PDA

View Full Version : [Pre-print] Chabu (Modern HG from SW Ethiopia)



blackflash16
01-12-2019, 07:52 PM
Source: Biorxiv (https://www.biorxiv.org/content/early/2019/01/11/517730)

Hunter-gatherer genomes reveal diverse demographic trajectories following the rise of farming in East Africa

Abstract

A major outstanding question in human prehistory is the fate of hunting and gathering populations following the rise of agriculture and pastoralism. Genomic analysis of ancient and contemporary Europeans suggests that autochthonous groups were either absorbed into or replaced by expanding farmer populations. Many of the hunter-gatherer populations persisting today live in Africa, perhaps because agropastoral transitions occurred later on the continent. Here, we present the first genomic data from the Chabu, a relatively isolated and marginalized hunting-and-gathering group from the Southwestern Ethiopian highlands. The Chabu are a distinct genetic population that carry the highest levels of Southwestern Ethiopian ancestry of any extant population studied thus far. This ancestry has been in situ for at least 4,500 years. We show that the Chabu are undergoing a severe population bottleneck which began around 40 generations ago. We also study other Eastern African populations and demonstrate divergent patterns of historical population size change over the past 60 generations between even closely related groups. We argue that these patterns demonstrate that, unlike in Europe, Africans hunter-gatherers responded to agropastoralism with diverse strategies.


https://i.postimg.cc/L8Gpjfvq/chabu.png

https://i.postimg.cc/2S1YqTdR/chabu-pca.png

NetNomad
01-12-2019, 10:01 PM
Ethiopia is more diverse than I thought originally, interesting.

Lank
01-13-2019, 04:35 PM
Interesting! Chabu even appear to lack the minor Afroasiatic-related ancestry detected in Mota in the ADMIXTURE analysis.

NetNomad
01-13-2019, 06:37 PM
Afroasiatic-related ancestry

The yellow Afroasiatic-related cluster seems unstable to me due to the lack of more North African and Arabian samples (the # of samples has influence on the final result). Only using Iberians (without Arabians) as an outgroup was kind of lazy on their part.

Source:

A tutorial on how not to over-interpret STRUCTURE and ADMIXTURE bar plots
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6092366/

''Case Study 3: Worldwide human data

An important consideration in any STRUCTURE analysis is sample size. This is vividly illustrated by the analyses of Friedlaender et al.18 who augmented a pre-existing microsatellite data set from a worldwide collection by a similar number of samples from Melanesia, in order to study genetic relationships between Melanesians, for which purpose their sample was excellent. For K = 2, their analysis infers Papua New Guinea (PNG) as one ancestral population and Western Eurasia and Africa as the other, with East Asians being represented as genetic mixtures (Fig. 5b). This analysis differs from that of Rosenberg et al.19 for K = 2 who had only a small number of Melanesians in their sample, and who found Native Americans rather than Melanesians to be the unadmixed group (Fig. 5f). For K = 6, both models distinguish between all 5 continental groups (Americans, Western Eurasians, Africans, East Asians, and Oceanians). Rosenberg et al. split Native American groups into two ancestral populations while Friedlaender et al. infer that Melanesians have two ancestral populations (Fig. 5a). Rosenberg et al.4 also found the Kalash, an isolated population in Pakistan, to be the sixth cluster.''

Also, this image is illustrative:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6092366/bin/41467_2018_5257_Fig2_HTML.jpg

Lank
01-13-2019, 07:04 PM
The yellow Afroasiatic-related cluster seems unstable to me due to the lack of more North African and Arabian samples (the # of samples has influence on the final result). Only using Iberians (without Arabians) as an outgroup was kind of lazy on their part.
You're right. But you might at least expect the overall results to reflect the most basic differences, such as the Eurasian vs. African affinity of the components. Chabu form one of the poles of variation in the PCA as well.

Another note: The Shekkacho (Omotic) samples are quite similar to their Oromo neighbors. Shekkacho have the highest Y-DNA J levels (52%) found in Ethiopia so far, followed by the closely related Kefa people. It would be interesting to see the main J subclades found in these groups.

pgbk87
01-18-2019, 02:40 AM
Is that "Afro-Asiatic" component similar to the alleged "Ethio-Somali" compenent? It appears to behave similarly.

On another note, it seems like the the South Sudan/Sudan/Ethiopia/Kenya border regions are full of genetically distinct people groups. The Nilo-Saharan "language family" is looking less credible on a genetic and linguistic level (at least in Ethiopia).

Awale
01-19-2019, 12:27 PM
Is that "Afro-Asiatic" component similar to the alleged "Ethio-Somali" compenent? It appears to behave similarly.

This looks pretty much exactly like Hodgson et al. but just with some new samples, really. It's a pretty nice result seeing the SW Ethiopian component appear in Horners, Somalis included, then suddenly not appear in areas like Sudan. Only the Beni-Amer Beja who are known to intermingle with Tigres show it. Be more interesting to toy with all these samples using something like nMonte, though.

