ancestry of Tanzania_Zanzibar_1400BP, 60% ± 6% of the ancestry of Malawi_Fingira_6100BP, and 65% ± 3% of the ancestry of Malawi_Fingira_2500BP (Figure 2A). Notably, the Khoe-San-related ancestry in ancient individuals from Malawi and Tanzania is symmetrically related to the two previously identified lineages present in the San (Z < 2; Figure S2), estimated to have diverged at least 20,000 years ago (Mallick et al., 2016, Pickrell et al., 2012, Schlebusch et al., 2012), implying that this was an ancient divergent branch of this group that lived in eastern Africa at least until 1,400 BP. However, it was not present in all eastern Africans, as we do not detect it in the ∼400-year-old individual from coastal Kenya nor in the present-day Hadza.

?

It's a shame the Bambuti Hunter-Gatherers from Congo (Mbuti) aren't in the Global-10 PCA, though. Would like to see what they could add to some of these models. Also, we're really in desperate need of really old ancient DNA from around Sudan, Chad and Kenya, to be honest. We need to find a more pristinely "Eastern-African" group.

Awale
11-06-2017, 07:58 AM
Okay... I just noticed something real interesting and odd yesterday while trying to make sure the later West-Asian influences showing in Somalis was real. First I tried to run the Luxmanda sample through a model where the input file contained only Dinkas, Natufians and Yemenite-Jews (realized I never ran this sort of model without South-Africa-1300ybp and got curious):

distance%=1.5042 / distance=0.015042

Tanzania-Luxmanda-3000ybp

Dinka 54.1
Natufian 40.0
Yemenite_Jew 5.9

Despite being a poor fit, this was initially troubling because it might've meant the later admixture in Somalis was somehow bogus but the Yemenite-Jew weirdly disappeared after I added South-Africa-1300ybp:

distance%=0.8087 / distance=0.008087

Tanzania-Luxmanda-3000ybp

Dinka 44.70
Natufian 42.45
South_Africa_1300BP 12.50
Yemenite_Jew 0.35

(or more like went down to a very negligible level)

Which makes sense as I've run this model before and the above result was what I got. Now I have ZERO inkling as to why South-Africa-1300ybp would cause such a change (they shouldn't really have any elevated Eurasian affinity?) but I wanted to know if maybe the reason she was showing Yemenite-Jew earlier was a sign of Mota-related affinities:

distance%=0.5303 / distance=0.005303

Tanzania-Luxmanda-3000ybp

Natufian 41.0
Mota 33.5
Dinka 25.5
Yemenite_Jew 0.0

And sure enough, and as you can see above, Mota being entered totally eliminated this "Yemenite-Jew" affinity which, to me, implied the Luxmanda sample was just showing some broadly Eurasian affinity like what Mota often shows and that does roughly seem to be the case:


distance%=1.4766 / distance=0.014766

Tanzania-Luxmanda-3000ybp

Dinka 53.65
Natufian 42.60
Iran_ChL 3.75

distance%=1.3864 / distance=0.013864

Tanzania-Luxmanda-3000ybp

Dinka 51.9
Natufian 42.9
Onge 5.3

distance%=1.5257 / distance=0.015257

Tanzania-Luxmanda-3000ybp

Dinka 54.5
Natufian 45.0
WHG 0.5


distance%=1.459 / distance=0.01459"

Tanzania-Luxmanda-3000ybp

Dinka 52.65
Natufian 44.40
MA1 2.95

distance%=1.5155 / distance=0.015155

Tanzania-Luxmanda-3000ybp

Dinka 54.2
Natufian 44.2
Han 1.5


distance%=1.4154 / distance=0.014154

Tanzania-Luxmanda-3000ybp

Dinka 51.95
Natufian 42.70
Tianyuan 5.35

distance%=0.7339 / distance=0.007339

Tanzania-Luxmanda-3000ybp

Dinka 43.50
Natufian 41.35
South_Africa_1300BP 11.90
Tianyuan 3.25

distance%=1.3864 / distance=0.013864"

