PDA

View Full Version : Autosomal DNA data from Sudan, Dobon et al. 2015



Awale
06-24-2015, 08:18 PM
Not many African members here but for anyone who's interested-> there are various new Sudanese samples from a new study:



East Africa is a strategic region to study human genetic diversity due to the presence of ethnically, linguistically, and geographically diverse populations. Here, we provide new insight into the genetic history of populations living in the Sudanese region of East Africa by analysing nine ethnic groups belonging to three African linguistic families: Niger-Kordofanian, Nilo-Saharan and Afro-Asiatic. A total of 500 individuals were genotyped for 200,000 single-nucleotide polymorphisms. Principal component analysis, clustering analysis using ADMIXTURE, FST statistics, and the three-population test were used to investigate the underlying genetic structure and ancestry of the different ethno-linguistic groups. Our analyses revealed a genetic component for Sudanese Nilo-Saharan speaking groups (Darfurians and part of Nuba populations) related to Nilotes of South Sudan, but not to other Sudanese populations or other sub-Saharan populations. Populations inhabiting the North of the region showed close genetic affinities with North Africa, with a component that could be remnant of North Africans before the migrations of Arabs from Arabia. In addition, we found very low genetic distances between populations in genes important for anti-malarial and anti-bacterial host defence, suggesting similar selective pressures on these genes and stressing the importance of considering functional pathways to understand the evolutionary history of populations.

Link (http://www.nature.com/srep/2015/150528/srep09996/full/srep09996.html)

I asked David/ Polako/ the author of Eurogenes (http://eurogenes.blogspot.ae/)/ Generalisimo via email to acquire the samples and run them against other East Africans and some West Asian populations (in PCAs) so we could see how they compare and well, I'll just copy and past what I ended up getting and posted on ABF (http://www.forumbiodiversity.com/showthread.php/44363-Sudan-autosomal-study/page5?highlight=Sudan+autosomal):


Well, ladies and gentlemen! Polako made two PCAs for us as well as a small dataset available in case one wants to run some of these new samples through an ADMIXTURE calculator.


The chip they used is really crap; it doesn't overlap well with the usual chips. But the attached PCA look OK, as long as none of the samples are mislabeled.

Btw, here you can download a small dataset with these new samples that you can run with ADMIXTURE.

Here is the Global PCA and it shows you each group's admixture levels rather well based on how much they pull toward West Asians and North Africans:

https://drive.google.com/file/d/0B2ARnUeK-Y8WSjNlNUZlYUdhWU0/view?usp=sharing

I apologize though for how messy it can get given how Polako labels each sample with its ethnic designation.

This is a sort of Pan Northeast Africa PCA with Levantines, Arabians & Maasai thrown in:

https://drive.google.com/file/d/0B2ARnUeK-Y8WS2JfNGYzelpZODQ/view?usp=sharing

And finally, here is the dataset Davidski-man/ the King of the Poles gave us to use for an ADMIXTURE analysis:

https://drive.google.com/file/d/0B9o3EYTdM8lQb0plX1AwYS1iV2M/view

---

A Katrilion Bajillion thanks to David/ Polako!




Interesting things to note, some of which I mention at my blog (http://anthromadness.blogspot.ae/2015/06/sudanese-arabs-beni-ameri-beja-and.html):


The Beni-Amer (https://en.wikipedia.org/wiki/Beni-Amer_people) Bejas (https://en.wikipedia.org/wiki/Beja_people) seem to perhaps share more ancestry with Horners (https://en.wikipedia.org/wiki/Horn_of_Africa) such as Habeshas (https://en.wikipedia.org/?title=Habesha_people) and form a cluster with Horners at large (https://drive.google.com/file/d/0B2ARnUeK-Y8WS2JfNGYzelpZODQ/view) while Sudanese Arabs (https://en.wikipedia.org/?title=Sudanese_Arabs) and Nubians (https://en.wikipedia.org/wiki/Nubians)form their own, however they (the Bejas) do sort of look to be intermediates between Horners and North Sudanese (Sudanese Arabs + Nubians). They also look to mostly be more West Eurasian than Habeshas though, putting them at >50% range of West Eurasian/ West Asian ancestry.