Ignis90
01-26-2019, 07:14 PM
Interesting case! Usually, we need ancient and ghost populations because a population outside modern variation is needed. Here, the living Chabu are forming a pole of the variation, at the expense of Mota.

Ignis90
01-26-2019, 08:23 PM
Interesting case! Usually, we need ancient and ghost populations because a population outside modern variation is needed. Here, the living Chabu are forming a pole of the variation, at the expense of Mota.

Michalis Moriopoulos
01-27-2019, 06:39 PM
Fascinating stuff. I looked up the Chabu. The nature of their language is controversial; some think it's an isolate while others say it's Nilo-Saharan. In any case, physically they resemble the Hadza to my eyes, which makes sense given their admixture results. I guess the purple component (maximized in Mota) is probably ~3/4 Ancient East African HG (roughly Dinka-like) + ~1/4 Southern African HG (San-like).

Notice that the Sudanese Arabs and Nubians have a lot of that yellow [Afro-Asiatic] "Horner" component, but the Nilotes don't. I really can't wait for ancient DNA from Nubia. I intuit that the West Eurasian admixture in the region predates the Arabization of NE Africa, and that the inhabitants of Nubia and North Sudan were at one point similar to Horn Africans, just with some extra "blue" Nilotic admixture (from hypothetical Nubian language bearers) and probably "green" Egyptian-related admixture, though I suppose some of the latter could have indeed come from the Arabs. Unlike the Nilotes, many NE African groups with the blue component don't have any "green" West African-related admixture, which suggests that the Nilotes were possibly affected by the Bantu migration (with the Nuer being the least affected). Any thoughts?

Megalophias
01-27-2019, 07:25 PM
That's more or less how I'd bet too - West Eurasian (prob confounded with North African) ancestry present in Sudan since the later phase of the Green Sahara (~5500 BC) and possibly much earlier. Main element of Mota/Omotic could be from Ethiopian Highland LGM refuge population while Nilotes perhaps come from savannas further west? - don't know relevant archaeology. And all likely a mix of deeply divergent lineages from before the LGM.

West African affinity in Nilotes need not relate to Bantu - there are other Niger-Congo-speaking people spread across north of the Bantu belt, e.g. Azande extending into South Sudan, and IIRC Central Sudanic speakers so far tested are also intermediate between West and East Africans. The transitional region (e.g. Central African Republic) is poorly sampled genetically though (and there's very little archaeological research either).

pgbk87
01-29-2019, 01:31 AM
Fascinating stuff. I looked up the Chabu. The nature of their language is controversial; some think it's an isolate while others say it's Nilo-Saharan. In any case, physically they resemble the Hadza to my eyes, which makes sense given their admixture results. I guess the purple component (maximized in Mota) is probably ~3/4 Ancient East African HG (roughly Dinka-like) + ~1/4 Southern African HG (San-like).

Notice that the Sudanese Arabs and Nubians have a lot of that yellow [Afro-Asiatic] "Horner" component, but the Nilotes don't. I really can't wait for ancient DNA from Nubia. I intuit that the West Eurasian admixture in the region predates the Arabization of NE Africa, and that the inhabitants of Nubia and North Sudan were at one point similar to Horn Africans, just with some extra "blue" Nilotic admixture (from hypothetical Nubian language bearers) and probably "green" Egyptian-related admixture, though I suppose some of the latter could have indeed come from the Arabs. Unlike the Nilotes, many NE African groups with the blue component don't have any "green" West African-related admixture, which suggests that the Nilotes were possibly affected by the Bantu migration (with the Nuer being the least affected). Any thoughts?

All great points, except the bolded. North Sudanese uniparental markers point to male driven Arabization of a predominantly African population within the past 1,500 years, whereas Horners have more balanced distribution of Horn African and MENA uniparental markers.

Michalis Moriopoulos
01-29-2019, 02:37 AM
Yeah, that's what's giving me pause. I thought that the prevalence of J (https://en.wikipedia.org/wiki/Y-DNA_haplogroups_in_populations_of_Sub-Saharan_Africa) in the Nubians (43.6%) and Sudanese Arabs (47.1%) might suggest Islamic era Arab gene flow, as well, but the Beja also have a high amount (38.1%). Is it reasonable to assume major Arab gene flow into the Beja, as well? Considering their proximity, it seems reasonable to assume that all of these groups got their J from the same source populations (Arabs or otherwise), unless the subclades tell a different story.

Angoliga
01-29-2019, 06:58 AM
Fascinating stuff. I looked up the Chabu. The nature of their language is controversial; some think it's an isolate while others say it's Nilo-Saharan. In any case, physically they resemble the Hadza to my eyes, which makes sense given their admixture results. I guess the purple component (maximized in Mota) is probably ~3/4 Ancient East African HG (roughly Dinka-like) + ~1/4 Southern African HG (San-like).