Tanzania-Luxmanda-3000ybp

Dinka 51.85
Natufian 42.75
Onge 5.25
Iran_ChL 0.15

distance%=0.6756 / distance=0.006756

Tanzania-Luxmanda-3000ybp

Dinka 43.2
Natufian 41.2
South_Africa_1300BP 11.8
Onge 3.7

I say roughly because, as you can see above, this affinity seems more ENA than Eurasian as a whole and matches best with Onges and Tianyuan (and in that order). I have no bloody clue what this affinity is (some East-African affinity to Eurasians?) but I did wanna know if the same sort of stuff would happen to Somalis (Yemenite-Jew ancestry being usurped by Mota and South-Africa-13000ybp) and that doesn't seem to be the case:


distance%=0.5468 / distance=0.005468

Somali

Dinka 55.4
Natufian 34.5
Yemenite_Jew 10.1
South_Africa_1300BP 0.0

distance%=0.5287 / distance=0.005287

Somali

Dinka 55.20
Natufian 40.25
Iran_ChL 4.55
South_Africa_1300BP 0.00

distance%=0.5402 / distance=0.005402

Somali

Dinka 53.65
Natufian 34.65
Yemenite_Jew 9.65
Mota 2.05

distance%=0.5261 / distance=0.005261

Somali

Dinka 54.1
Natufian 40.2
Iran_ChL 4.4
Mota 1.3

distance%=0.5453 / distance=0.005453

Somali

Dinka 54.35
Natufian 34.40
Yemenite_Jew 9.80
Ari-Blacksmith 1.45

distance%=0.5287 / distance=0.005287

Somali

Dinka 54.95
Natufian 40.20
Iran_ChL 4.50
Ari-Blacksmith 0.35

Whether Mota, South-Africa-1300ybp or Ari-Blacksmiths were entered alongside Yemenite-Jews and Iran_Chl; this ancestry never disappeared in Somalis, it might've ended up being a little depleted for whatever reason but it never went away, unlike with the South-Erythraean which, to me, implies it's genuine admixture in Somalis.

There's something a little odd amok, though. The ENA affinity pretty much remains even if you include Mota or Ari-Blacksmiths in the runs and that goes for both Somalis and the South-Erythraean:


distance%=0.5025 / distance=0.005025

Somali

Dinka 54.15
Natufian 35.85
Yemenite_Jew 7.55
Onge 1.95
Mota 0.50

distance%=0.5294 / distance=0.005294"

Somali

Dinka 53.90
Natufian 35.05
Yemenite_Jew 8.70
Mota 1.20
Tianyuan 1.15

distance%=0.5029 / distance=0.005029

Somali

Dinka 54.5
Natufian 35.9
Yemenite_Jew 7.6
Onge 2.0
Ari-Blacksmith 0.0

distance%=0.5313 / distance=0.005313

Somali

Dinka 54.60
Natufian 35.05
Yemenite_Jew 8.70
Tianyuan 1.30
Ari-Blacksmith 0.35

distance%=0.5138 / distance=0.005138"

Tanzania-Luxmanda-3000ybp

Natufian 40.65
Mota 32.30
Dinka 25.90
Onge 1.15
Yemenite_Jew 0.00

distance%=0.5287 / distance=0.005287"

Tanzania-Luxmanda-3000ybp

Natufian 40.9
Mota 33.1
Dinka 25.6
Tianyuan 0.4
Yemenite_Jew 0.0

distance%=0.6756 / distance=0.006756

Tanzania-Luxmanda-3000ybp

Dinka 43.2
Natufian 41.2
South_Africa_1300BP 11.8
Onge 3.7
Yemenite_Jew 0.0

distance%=0.7339 / distance=0.007339

Tanzania-Luxmanda-3000ybp

Dinka 43.50
Natufian 41.35
South_Africa_1300BP 11.90
Tianyuan 3.25
Yemenite_Jew 0.00

However, when you remove the Onge or Tianyuan; Mota-related ancestry increases. This actually reminds me a bit of how "Dinka + Mota + Natufian + Iran_Chl" actually fits better for Somalis than "Dinka + Mota + Natufian + Yemenite Jew" and how, when given a chance, the simulation totally favors Iran_Chl over Yemenite Jews:

distance%=0.5402 / distance=0.005402

Somali

Dinka 53.65
Natufian 34.65
Yemenite_Jew 9.65
Mota 2.05

distance%=0.5261 / distance=0.005261

Somali

Dinka 54.1
Natufian 40.2
Iran_ChL 4.4
Mota 1.3

distance%=0.5261 / distance=0.005261

Somali

Dinka 54.1
Natufian 40.2
Iran_ChL 4.4
Mota 1.3
Yemenite_Jew 0.0

The fit isn't improved by THAT much but it is improved. But of course, it's highly unlikely that Chalcolithic-Zagrosians just hobbled on over to the Horn and intermixed with the populations there so this is very a much a case where a better fitting model is not necessarily the more sensible model as something else most likely brought the Zagros-Chalcolithic related ancestry. And I think the the ENA affinity is the same, it may actually usurp and/or deplete Mota-related ancestry and produce better fits in models but, of course, actual "Onge" or "Tianyuan" admixture in Somalis, Tigrinyas, Agaws and the like makes little sense and it's probably owed to the broadly Eurasian (more ENA) affinities Mota is showing below:

distance%=1.1196 / distance=0.011196

Mota

Dinka 53.00
South_Africa_1300BP 37.45
Onge 7.80
Natufian 1.75
South_Africa_2000BP 0.00
MA1 0.00
Tianyuan 0.00
WHG 0.00


(it's important to note that I've only seen this affinity appear in PCAs and perhaps also ADMIXTURE (differently between the two, if I recall correctly) and maybe a bit in tree-mix given how Mota sits closer on the tree to Eurasians than other Africans do but I've never seen this affinity in straight d-stats, for example)

It's all still a bit confusing and interesting but I'd say this is probably the most sensible model for Somalis for now:

distance%=0.5402 / distance=0.005402

Somali

Dinka 53.65
Natufian 34.65
Yemenite_Jew 9.65
Mota 2.05

And for those wondering, other Horn-Africans like Tigrinyas and Amharas do show the same ENA affinity as Somalis and show this even when Mota is placed in the run but they arguably have too much Mota-related ancestry in comparison to Somalis for the total usurping effect to take place:

distance%=0.9216 / distance=0.009216

Tigray-Tigrinya

Natufian 34.85
Dinka 34.70
Yemenite_Jew 19.80
Mota 9.80
Onge 0.85

-

distance%=0.5734 / distance=0.005734

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 40.15
Natufian 33.85
Yemenite_Jew 20.35
Mota 3.60
Onge 2.05

But, I wonder if this implies the South-Erythraean has some actual Mota-related ancestry or if the South-Africa-1300ybp samples somehow also cause an ENA affinity because lets not forget how the Yemenite-Jewish ancestry in her mostly disappeared when South-Africa-1300ybp was used in the model:

distance%=0.8087 / distance=0.008087

Tanzania-Luxmanda-3000ybp

Dinka 44.70
Natufian 42.45
South_Africa_1300BP 12.50
Yemenite_Jew 0.35

However, to be fair, it doesn't entirely make the ENA affinity go away:

distance%=0.6756 / distance=0.006756

Tanzania-Luxmanda-3000ybp

Dinka 43.2
Natufian 41.2
South_Africa_1300BP 11.8
Onge 3.7

However, to be fair once again, it didn't actually make the Yemenite-Jew affinity go away entirely either (just drastically reduced it down to a totally negligible level). So idk, this woman's ancestors might have actually been part Mota-related:

distance%=0.5138 / distance=0.005138

Tanzania-Luxmanda-3000ybp

Natufian 40.65
Mota 32.30
Dinka 25.90
Onge 1.15
South_Africa_1300BP 0.00

-

distance%=0.5303 / distance=0.005303

Tanzania-Luxmanda-3000ybp

Natufian 41.0
Mota 33.5
Dinka 25.5

This looks like too much Mota-related ancestry, though. I wonder if she actually has admixture more related to the two Malawi samples, for example:

distance%=0.8089 / distance=0.008089

Tanzania-Luxmanda-3000ybp

Dinka 44.70
Natufian 42.80
South_Africa_1300BP 12.25
Malawi_Fingira_2500BP 0.25
Malawi_Hora_Holocene 0.00

No give, apparently... Yeah, I'll keep modeling another time and see if I can get back to this but, for the time being, I think the gist is that the later West-Asian ancestry in Somalis seems pretty real and the South-Erythraean might be part Mota-related rather than a straight "Dinka-like + Southern African + Natufian" mixture.

NetNomad
11-06-2017, 08:26 AM
So what kind of people lived in Somalia around Mota's time? It seems like Somalis are replacement population from East Sudan rather than ancient to Somalia. This might also explain the high E-M78 and lack of E-M329.

Awale
11-06-2017, 09:44 AM
So what kind of people lived in Somalia around Mota's time? It seems like Somalis are replacement population from East Sudan rather than ancient to Somalia. This might also explain the high E-M78 and lack of E-M329.