---

The Nagada Copts (https://en.wikipedia.org/wiki/Copts) are basically the descendants of immigrants to Sudan within the last 2 centuries and are more West Eurasian than your average Muslim Egyptian; they also seem to lack ANE & WHG input. Granted, they show trace levels of WHG but that as David/ Generalismo notes after running them through the K8 ADMIXTURE analysis; is probably not real WHG but something WHG-like from West Asia ("UHG"). At any rate; the study itself confirms that they mostly lack Niger-Congo input (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png) (the "YRI"/ Yoruba (https://en.wikipedia.org/wiki/Yoruba_people) component at K=2) unlike their Muslim Egyptian counterparts who do have Niger-Congo admixture. These Copts basically only have East African (http://anthromadness.blogspot.ae/2015/02/the-east-african-cluster.html) admixture so they're more or less about ~85% "ENF (http://eurogenes.blogspot.com.au/2014/12/the-fateful-triangle.html)" and ~15% East African much like a Copt whose results I touch upon at my blog (http://anthromadness.blogspot.ae/2015/06/copts-example-of-pre-islamic-and-arab.html) whilst wondering if these Copts can be taken as a good example of Pre-Islamic and Pre-Arab Conquest Egypt-> What do you think of them?

---

David ran some of these Sudanese samples through Eurogenes K=8 & Sudanese Arabs, Nubians & Beni-Amer Bejas seem to be bereft of ANE & WHG ancestry much like Horners such as Somalis & Habeshas (http://www.forumbiodiversity.com/showthread.php/43501-Lack-of-ANE-in-the-Horn?highlight=). They like Horners seem to be mostly just "ENF" admixed. But some of the samples from what he's said do show trace amounts of WHG (Horners don't show this) but that again probably isn't real WHG.

---

There seems to be a good amount of heterogeneity amongst Sudanese Arabs and Nubians in terms of West Eurasian ancestry levels. Some few are about as West Eurasian as Somalis (~40%), some seem to sit between Somalis & Habeshas, others are at Habesha levels whilst many such as Shukrya Arabs (https://en.wikipedia.org/wiki/Shukria_clan) seem more admixed than Habeshas (>50% West Eurasian)

---

The Fulani (https://en.wikipedia.org/wiki/Fula_people) samples seem to be of two separate groups, one that seems practically as African as the Gumuz (https://en.wikipedia.org/wiki/Gumuz_people) and one that looks more blatantly West Eurasian/ Northwest African admixed and pulls toward North Africans and West Asians on a PCA (https://drive.google.com/file/d/0B2ARnUeK-Y8WS2JfNGYzelpZODQ/view) however when Northwest Africans are thrown into this PCA (https://drive.google.com/file/d/0B2ARnUeK-Y8WWVNEdGpSU2VsUEk/view); both the less admixed Fulanis and more admixed Fulanis pull downward on the Y axis as Northwest Africans do; displaying a clear affinity for them from what I can tell. This is in line with how Fulanis tend to show "Maghrebi/ Northwest African" admixture in other studies like Hodgson et al. 2014 (http://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1004393): [Hogson et al. 2014 ADMIXTURE run (http://s3-eu-west-1.amazonaws.com/previews.figshare.com/1533296/preview_1533296.jpg)] [names of the various components (http://www.plosgenetics.org/article/fetchObject.action?uri=info:doi/10.1371/journal.pgen.1004393.g002&representation=PNG_L)]. David ran them through Eurogenes K=8 and it seems that like Northwest Africans/ Maghrebis (https://docs.google.com/spreadsheets/d/1x8pm8sVcHqceiNFJMO082kxaBF5ePr4__bAK05VQRFw/edit#gid=2145681294) the more West Eurasian admixed Fulani samples are showing WHG ancestry (~5-6%). So it looks to me that Fulanis (at least the more clearly Northwest African admixed ones) do carry some WHG like Northwest Africans do.



Some "mappings" I made of the various groups that form clusters together in the PCAs:

http://3.bp.blogspot.com/-kQtBYwNbU1I/VXheE4Eod1I/AAAAAAAACtM/p7ifqc_PHe8/s1600/28mk2sy.jpg.png

.

http://oi58.tinypic.com/90bx9k.jpg


Detailed info on the samples I acquired from the paper's authors via email:

https://docs.google.com/spreadsheets/d/1XfldiJOG4T4rKIX20Twku89hYMYiGilUSD47uhT41Fo/edit#gid=1847530189

For those of you capable of tinkering with them; David put together a small dataset of samples from this new study someone could run through an ADMIXTURE analysis if they cared to:

https://drive.google.com/file/d/0B9o3EYTdM8lQb0plX1AwYS1iV2M/view

--

Many thanks to David btw! And while I don't see too many members interested in African genetics here; these Copts are certainly interesting so... Discuss...

Ezana
06-24-2015, 08:53 PM
Great work, Awale! Awesome first post!

Awale
06-24-2015, 09:30 PM
Great work, Awale! Awesome first post!

