PDA

View Full Version : Africa before ancient DNA: The deepest splits



Awale
12-23-2016, 01:14 PM
Alright, I plan to make a somewhat short post about this at Anthromadness in due time but Uni-work and other posts due before it have had me distracted for the time being but here goes:


http://oi64.tinypic.com/1ok5dg.jpg


I basically asked David to run the South-Sudanese, the Gumuz, Yorubas & Mandenkas, Mbutis & Biakas and, finally, the "San" to sort of try and represent the deepest splits we so often notice via ADMIXTURE in PCA form:

https://lh5.googleusercontent.com/-r449SBzMGHQ/VOguSxdVUDI/AAAAAAAACFc/r3hH6Yc-PNcuo8G7MDVWIXgs2kB67QuqACL0B/w633-h874-no/journal.pgen.1004393.g002.png


So... My two cents based on the above would be:

http://oi68.tinypic.com/309hgrt.jpg

Sent to an old friend (Lank) in private via email

So... Theorize away. What do you think ancient DNA from across Africa will do to the diversity that PCA displays. Have you noticed other deep splits across the continent that I haven't? Share away if so. Consider this a place to dump some theories and discuss said theories whilst gauging how probable they are (share evidence you think back up your theories as well). It'll be interesting to see how much many of you have picked up on before we see plenty of aDNA from Africa someday.

pgbk87
12-23-2016, 02:44 PM
Not a theory, but I'm curious why you did not include MOTA. I'm also curious to see whether those upcoming ancient Turkana samples would deviate further to the north than contemporary Nilotes.

Nonetheless, I'd be curious if the genetic distance we see between West Africans and East Africans is drift or admixture, or both. West Africans carry a lot of Y-DNA and mtDNA lineages that are subsets of East African ones, but also very unique lineages to West Africa. There was probably a (now extinct?) hunter-gatherer population in West Africa that was absorbed by E-V38, mtDNA L2* + L3b'd'e "Proto West Africans". L1b* and L2a1 are hella old, and E-V33 is from the southwestern Sahara.

I find it interesting that the Pygmy groups deviate from one another so drastically. Are the Mbuti closer to the San than the Biaka? Bantu admixture in the Biaka I'd guess.

beyoku
12-23-2016, 02:52 PM
I am interested in seeing if populations that seem like a composite of ancestries still look the same in ancient DNA.

pgbk87
12-23-2016, 03:09 PM
I am interested in seeing if populations that seem like a composite of ancestries still look the same in ancient DNA.

Ie. South Sudanese/Anuak vs Gumuz.

SWAHILLI_PRINCE16
12-23-2016, 03:14 PM
Interesting! are there any genetic studies on coastal East Africans like the Swahili? it will be interesting to see the genetic composition of different Swahili tribes as well as the ancient component. Not much genetic research has been done on Swahili people which is a shame considering the multi-ethnic and multicultural diversity in the Coast of East Africa.

Before the Swahili became Swahili they were bantu people living in Eastern Congo. There were Cushitic and Nilotic groups living in the East Coast of Africa at that time. The migration towards Ethiopia and Southern Somalia was due to the bantu migration as well as the harsh tropical climate. These Bantu people converted to Islam long before the Arabs and Persians were in the East Coast of Africa. This could suggest that the Bantu people living in Southern Somalia or Shungwaya had relations to the Axumite and Cushitic East African kingdoms in East Africa?

The migration with the Bantu from Southern Somalia towards South Eastern Africa was due to war and conflict with the Oromo. These Bantu people became Swahili meaning people living in the East Coast of Africa. Later on some Swahili people intermarriage over time with first Persian Shirazis, Yemenis, East Asians then Omanis, Portuguese and South Asians.

It really would be interesting to support this historical research with Genealogy.

Cheers Abdurashid

Power77
12-23-2016, 03:20 PM
Such a shame that we lack African ancient DNA so far:(. However, I wouldn't be surprised if E1a (my own Y-DNA:)) turns out to be dominant in Pre-Neolithic samples from the Maghreb and West Africa:nod:.

xKeleix
12-23-2016, 05:18 PM
I'd wager that there's much diversity not being captured between San and Mbuti. Though they're both Hunter Gatherer groups and have some sort of relation which is well established, that PCA placed them in a position which isn't really intuitive. There's definitely something else going on there. The first dimension is not showing their divergence in the way that it shows HG vs. Farmer divergence, but I'd expect it to be overt simply because these populations allegedly separated very early.