Well, Somalis, as an ethno-linguistic group, are a relatively recent development (last ~2,000-2,500 or so years) but are probably quite descended from even some of the earliest inhabitants of the Peninsula and the Ogaden area and it's the presence of these likely Erythraeic speaking pastoralists in the overall region that's important and that seems to stretch back to about ~3000 BCE (see here (https://link.springer.com/article/10.1007/s10437-015-9184-9)) in the Somali Peninsula (particularly it's north-central half) and probably just somewhat earlier than that in Ethiopia and Eritrea as a whole. In the highlands, it seems like people like Mota were probably the earlier population before these pastoralists (and probably Omotic speakers before them), whose origin is most likely in Sudan (as part of the Sudanese Neolithic), started showing up. But I can't really say in the case of the Somali Peninsula...

In the southern parts of the Peninsula, there's apparently evidence of "Khoisan"-like peoples who were Hunter-Gatherers being the earlier, indigenous people until they were displaced and/or assimilated by Erythraeic speaking agro-pastoralists (some supposedly remained until relatively recently when they were seemingly assimilated like the Eyle (https://en.wikipedia.org/wiki/Buur_Heybe#Demographics)) but in the north; I don't know. We'd need more archaeological study, I think. I haven't seen much on what the place was like before Erythraean (probably genetically "Luxmanda minus the South-Africa-1300ybp-like ancestry") pastoralists showed up. But I highly doubt people like Mota were common as Somalis show too little of such ancestry. I think Mota-like people were more strictly present along the Ethiopian Highlands.

Lank
11-06-2017, 04:46 PM
The South_Africa_1300BP sample (who really dates to 1200 BP; 1300 is the uncalibrated date) has significant ancestry from the pastoralists who reached South Africa around 2000 years ago. She's modeled in the paper as deriving 40% of her ancestry from a group related to the Tanzania Luxmanda 3100 BP sample, and also shows significant Cushitic admixture in the ADMIXTURE run. I wouldn't use her in these models for this reason (other than modeling later southern Africans), but rather someone like the South_Africa_2000BP sample, who looks like a relatively unadmixed Khoisan. Of course, ideal in the case of the 3100 BP Tanzanian would be to use a representative East African HG sample, rather than distantly related southern Africans. How about the Tanzanian from Pemba 1400 BP? She's a bit recent and may have very minor Cushitic admixture, so she's not entirely ideal, but would be interesting to test, perhaps also alongside the South African from 2000 BP. The Tanzania 1400 BP sample carries mtDNA L0a4, which is found at a very high frequency in some modern East African hunter-gatherers.

The recent West Asian signal in Somalis is a very interesting finding, and there are several plausible scenarios to explain its origin. Some ancient DNA would be great, but it makes sense that they may have received recent admixture compared to the early Tanzanian South Cushite, who also predates the spread of camel pastoralism in Africa.

Awale
11-06-2017, 05:35 PM
The South_Africa_1300BP sample (who really dates to 1200 BP; 1300 is the uncalibrated date) has significant ancestry from the pastoralists who reached South Africa around 2000 years ago. She's modeled in the paper as deriving 40% of her ancestry from a group related to the Tanzania Luxmanda 3100 BP sample, and also shows significant Cushitic admixture in the ADMIXTURE run. I wouldn't use her in these models for this reason (other than modeling later southern Africans), but rather someone like the South_Africa_2000BP sample, who looks like a relatively unadmixed Khoisan.

Shit, I actually didn't catch this. I'll stick to 2000ybp from now on, then. Thank you for pointing this out, it makes a lot more sense that the Yemenite-Jew-like ancestry disappears when that sample is used now.

Here's Mota remodeled using the 2000ybp sample for now:

distance%=1.1744 / distance=0.011744

Mota

Dinka 58.20
South_Africa_2000BP 24.95
Onge 9.45
Natufian 7.40

The Eurasian signal is looking more broad now too which is great because this is how I recall it looking in ADMIXTURE. I'll re-run every run where I used the 1300ybp sample with the 2000ybp sample and share the results soon. And the study seems quite right about the 1300ybp sample:

distance%=0.6989 / distance=0.006989

South_Africa_1300BP

South_Africa_2000BP 63.7
Tanzania_Luxmanda_3000BP 36.3

I see this individual was a pastoralist too, makes good sense.