Heh, thanks. Btw, Lol_Race (he's told me a bit about you so I'm sure you're familiar with him :-)) and I noticed something a bit OT bit still interesting about the "Northeast Africa" plot:

https://drive.google.com/file/d/0B2ARnUeK-Y8WS2JfNGYzelpZODQ/view

It's that Somalis alongside Northern Highlanders (Habeshas + Beta Israels) & Oromos are seemingly pulling toward Aris and showing what looks like an Omotic (http://anthromadness.blogspot.ae/2015/01/the-omotic-cluster-horn-specific-cluster.html) affinity. Lol_Race asked David to throw in Hadzas to perhaps introduce something similar to Aris that may have been able to slightly differentiate Somalis from Omotic admixed Horners but the pulling toward Aris seemed to stick:

https://drive.google.com/file/d/0B2ARnUeK-Y8WWVRIYXdPeGRuOWs/view

Why is this weird? Well... It's making it seem like Somalis may actually have Omotic admixture. As you may know; Bandar from the old ABF days created a run (https://docs.google.com/spreadsheets/d/1HRFbBy3WiFGq7AX60E1LZqqYrGfizObQgkjolfZ6Qa8/edit#gid=5) that was pretty good at spotting Omotic by removing the most inbred Ari Blacksmiths & in this run Somalis generally came up 0% Omotic except for Ethiopians Somalis who had some folks among them (alongside two heavily Omotic admixed outliers) showing Omotic admixture. Now, David suggested that the lower number of markers he was forced to use for these PCAs (due to the rather poor overlap between Dobon et al.'s chip and the chips used by other studies) might just change this up a bit.

So I recently asked him to try and replicate those two above PCAs but exclude the Sudanese samples entirely so he could then use the usual high number of markers-> if things stay the same if and when he turns up with those PCAs and Somalis keep pulling toward Aris alongside other Horners-> Do you think it could mean Bandar's run was wrong and Somalis perhaps do have at least some Omotic admixture?

I wonder if these Bejas also have Omotic admixture; I asked Ethiohelix (http://ethiohelix.blogspot.ae/2015/06/more-ethiopian-uniparental-data-more.html) to run them through an ADMIXTURE analysis and while he said he was pretty rusty; he does know how to spot Omotic admixture IIRC so we'll see about these Bejas, I guess... If they lack Omotic admixture then it does lend some good credence (as Lol_Race mentioned on ABF) that Cushitic speakers originally came from further up north in Northeast Africa. Granted, that idea makes enough good sense to me from a linguistic standpoint alone..

tamilgangster
06-25-2015, 08:33 AM
1) On the graphs whats the different between dark blue and green, they both appear to be west eurasian related. Is the green a berber specific component?
2) Also take into consideration noth african populations in algeria and morroco have WHG around 15%, so its very possible that the WHG found in copts is from that source and not UHG/EHG related

Awale
06-25-2015, 01:10 PM
1) On the graphs whats the different between dark blue and green, they both appear to be west eurasian related. Is the green a berber specific component?

You mean this graph (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png)? The dark blue one just looks like a Southwest Asian-esque component peaking in Qataris in this dataset (mostly ENF with some East African, I'd wager).

The dark green component definitely doesn't seem to have much to do with Northwest Africans; otherwise it would have been perhaps much more profuse in the Fulani samples. Seems like these Copts are just forming their own component...


2) Also take into consideration noth african populations in algeria and morroco have WHG around 15%, so its very possible that the WHG found in copts is from that source and not UHG/EHG related

Hmm, perhaps. But keep in mind that what WHG they do have is very minimal (1-3%). One thing I would notice though is the PCA David made with Northwest Africans present:

https://drive.google.com/file/d/0B2ARnUeK-Y8WWVNEdGpSU2VsUEk/view

Notice how all Northwest African/ Maghrebi admixed groups (Morrocan Jews who have been known to have some Northwest African admixture [-] (http://www.readcube.com/articles/10.1038%2Fnature09103) & Fulanis) pull down on the Y axis with Northwest Africans-> displaying what looks to be their affinity for them. And notice that these Copts don't do the same... I can't totally deny the possibility but I doubt they have meaningful Northwest African admixture or got their trace amounts of WHG from them. If they share any ancestry with Northwest Africans; my money would be that this would mostly be much more ancient shared ancestry-> owed perhaps to the original Berber speakers who migrated into Northwest Africa who probably weren't WHG admixed at all.

DMXX
06-25-2015, 04:14 PM
Excellent work, Awale. A worthwhile read.

My main impression is that the PCA plots tend to mirror geographic clines in a broad sense. Additionally, I'm taken aback (in a positive way) regarding the clustering patterns here; the summation of North Africa, the Levant and the Arabian Peninsula occupy a proportionally smaller clustering space than the neighbouring non-North African groups. I'm immediately reminded of the coined phrase regarding Eurasia being a "subset of African genetic diversity".