I'm also really curious about West African ancients and wonder if there are some extreme populations hidden within the region. Although there is marked modern diversity, the continent doesn't seem to meet my expectations as "the most diverse" when Eurasian admixture is accounted for. Again, something else is going on here, or I'm missing something (ascertainment bias?).

Edit: Further, the enormity of African diversity which was mentioned frequently say, 5 or so years ago, hasn't really been captured autosomally, as far as I can tell anyway. One must wonder why this intra-African diversity isn't as apparent as we'd expect given the origin of Homo Sapiens.

Lank
12-23-2016, 07:53 PM
I'll paste my response here too then. :P



Agreed on all counts. I definitely think the admixture in African prehistory will be at least as complicated as in Eurasia, given the much longer history of modern humans on the continent. Once technology and available resources permit ancient DNA analysis more widely in Africa, we will certainly find a very complicated history, including but not at all limited to archaic humans. We will be seeing divergence on the order of hundreds of thousands of years, not merely tens of thousands of years as in Eurasian modern human structure and admixture.


The common ancestor of Neanderthals/Denisovans and Homo sapiens, Homo heidelbergensis, is believed to have migrated out of Africa around 600 kya. Homo sapiens is estimated to be 200 ky old based on a combination of fossils (Omo 1), mtDNA and Y-DNA, although the A00 Y found in recent years is a bit older than 200 ky). But this division between modern and archaic is somewhat arbitrary, since modern humans were at least somewhat compatible with Neanderthals/Denisovans, and evolved in Africa continuously ever since that split. And the "modern behavioral package", i.e. LSA/UP technology, only dates to 50 kya, so there are those who only consider humans since 50 kya to be "modern" (although Khoisan, who split from the rest 200 kya, are also "behaviorally modern").

However, if we accept that 200 kya limit for AMH, it is quite likely that Africans have some ancestry from groups more divergent than that. There are plenty of African fossils with a mixture of modern and archaic traits, and there is a lot of variety in Middle Stone Age African cultures. This fits with the population boom we see in proto-Sapiens from genetic data (http://www.nature.com/nature/journal/v505/n7481/images_article/nature12886-f4.jpg) after the split from Neanderthals/Denisovans, whereas the latter had a population crash, no doubt due to the harsh Eurasian climate. There were a lot of different groups. Many of these would have been more archaic than the 200 kya TMRCA, but not at all as divergent as Neanderthals (e.g. the A00 dudes). And I wouldn't be surprised if some who were even more divergent than Neandeerthals survive in some of our DNA (as has been reported at least in Pygmies)."


Mota's sequence came as a pleasant surprise, somewhat confirming earlier speculation that Paleolithic Ethiopians were Omotic-like (without Eurasian admixture). Let's hope we have more such surprises in 2017! Some of these, such as Epipaleolithic (E-M35?) North Africans may be relevant for Eurasians as well. Late MSA Africans, forming the predominant ancestry of all modern humans, certainly would be. Hofmeyr, the LSA South African who supposedly had craniometric similarities to Upper Paleolithic Eurasians, was apparently sent to the aDNA lab in Copenhagen over two years ago (https://www.pressreader.com/south-africa/the-herald-south-africa/20140627/281586648685884).

Lank
12-23-2016, 08:02 PM
I'm also really curious about West African ancients and wonder if there are some extreme populations hidden within the region. Although there is marked modern diversity, the continent doesn't seem to meet my expectations as "the most diverse" when Eurasian admixture is accounted for. Again, something else is going on here, or I'm missing something (ascertainment bias?).

Edit: Further, the enormity of African diversity which was mentioned frequently say, 5 or so years ago, hasn't really been captured autosomally, as far as I can tell anyway. One must wonder why this intra-African diversity isn't as apparent as we'd expect given the origin of Homo Sapiens.
One important point to make is that Africans are less drifted, due to higher effective population sizes. This is why Fst distances between African groups are relatively low; much of the great African diversity is found within groups rather than between them. Yet analyses based on IBD, mtDNA and Y-DNA reveal fairly deep divergences do exist, even between the less isolated East and West Africans (after accounting for Eurasian admixture in the former).

rozenfeld
12-23-2016, 08:09 PM
Such a shame that we lack African ancient DNA so far:(. However, I wouldn't be surprised if E1a (my own Y-DNA:)) turns out to be dominant in Pre-Neolithic samples from the Maghreb and West Africa:nod:.