Awale
11-07-2017, 12:43 PM
Well, it looks like the broadly Eurasian affinity is actually a thing among all the Mota-like individuals (along the East-South cline):

distance%=0.7567 / distance=0.007567

Malawi-Fingira-2500ybp

South-Africa-2000ybp 71.30
Dinka 22.60
Onge 4.15
Natufian 1.95

distance%=0.7851 / distance=0.007851

Malawi-Hora-Holocene

South-Africa-2000ybp 68.6
Dinka 18.8
Onge 7.0
Natufian 5.6

So the idea that the South-Erythraean might have specifically Mota-like ancestry from the Horn itself is shaky now cos she could've gotten that odd ENA affinity from populations south of the Horn too.

--

And, oddly, South-Africa-2000ybp still causes the Yemenite-Jew in the South-Erythraean to become negligible:

distance%=0.8921 / distance=0.008921

Tanzania-Luxmanda-3000ybp

Dinka 47.0
Natufian 44.1
South-Africa-2000ybp 8.0
Yemenite-Jew 0.9

Odd that this sample would have some sort of Eurasian affinity... But the ENA affinity doesn't go away with either Mota or the Southern African HG being used (it does lessen, though):

distance%=1.3864 / distance=0.013864

Tanzania-Luxmanda-3000ybp

Dinka 51.9
Natufian 42.9
Onge 5.3

distance%=0.7307 / distance=0.007307

Tanzania-Luxmanda-3000ybp

Dinka 45.05
Natufian 43.00
South-Africa-2000ybp 7.70
Onge 4.25


distance%=0.5138 / distance=0.005138

Tanzania-Luxmanda-3000ybp

Natufian 40.65
Mota 32.30
Dinka 25.90
Onge 1.15
South-Africa-2000yp 0.00
Yemenite_Jew 0.00

But the Yemenite-Jew/Iran_Chl in Somalis and Aris seems totally different from this and doesn't go away if you introduce ENAs (Onge & Tianyuan), Mota or South-Africa-2000ybp:


distance%=0.5025 / distance=0.005025

Somali

Dinka 54.15
Natufian 35.85
Yemenite_Jew 7.55
Onge 1.95
Mota 0.50
Tianyuan 0.00
South-Africa-2000ybp 0.00

-


distance%=0.2828 / distance=0.002828

Ari-Blacksmith

Mota 37.60
Dinka 32.60
Natufian 14.80
South-Africa-2000ybp 7.35
Yemenite_Jew 6.80
Onge 0.85
Tianyuan 0.00

--

distance%=0.3923 / distance=0.003923

Ari-Cultivator

Mota 55.8
Dinka 22.7
Natufian 13.0
Yemenite_Jew 7.1
South-Africa-2000ybp 1.4
Onge 0.0
Tianyuan 0.0

--

distance%=0.5734 / distance=0.005734

Xamtanga-Agaw ("Ethiopian-Afar")

Dinka 40.15
Natufian 33.85
Yemenite_Jew 20.35
Mota 3.60
Onge 2.05

--

distance%=0.9216 / distance=0.009216

Tigray-Tigrinya

Natufian 34.85
Dinka 34.70
Yemenite_Jew 19.80
Mota 9.80
Onge 0.85


So it still looks real.


How about the Tanzanian from Pemba 1400 BP? She's a bit recent and may have very minor Cushitic admixture, so she's not entirely ideal, but would be interesting to test, perhaps also alongside the South African from 2000 BP. The Tanzania 1400 BP sample carries mtDNA L0a4, which is found at a very high frequency in some modern East African hunter-gatherers.

Here you go, 7awey:

distance%=0.8197 / distance=0.008197

Tanzania-Luxmanda-3000ybp

Natufian 41.8
Dinka 38.5
Tanzania-Pemba-1400ybp 19.8
Yemenite-Jew 0.0

distance%=0.8196 / distance=0.008196

Tanzania-Luxmanda-3000ybp

Natufian 41.85
Dinka 38.85
Tanzania-Pemba-1400ybp 18.95
South-Africa-2000ybp 0.35
Yemenite-Jew 0.00

I'll leave it at that for now. Gonna be a bit busy over the next few days but I'll try and do some stuff with removing some of the PCA dimensions and see how that changes things up.