I'd be curious to see what the ANE frequencies are among these folks. It does appear, as you've covered in your earlier blog entries and stated definitively here, that almost all of the West Eurasian-derived scores among NE Africans is ENF. Given the presence of R1b-V88 in Chadic speakers and the current popular working hypothesis regarding the Y-DNA R1-M173 bifurcation in the Eurasian steppe, I do wonder whether even a trace of this is anywhere to be seen in NE Africa, as it lies more or less en route between Eurasia and Cameroon.

In that respect, I wonder if anyone with ADMIXTURE familiarity is interested in running these samples to see if anything comes up.

Awale
06-25-2015, 09:28 PM
My main impression is that the PCA plots tend to mirror geographic clines in a broad sense. Additionally, I'm taken aback (in a positive way) regarding the clustering patterns here; the summation of North Africa, the Levant and the Arabian Peninsula occupy a proportionally smaller clustering space than the neighbouring non-North African groups. I'm immediately reminded of the coined phrase regarding Eurasia being a "subset of African genetic diversity".

Well, a good amount of modern African diversity is also ultimately owed to Eurasian input, so I'd keep that in mind. One of the main things (if not in some cases the main thing) pulling Horners, North Sudanese, the Maasai and various Fulanis away from where groups like Nilotes are clustering is ultimately their West Eurasian ancestry.


It does appear, as you've covered in your earlier blog entries and stated definitively here, that almost all of the West Eurasian-derived scores among NE Africans is ENF.

Yes, this is often why Peninsular Arabians and now these Copts tend to seem like the best example of the kinds of West Asians the East African ancestors of Horners mixed with episodically; it's because these new Copt samples and Arabians tend to mostly just be ENF (mostly lacking ANE & WHG-UHG) when it comes to their West Eurasian ancestry.


Given the presence of R1b-V88 in Chadic speakers and the current popular working hypothesis regarding the Y-DNA R1-M173 bifurcation in the Eurasian steppe, I do wonder whether even a trace of this is anywhere to be seen in NE Africa, as it lies more or less en route between Eurasia and Cameroon.

Hmm, there are people better suited to helping you figure this out (Beyoku, Lol_Race, Pgbk87 et al. ; I'll message them about your musings ;))


In that respect, I wonder if anyone with ADMIXTURE familiarity is interested in running these samples to see if anything comes up.

Yeah, it would be very interesting to see what would turn up. Adding Somalis into the mix would be very interesting so we could see just how much ancestry would prove to look "shared" between Somalis & Bejas (fellow Cushitic speakers) given that at K=7 (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png) what looks to be a sort of pseudo-Cushitic component like "Ethio-Somali (http://anthromadness.blogspot.ae/2014/12/ethio-somali-is-farce.html)" forms (making up a good degree of the Maasai's ancestry and a whole lot of the Beja and Ethiopian samples' ancestry) but that ADMIXTURE run as Lol_Race noted to me at ABF is rather unstable:


That Cushitic component doesn't look particularly stable, probably due to the mere 14K SNPs used, as well as the lack of important samples, like Somalis. Other than the major Cushitic contribution, Maasai should also show significant Nilo-Saharan and Bantu, but that K=7 run isn't picking up on the NS at all. It makes them look predominantly Niger-Congo, which is misleading (should be closer to 20%).

The "Fulani" component is also problematic, and not informative for East African populations.

Chad Rohlfsen
06-25-2015, 09:48 PM
I can add these to my dataset and keep whatever passes the 50% threshold for coverage. I'll run them in admixture after a while. I'm plugging in some other stuff at the moment.

Awale
06-25-2015, 09:54 PM
I can add these to my dataset and keep whatever passes the 50% threshold for coverage. I'll run them in admixture after a while. I'm plugging in some other stuff at the moment.

That would be great but do you have any Somali & Habesha samples? Or at least the Horner samples from Pagani et al. (http://www.cell.com/ajhg/abstract/S0002-9297%2812%2900271-6); running them alongside the "Afar (http://anthromadness.blogspot.ae/2014/12/afar-samples-from-autosomal-dna-studies.html)", Somali and Habesha samples from that study is pretty key.

Chad Rohlfsen
06-25-2015, 10:10 PM
I have tons of Africans, including Somalis, Ethiopian Jews, etc. I've got like 2700 samples, of which 200 are ancient. I even have Indian tribals and the Onge.

Awale
06-25-2015, 10:15 PM
I have tons of Africans, including Somalis, Ethiopian Jews, etc. I've got like 2700 samples, of which 200 are ancient. I even have Indian tribals and the Onge.

Well, shit on a stick-> that should be good. Thanks for offering to do this.

Chad Rohlfsen
06-25-2015, 10:47 PM
I don't know if these samples have enough coverage. I'm only seeing 15k snp's. That's way too low. I don't think they'll fly with my SNP list.