To be pedantic: There is one ancient African genome, namely Mota from Ethiopia: http://science.sciencemag.org/content/early/2015/10/07/science.aad2879

Power77
12-23-2016, 08:25 PM
I wonder if we have any Y-DNA data from the Iberomaurusian culture (I know some scientists published a few mtDNA results from this culture):unsure:. Do any of you here have any information on this issue:confused:?

Awale
12-23-2016, 08:46 PM
Not a theory, but I'm curious why you did not include MOTA.

Simple:


His African elements just seem like some sort of composite between San-like and "East African" elements (basically similar to Aris (http://anthromadness.blogspot.ae/2016/06/the-african-west-eurasian-elements-in.html) and the Gumuz) so I wasn't that interested in seeing how he'd turn out as a result.
And he shows some overt Eurasian affinities. I doubt he actually has Eurasian admixture but if there's some chance that the 10-20% Eurasian admixture he often displays has some merit to it (and even if it doesn't); he just wouldn't fit the criteria I wanted for the PCA... Which was to only throw in African populations mostly lacking (<5%) in overt Eurasian admixture and/or affinities.



Nonetheless, I'd be curious if the genetic distance we see between West Africans and East Africans is drift or admixture, or both. West Africans carry a lot of Y-DNA and mtDNA lineages that are subsets of East African ones, but also very unique lineages to West Africa. There was probably a (now extinct?) hunter-gatherer population in West Africa that was absorbed by E-V38, mtDNA L2* + L3b'd'e "Proto West Africans". L1b* and L2a1 are hella old, and E-V33 is from the southwestern Sahara.

I don't think all their differences are caused by drift and pretty much went into that here (http://oi68.tinypic.com/309hgrt.jpg). I think there are pre-historic lineages shared between "East African" and "West African" and that they're possibly going to turn out to be mixtures like various West Eurasian clusters did after we started getting ancient DNA and I feel that way because of reasons similar to what you noted here:


West Africans carry a lot of Y-DNA and mtDNA lineages that are subsets of East African ones, but also very unique lineages to West Africa. There was probably a (now extinct?) hunter-gatherer population in West Africa that was absorbed by E-V38, mtDNA L2* + L3b'd'e "Proto West Africans". L1b* and L2a1 are hella old, and E-V33 is from the southwestern Sahara.


I find it interesting that the Pygmy groups deviate from one another so drastically. Are the Mbuti closer to the San than the Biaka? Bantu admixture in the Biaka I'd guess.

Yes, the "Pygmies" splitting up that much surprised me too. You're probably right and there's some "West African" related admixture in the Biaka, makes sense given that they're pulling westward toward Yorubas and Madenkas.

beyoku
12-24-2016, 12:16 PM
Ie. South Sudanese/Anuak vs Gumuz.

Even Fulani, Horners and Nilo/Cushitic composite populations. The fact that they have recent admixture does not exclude an ancient admixed substratum. Also interested in seeing Nilotic influences in Mbuti and Mangbetu which seemed to have totally disappeared. Lastly anything associated with Haplogroup B lineages which seem to be the best candidates for populations that have undergone language loss of their isolates. Besides hunter gatherers notice the Laal and Gumuz carry divergent B lineages. Their languages have or in the case of Laal, still are classified as isolate. I would argue that B2a fits this description prior to its association with Nilotic.

beyoku
12-24-2016, 12:19 PM
@ SWAHILLI_PRINCE16 - There are one or two preliminary abstracts on ancient DNA in that area. I will see if I can locate them.

SWAHILLI_PRINCE16
12-24-2016, 02:50 PM
Heres what I found based on my fellow Swahilis. I couldn't find out the specific ancestry composition however.