Chad Rohlfsen
11-07-2017, 10:14 PM
Awale, do you also have Levant_N and Anatolia for your runs? They might help your fits with regards to Eurasian ancestry in Horners. Also, the broadly Eurasian affinity of Mota shouldn't be a big concern. Considering the introgression of E and the dates of those branches, it may be older than Ust_Ishim, in regards to the type of Eurasian ancestry. There's even a chance it is Basal to all, including Basal Eurasian, so it can't choose between Natufian, WHG, and Onge. One thing about using WHG is that they are so inbred, they are going to be hard to fit in anywhere that doesn't have a sizeable contribution from them. I've found some wild stuff with qpGraph and will be going deeper into it on my African thread in the aDNA section. Also, the 2000BP samples we have might already be pastoralist admixed. There are samples up to 9000 years old out there from South Africa. Let's hope someone tests them.

Michalis Moriopoulos
11-09-2017, 04:07 AM
Hey, guys, it's Mike from the old days. Yeah, I'm back... http://www.anthrogenica.com/showthread.php?12616-Behold-the-glorious-return-of-wait-who&p=308087#post308087

Just want to say that this is a very interesting study, but I'm having quite a time making heads or tails of some of it.

Are the Hadza supposed to be direct descendants of Mota? Descendants of a closely-related side-branch? They certainly share a lot of alleles with Mota, but I'm curious if their exact relationship with the sample has been resolved to anyone's satisfaction...

Awale
11-20-2017, 11:31 PM
Awale, do you also have Levant_N and Anatolia for your runs?

They're in the Global-10 PCA but no, I don't use them for my runs. They tend to be totally edged out once Natufians are present:

distance%=0.5928 / distance=0.005928

Somali

Dinka 54.05
Natufian 36.45
Levant_BA 7.35
Mota 2.15
Levant_N 0.00
Barcin_N 0.00

And using Levant_N instead of Natufians makes the fits a lot worse (would make the above one jump up to about 1.3 or so) so I just avoid throwing them in and use Natufians instead.


They might help your fits with regards to Eurasian ancestry in Horners. Also, the broadly Eurasian affinity of Mota shouldn't be a big concern. Considering the introgression of E and the dates of those branches, it may be older than Ust_Ishim, in regards to the type of Eurasian ancestry. There's even a chance it is Basal to all, including Basal Eurasian, so it can't choose between Natufian, WHG, and Onge. One thing about using WHG is that they are so inbred, they are going to be hard to fit in anywhere that doesn't have a sizeable contribution from them.

Hmm, I personally think this Eurasian affinity has something more to do with sharing descent from pre-historic East-Africans with Eurasians, but who knows. It doesn't shock me too much either. Interesting point about WHGs, though. I'll be waiting on those qpGraph runs of yours. I wanna see some good qpGraphs for Africans big time. :P


Are the Hadza supposed to be direct descendants of Mota? Descendants of a closely-related side-branch? They certainly share a lot of alleles with Mota, but I'm curious if their exact relationship with the sample has been resolved to anyone's satisfaction...

Welcome back, co-founder. Arash-meng has told me much about you. Anyway, Hadzas are not descendants of Mota. Probably descended from pre-historics in Southeast Africa who were rather similar to him on a basal level like the Malawi samples (existing along the "East-South cline"), though. And they definitely seem to have some Horn-African admixture as well but more along the Erythraeaic (Cushitic) speaking pastoralist variety than the Mota-like variety.

Michalis Moriopoulos
11-22-2017, 03:50 AM
Thanks for the explanation, Awale. I've seen a few African populations modeled in qpGraph, but only in very specific contexts (i.e., pastoralist geneflow into South Africa). It would be nice to see a more inclusive project where all the major players were included. That's why I'm excited for Chad's project.

We need all the ancient SSA DNA we can get. One wonders if the most famous fossils from the region are just too damaged to test. I'm talking about Hofmeyr, Jebel Sahaba, Iwo Eleru, and Asselar. If I worked for an academic genomics lab, those would be among the first fossils I'd want to check out.

blackflash16
11-22-2017, 02:15 PM
We need all the ancient SSA DNA we can get. One wonders if the most famous fossils from the region are just too damaged to test. I'm talking about Hofmeyr, Jebel Sahaba, Iwo Eleru, and Asselar. If I worked for an academic genomics lab, those would be among the first fossils I'd want to check out.

In the case of Jebel-Sahaba, the Nataruk remains (http://www.anthrogenica.com/showthread.php?6271-c-Inter-group-violence-among-early-Holocene-hunter-gatherers-of-West-Turkana-Kenya) samples might be a good stand-in for Mesolithic Nubia/Jebel-Sahaba site considering the presence of pottery at the massacre site as well as previous finds of wavy-line pottery at Lake Turkana.