Awale
06-25-2015, 10:52 PM
I don't know if these samples have enough coverage. I'm only seeing 15k snp's. That's way too low. I don't think they'll fly with my SNP list.

Yes, that's what they made their own ADMIXTURE runs with like this one (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png)... Their chip doesn't overlap well with the ones used in other studies. See what you can do though...

Chad Rohlfsen
06-25-2015, 11:05 PM
What I've got will give a better picture. 15k SNPs is kind of a waste. Results will likely be rather dubious. My Africans have about 340k SNPs.

Awale
06-25-2015, 11:08 PM
15k SNPs is kind of a waste. Results will likely be rather dubious.

Indeed but do it anyway... But if you want to do a run of your own with the East African samples you already have; go ahead, we have plenty of good runs on them but it'd still be interesting to see, I guess.

Chad Rohlfsen
06-25-2015, 11:19 PM
I'll see if it allows me to merge and run them after bit.

Chad Rohlfsen
06-26-2015, 02:30 AM
Okay. If all goes well, I've got about 1400 samples in the test. I'd say close to 1100 are African of some kind. I'll see if it allows me to run the test in just a bit.

tamilgangster
06-26-2015, 10:17 AM
You mean this graph (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png)? The dark blue one just looks like a Southwest Asian-esque component peaking in Qataris in this dataset (mostly ENF with some East African, I'd wager).

The dark green component definitely doesn't seem to have much to do with Northwest Africans; otherwise it would have been perhaps much more profuse in the Fulani samples. Seems like these Copts are just forming their own component...



Hmm, perhaps. But keep in mind that what WHG they do have is very minimal (1-3%). One thing I would notice though is the PCA David made with Northwest Africans present:

https://drive.google.com/file/d/0B2ARnUeK-Y8WWVNEdGpSU2VsUEk/view

Notice how all Northwest African/ Maghrebi admixed groups (Morrocan Jews who have been known to have some Northwest African admixture [-] (http://www.readcube.com/articles/10.1038%2Fnature09103) & Fulanis) pull down on the Y axis with Northwest Africans-> displaying what looks to be their affinity for them. And notice that these Copts don't do the same... I can't totally deny the possibility but I doubt they have meaningful Northwest African admixture or got their trace amounts of WHG from them. If they share any ancestry with Northwest Africans; my money would be that this would mostly be much more ancient shared ancestry-> owed perhaps to the original Berber speakers who migrated into Northwest Africa who probably weren't WHG admixed at all.

THe reason why copts form a seperate component could be due to inbreeding and genetic isolation, from other communities, in the area. The WHG they have is minute and could be the same type found among other east med populations, not EHG related in any way

tamilgangster
06-26-2015, 10:23 AM
Well, a good amount of modern African diversity is also ultimately owed to Eurasian input, so I'd keep that in mind. One of the main things (if not in some cases the main thing) pulling Horners, North Sudanese, the Maasai and various Fulanis away from where groups like Nilotes are clustering is ultimately their West Eurasian ancestry.



Yes, this is often why Peninsular Arabians and now these Copts tend to seem like the best example of the kinds of West Asians the East African ancestors of Horners mixed with episodically; it's because these new Copt samples and Arabians tend to mostly just be ENF (mostly lacking ANE & WHG-UHG) when it comes to their West Eurasian ancestry.



Hmm, there are people better suited to helping you figure this out (Beyoku, Lol_Race, Pgbk87 et al. ; I'll message them about your musings ;))



Yeah, it would be very interesting to see what would turn up. Adding Somalis into the mix would be very interesting so we could see just how much ancestry would prove to look "shared" between Somalis & Bejas (fellow Cushitic speakers) given that at K=7 (http://4.bp.blogspot.com/-XuGP_oNZVNM/VXhjc05uL1I/AAAAAAAACtg/7cJhcWzMJHY/s1600/2akh3k2.jpg.png) what looks to be a sort of pseudo-Cushitic component like "Ethio-Somali (http://anthromadness.blogspot.ae/2014/12/ethio-somali-is-farce.html)" forms (making up a good degree of the Maasai's ancestry and a whole lot of the Beja and Ethiopian samples' ancestry) but that ADMIXTURE run as Lol_Race noted to me at ABF is rather unstable:



Isnt most of the Eurasian admixture in Africa basal eurasian related. If so, one can isolate the nonafrican portion of DNA from a population like Masai or some other nilotic group inorder to obtain a "pure" basal eurasian sample. Unlike horners who might also have ENF type eurasian admixture, im guessing the eurasian found among Masai is purely basal eurasian

Awale
06-26-2015, 06:01 PM
THe reason why copts form a seperate component could be due to inbreeding and genetic isolation, from other communities, in the area. The WHG they have is minute and could be the same type found among other east med populations, not EHG related in any way

Well, yeah. That tends to be why a group forms it's own cluster-> isolation and such. Though I should note that they removed the most inbred members of each population/ they removed individuals who were clearly related:


A sample without enough ethnic and linguistic information was removed (211-5344). Duplicated or/and related individuals were identified by identity by descent (IBD) matrix. Nine samples were identified as being second or third-degree relatives (IBD > 0.185) with respect to other individual and were removed: 4 Copts (2115103, 211-5159, 211-5110, 211-5134), 4 Arabs (211-4947, 211-4898, 211-4921, 21022425) and 1 Beja (211-5319).