Source: https://en.wikipedia.org/wiki/Shirazi_people

According to genetic analysis by Msaidie et al. (2010), who sampled the uniparental DNA of the inhabitants of the Comoros archipelago (Grand Comore: 170 men, 67 women; Anjouan: 104 men, 69 women; Moheli: 107 men, 60 women), the most common paternal haplogroups among the Comorian Shirazi are E1b1a-M2 (41%) and E2-M90 (14%). These Y-DNA clades are frequent among other Bantu-speaking populations on the east African mainland, which points to shared origins. The remaining Comorians primarily carry the haplogroups E1b1b-V22, E1b1b-M123, F*(xF2, GHIJK), G2a, I, J1, J2, L1, Q1a3, R1*, R1a*, R1a1 and R2 (29.7%). Of these latter clades, the particular haplotypes that are found in Comoros were observed to be most closely related to those in South Iran. This suggests that these northern lineages were brought by early Shirazi merchants from Persia between 1200-1300 CE, as they established local trading posts on the Comoros islands. Around 6% of the Comorians also bear the O1 haplogroup, which indicates a minor Southeast Asian influence.[18]

Source: http://core.tdar.org/document/398421/decoding-the-swahili-ancient-dna-studies-on-the-kenyan-coast

Our project examines the role of migration in the development of the large autonomous Swahili towns and city-states that grew out of small fishing, agrarian, and pastoral settlements on the East African coast in the late first millennium CE. Our sample is comprised of 97 individuals from three sites on the Kenya coast: Mtwapa (N=72; 900-1732 BCE) near Mombasa, and two sites in the Lamu archipelago, Manda (N=16; 800-1400 BCE), and Shanga (N=9; 800-1400 BCE). The teeth were well preserved and about 80% of the samples yield genetic material. We sequenced the HVRI of the mitochondrial control region. We found that, while early Swahili populations were primarily of African origin and consisted of haplogroups commonly found throughout East Africa and in Bantu speaking groups in sub-Saharan Africa, but the Swahili were much more diverse than has been commonly supposed and included haplotypes less commonly found in the area today.
SAA 2015 abstracts made available in tDAR courtesy of the Society for American Archaeology and Center for Digital Antiquity Collaborative Program to improve digital data in archaeology. If you are the author of this presentation you may upload your paper, poster, presentation, or associated data (up to 3 files/30MB) for free. Please visit http://www.tdar.org/SAA2015 for instructions and more information.

pgbk87
12-25-2016, 12:49 PM
Even Fulani, Horners and Nilo/Cushitic composite populations. The fact that they have recent admixture does not exclude an ancient admixed substratum. Also interested in seeing Nilotic influences in Mbuti and Mangbetu which seemed to have totally disappeared. Lastly anything associated with Haplogroup B lineages which seem to be the best candidates for populations that have undergone language loss of their isolates. Besides hunter gatherers notice the Laal and Gumuz carry divergent B lineages. Their languages have or in the case of Laal, still are classified as isolate. I would argue that B2a fits this description prior to its association with Nilotic.

Agreed on haplogroup B. What mtDNA lineages would you associate with it?

SWAHILLI_PRINCE16
12-25-2016, 05:09 PM
Even Fulani, Horners and Nilo/Cushitic composite populations. The fact that they have recent admixture does not exclude an ancient admixed substratum. Also interested in seeing Nilotic influences in Mbuti and Mangbetu which seemed to have totally disappeared. Lastly anything associated with Haplogroup B lineages which seem to be the best candidates for populations that have undergone language loss of their isolates. Besides hunter gatherers notice the Laal and Gumuz carry divergent B lineages. Their languages have or in the case of Laal, still are classified as isolate. I would argue that B2a fits this description prior to its association with Nilotic.

Ill be the 2nd to agree with that, I wouldn't say B2a is Nilotic but a widespread East Central Africa haplogroup.

Awale
12-25-2016, 10:13 PM
One important point to make is that Africans are less drifted, due to higher effective population sizes. This is why Fst distances between African groups are relatively low; much of the great African diversity is found within groups rather than between them. Yet analyses based on IBD, mtDNA and Y-DNA reveal fairly deep divergences do exist, even between the less isolated East and West Africans (after accounting for Eurasian admixture in the former).

Yep. I actually never got or knew this properly until you explained it to me ages back and I then went onto point this out in a post about Aris (http://anthromadness.blogspot.ae/2016/06/the-african-west-eurasian-elements-in.html). But yeah, a lot of African groups' Fst distances and IBS results will really fool you in terms of how "diverged" they seem but looking at stuff like haplotypic data will show you some real deep splits in terms of time divergence and this indeed seems to be because various African macro-populations had larger population sizes during various points in Human pre-history (= Less conducive toward drift. But there is still some pretty substantive drift within Africa, nevertheless).