These Copts have obviously been endogamous for a while now.


The WHG they have is minute and could be the same type found among other east med populations

Could be indeed.


Isnt most of the Eurasian admixture in Africa basal eurasian related. If so, one can isolate the nonafrican portion of DNA from a population like Masai or some other nilotic group inorder to obtain a "pure" basal eurasian sample. Unlike horners who might also have ENF type eurasian admixture, im guessing the eurasian found among Masai is purely basal eurasian

The Eurasian admixture in groups like the Maasai is basically just "ENF" (which ultimately carries Basal Eurasian) like in Horners... The Maasai (as well as Datogs, Kikuyu et al.) are basically Horner admixed: http://anthromadness.blogspot.ae/2015/05/south-cushitic-admixture-in-southeast.html

Chad Rohlfsen
06-27-2015, 07:30 AM
Preliminary K10, with my data first. I'm looking at things on a global scale. This should help to increase the accuracy. I've added many new pops, including Africans, to increase the diversity. I will post those (K9-11) tomorrow if they look okay.

https://docs.google.com/spreadsheets/d/1N5AwPoQ8X2AK-r72lqqxhLMIAZ1OXdwWM4ONoeSzQMg/edit?usp=sharing

Chad Rohlfsen
06-27-2015, 06:46 PM
This K10 is a little more interesting. Dstats are matching this, for Europe. MN and Steppe equal, closer to WHG than Yamnaya, in Lithuanians. The Near East component looks more Near Eastern than ENF. WHG is higher in Egyptians than Bedouins. I think the first farmers had a good amount of WHG/UHG and dropped this Near Eastern cluster. Maybe 30-40%? Have a gander. I just added another 1000-1200 samples. It should be interesting.

https://docs.google.com/spreadsheets/d/16NZ7jw3T45q-AvpIKCMzpsYfD3gKKu3msnM5SgAjyxw/edit?usp=sharing

Kale
06-27-2015, 07:07 PM
This K10 is a little more interesting. Dstats are matching this, for Europe. MN and Steppe equal, closer to WHG than Yamnaya, in Lithuanians. The Near East component looks more Near Eastern than ENF. WHG is higher in Egyptians than Bedouins. I think the first farmers had a good amount of WHG/UHG and dropped this Near Eastern cluster. Maybe 30-40%? Have a gander. I just added another 1000-1200 samples. It should be interesting.

https://docs.google.com/spreadsheets/d/16NZ7jw3T45q-AvpIKCMzpsYfD3gKKu3msnM5SgAjyxw/edit?usp=sharing

I don't know if it's a flaw or I'm just being picky...but Yamnaya have Basal/ENF ancestry...they are 100% Steppe in that calculator. EHG have no Basal/ENF ancestry, they are 60% Steppe in that calculator. It's a muddy component, I don't like how it is called upon to express only parts of its true nature in certain circumstances.

Chad Rohlfsen
06-27-2015, 07:12 PM
There just aren't enough EHG samples to make a cluster, so it fits them in. The Steppe component is pretty solid. Lithuanians are equally close to MN and Yamnaya, with extra WHG, which this has. This steppe component has a slight tilt towards WHG and EHG, which is what fits with Corded Ware equally close to EHG as Yamnaya and Sintashta having a lot of EHG. It all evens out. This is pretty damn close, in my opinion. My new run will have Sintashta and Andronovo.

Chad Rohlfsen
06-27-2015, 07:27 PM
Here's a K11 with EHG coming out. I may need to go up to 12, and maybe WHG will come out. The Neo cluster has probably 10%+ WHG. The Steppe HG is probably 10-15% WHG.

https://docs.google.com/spreadsheets/d/14GCDcQ2m5aL_dNP95KLmMOwnd7NgB-UDdLDb4_EH3V8/edit?usp=sharing

Kale
06-27-2015, 07:59 PM
There just aren't enough EHG samples to make a cluster, so it fits them in. The Steppe component is pretty solid. Lithuanians are equally close to MN and Yamnaya, with extra WHG, which this has. This steppe component has a slight tilt towards WHG and EHG, which is what fits with Corded Ware equally close to EHG as Yamnaya and Sintashta having a lot of EHG. It all evens out. This is pretty damn close, in my opinion. My new run will have Sintashta and Andronovo.

I think it's more of an artifact of trying to impose an ancient sample onto more modern clusters. It's just beyond the more modern poles of variation reduced through mixture over time. Plus now WHG are turning out 1/3 neolithic?

Edit: When trying to base out Bronze Age/Iron Age ancestry...I think the way to go is to cut deep into the phylogeny (into paleolithic times) with an Admixture run, and decipher more recent events with an oracle tool.

Chad Rohlfsen
06-27-2015, 08:12 PM
As I said, I can go up to K12 and WHG will probably separate from EHG. However, I'm working on that K10 with more pops. I'll get back to that later. Using the Paleolithic is useless. We don't contain much from them. Kostenki, Oase1, Ust_Ishim, and MA1 make that pretty clear. There's no point in bothering with them. Only samples under 8k years.

Kale
06-27-2015, 08:26 PM
Take a second look at the distribution of N-Caucasian in those Bell Beaker samples you have...
21%...only a little less than Corded Ware, and about half of Yamnaya, probably an early BB.
14%, 10%, 10%...Three out of Five relatively close, seems like normal distribution.
2.5%...What happened here?

Ezana
07-01-2015, 05:48 PM
Heh, thanks. Btw, Lol_Race (he's told me a bit about you so I'm sure you're familiar with him :-)) and I noticed something a bit OT bit still interesting about the "Northeast Africa" plot:

https://drive.google.com/file/d/0B2ARnUeK-Y8WS2JfNGYzelpZODQ/view

It's that Somalis alongside Northern Highlanders (Habeshas + Beta Israels) & Oromos are seemingly pulling toward Aris and showing what looks like an Omotic (http://anthromadness.blogspot.ae/2015/01/the-omotic-cluster-horn-specific-cluster.html) affinity. Lol_Race asked David to throw in Hadzas to perhaps introduce something similar to Aris that may have been able to slightly differentiate Somalis from Omotic admixed Horners but the pulling toward Aris seemed to stick:

https://drive.google.com/file/d/0B2ARnUeK-Y8WWVRIYXdPeGRuOWs/view

Why is this weird? Well... It's making it seem like Somalis may actually have Omotic admixture. As you may know; Bandar from the old ABF days created a run (https://docs.google.com/spreadsheets/d/1HRFbBy3WiFGq7AX60E1LZqqYrGfizObQgkjolfZ6Qa8/edit#gid=5) that was pretty good at spotting Omotic by removing the most inbred Ari Blacksmiths & in this run Somalis generally came up 0% Omotic except for Ethiopians Somalis who had some folks among them (alongside two heavily Omotic admixed outliers) showing Omotic admixture. Now, David suggested that the lower number of markers he was forced to use for these PCAs (due to the rather poor overlap between Dobon et al.'s chip and the chips used by other studies) might just change this up a bit.

So I recently asked him to try and replicate those two above PCAs but exclude the Sudanese samples entirely so he could then use the usual high number of markers-> if things stay the same if and when he turns up with those PCAs and Somalis keep pulling toward Aris alongside other Horners-> Do you think it could mean Bandar's run was wrong and Somalis perhaps do have at least some Omotic admixture?

I wonder if these Bejas also have Omotic admixture; I asked Ethiohelix (http://ethiohelix.blogspot.ae/2015/06/more-ethiopian-uniparental-data-more.html) to run them through an ADMIXTURE analysis and while he said he was pretty rusty; he does know how to spot Omotic admixture IIRC so we'll see about these Bejas, I guess... If they lack Omotic admixture then it does lend some good credence (as Lol_Race mentioned on ABF) that Cushitic speakers originally came from further up north in Northeast Africa. Granted, that idea makes enough good sense to me from a linguistic standpoint alone..

Interesting...more data is needed, then, if the two results contradict each other even with more markers. My own gut feeling is that there's probably some low-level "Omotic" ancestry in nearly all East African groups. Not necessarily because of later admixture, however. My guess is that proto-AA probably had multiple splits and migrations and back-migrations over the millennia. Just a guess, though. A3b2 has been found at high levels in Ethiopians (~10%) in both Semitic and Cushitic speakers, as much as 15% in one sample of Amharas. It's supposedly representative of Nilotic populations and exists in low levels (~3%) in Egyptians as well. However, Somalis very rarely have this clade. IIRC, a similar pattern exists with haplogroup B as well, although less prominently (like 2-3%?). It might be explained away with cultural considerations, given the importance of clans. Smaller haplogroups could have been bred out. 75% paternal lineage in one haplogroup is not the norm for populations that have existed for thousands of years.

DMXX
07-01-2015, 05:52 PM
Just a quick reminder, folks; this thread is with respect to African results (specifically Sudani auDNA).

We have countless threads dissecting Eurasian genetic history, Yamnaya etc. as it is.

Please continue those off-topic tangents elsewhere on the forum. Thanks in advance.

Awale
07-08-2015, 09:55 AM
However, Somalis very rarely have this clade.

Yes, mostly cos Somalis have bottle-necked when it comes to E-V32 and T (https://en.wikipedia.org/wiki/Haplogroup_T-M184). T seems to be very common among Dir and Isaaq clan Somalis (basically Northwestern Somalis) and present at a rate of like 20-30% among other clans (with E-V32 being the dominant marker) whilst E-V32 seems to hold the same position among Dir folks for example (basing this on those old Dire Dawan Dirs who were tested & the recent small sample size from Djibouti (http://ethiohelix.blogspot.ae/2015/06/improved-resolution-of-e-m215-aka-e3b.html)-> Djiboutian Somalis are mostly Dir clan members). As a result of all this; J1 and A-M13/ A3b2 which still exist among Somalis have been beaten down into near obscurity. But there still are a noticeable number of J1 and A3b2 Somalis (there are even J1 carriers among the Ethiopian Somali samples you usually see thrown around autosomal studies (http://ethiohelix.blogspot.ae/2015/06/more-ethiopian-uniparental-data-more.html)); A3b2 is more rare though, in my experience.



Interesting...more data is needed, then, if the two results contradict each other even with more markers.

Yes, I'll get to work on getting PCAs with a high number of markers. I'll make sure the Sudanese are excluded this time to avoid lowering the markers for them... That way we can see if Somalis still keep displaying a sort of pull toward Omotics (Aris).

I also wanna try and basically throw the "Ethiopian" samples from Dobon et al. into a PCA with Horners-> that way we can see if they aren't somehow "Sudanese admixed" (highly doubt they are) and then if they prove not to be; we can use them in a PCA with these Sudanese samples using a high number of markers as we'll be limited to only this study's populations which it sampled for about 200k SNPs. Things might not be ideal given the lack of Aris who seem to help in the creation of a "Horner cluster" so the Ethiopians may not end up separating much from the Sudanese but it'll be interesting to see a PCA like that nonetheless.


My own gut feeling is that there's probably some low-level "Omotic" ancestry in nearly all East African groups. Not necessarily because of later admixture, however. My guess is that proto-AA probably had multiple splits and migrations and back-migrations over the millennia.

I somewhat hold a similar view. I think a lot of Afro-Asiatic groups share pretty profusely ancient ancestry similar to how various Indo-European speakers do but perhaps in an even more complex manner as you suggest. For example-> Horners often turn up with "Maghrebi" ancestry at the lower Ks of some runs like Hodgson et al.'s (http://www.plosgenetics.org/article/fetchObject.action?uri=info:doi/10.1371/journal.pgen.1004393.g002&representation=PNG_L) but of course we seem to lack WHG so this is not something recent-> instead I tend to think this is a sign of very ancient shared ancestry between Berbers & Cushites. I began to think this even more after a good friend you have here (Agamemnon) told me the following recently:



Btw... Are Cushitic, Berber and Semitic close in any particular way. I thought the grouping was Berber-Semitic-Egyptian?


Hehehehe, yeah that's also what I used to think as well but upon closer inspection it's pretty f*cking obvious that Berber, Semitic & Cushitic derive from the same AA node. The whole Berber-Egyptian-Semitic subgrouping is an illusion produced by the vast amount of research in these three branches. I'm probably gonna surprise you but Cushitic is more pristinely AA than Egyptian. I disagree with Huehnergard (one of the greatest specialists in Semitic studies) on this point, as he's still parroting the Berber-Egyptian-Semitic claim. Semitic and Berber are the closest relatives within this node though, both of them share the exact same verbal morphological patterns with Cushitic (and this is a big argument in favour of a Cushitic-Berber-Semitic node). IMO the community which spoke this early AA dialect lived in the Nile Valley, and this was a foraging community.

Awale
10-30-2015, 02:58 PM
Made a sort of "follow-up" post at my blog for this study but this time by comparing the results of this study (and those PCAs owed to David) to the Gedmatch results of a Sudanese Arab chap I encountered on 23andme:

Inferences that can be made from a Sudanese Arab's Gedmatch results (http://anthromadness.blogspot.ae/2015/10/inferences-that-can-be-made-from.html)

ThaYamamoto
02-05-2021, 02:34 AM
Assuming these are also too low overlap with the g25? I found em and thought I'd hit tha jackpot lol