Notice that the Sudanese Arabs and Nubians have a lot of that yellow [Afro-Asiatic] "Horner" component, but the Nilotes don't. I really can't wait for ancient DNA from Nubia. I intuit that the West Eurasian admixture in the region predates the Arabization of NE Africa, and that the inhabitants of Nubia and North Sudan were at one point similar to Horn Africans, just with some extra "blue" Nilotic admixture (from hypothetical Nubian language bearers) and probably "green" Egyptian-related admixture, though I suppose some of the latter could have indeed come from the Arabs. Unlike the Nilotes, many NE African groups with the blue component don't have any "green" West African-related admixture, which suggests that the Nilotes were possibly affected by the Bantu migration (with the Nuer being the least affected). Any thoughts?



That's more or less how I'd bet too - West Eurasian (prob confounded with North African) ancestry present in Sudan since the later phase of the Green Sahara (~5500 BC) and possibly much earlier. Main element of Mota/Omotic could be from Ethiopian Highland LGM refuge population while Nilotes perhaps come from savannas further west? - don't know relevant archaeology. And all likely a mix of deeply divergent lineages from before the LGM.

West African affinity in Nilotes need not relate to Bantu - there are other Niger-Congo-speaking people spread across north of the Bantu belt, e.g. Azande extending into South Sudan, and IIRC Central Sudanic speakers so far tested are also intermediate between West and East Africans. The transitional region (e.g. Central African Republic) is poorly sampled genetically though (and there's very little archaeological research either).

- very tantalizing dialogue here :)

I didn't want to comment since it's mostly conjecture but maybe it'll lead somewhere meaningful, I'd like to know your opinion:




I wonder if some of Greenberg's controversial "wastebasket" languages within Nilo-Saharan are actually relics of "proto-nilosaharan" pops that didn't migrate N/W into the green sahara following the LGM.

The Kenyan A-M13 isolate on Yfull (A-YP4751* (https://www.yfull.com/tree/A-YP4751*/)) might be indicative of this; the subclade's TMRCA relative to all other A-M13s is as old as the hg itself (11,000 ybp) which would've originated in and around the Lake Turkana basin bordering South-Sudan and Ethiopia.

The tentatively placed Kuliak Languages in NS would be another example, similar to the Chabu, the Nyangia (Kuliak speakers) are also hunter-gather pops.
I'm guessing these pops didn't follow or discontinued practicing the aqualithic traditions of the major NS groups that pushed north-westerly throughout watercourses in the Nile/Chari-valley and southern-sahara.
The isolation of these HGs for over 10 millennia could be what's making them so hard to classify within a language group; their lexicon would've developed in a completely different trajectory.

It's unfortunate we don't have any of their Y-DNA samples but I'd wager the Gumuz, who also speak a language controversially placed within Nilo-Saharan, would fit neatly in this "left-behind" East-African LGM category -- the Gumuz lack of L2 mtdna and negligible West-African affinity makes them a good candidate.


Due to their lack of substantial West-African, at least relative to Central-Sudanics (agro-pastoralists), I used to think upper-nilotes like the Dinka and Nuer didn't directly participate in cross-"green-saharan" migrations -- I associated their West-African affinity with a more archaic less-obvious connection with W.African pops. Though nowadays I'm coming across more findings demonstrating upper-nilotes (more "full-blown" pastoralists) occupying a wider, more westerly pastoral range of the southern-sahara, surrounding the far peripheral of lake mega chad -- the ancient saharan cattle cults and spread of earlier-LGM pottery traditions could make this plausible.

Another possible indicator, though non-archaeological: I'm aware there's other factors to skin-tone but a more northerly latitude within the southern-sahara might explain why presentday upper-nilotes are among the most melanin rich pops in the world.


Global Solar Radiation Map
https://i.imgur.com/J5mPQzd.png

Lank
01-29-2019, 09:32 PM
- very tantalizing dialogue here :)

I didn't want to comment since it's mostly conjecture but maybe it'll lead somewhere meaningful, I'd like to know your opinion:




I wonder if some of Greenberg's controversial "wastebasket" languages within Nilo-Saharan are actually relics of "proto-nilosaharan" pops that didn't migrate N/W into the green sahara following the LGM.

The Kenyan A-M13 isolate on Yfull (A-YP4751* (https://www.yfull.com/tree/A-YP4751*/)) might be indicative of this; the subclade's TMRCA relative to all other A-M13s is as old as the hg itself (11,000 ybp) which would've originated in and around the Lake Turkana basin bordering South-Sudan and Ethiopia.

The tentatively placed Kuliak Languages in NS would be another example, similar to the Chabu, the Nyangia (Kuliak speakers) are also hunter-gather pops.
I'm guessing these pops didn't follow or discontinued practicing the aqualithic traditions of the major NS groups that pushed north-westerly throughout watercourses in the Nile/Chari-valley and southern-sahara.
The isolation of these HGs for over 10 millennia could be what's making them so hard to classify within a language group; their lexicon would've developed in a completely different trajectory.

It's unfortunate we don't have any of their Y-DNA samples but I'd wager the Gumuz, who also speak a language controversially placed within Nilo-Saharan, would fit neatly in this "left-behind" East-African LGM category -- the Gumuz lack of L2 mtdna and negligible West-African affinity makes them a good candidate.


Due to their lack of substantial West-African, at least relative to Central-Sudanics (agro-pastoralists), I used to think upper-nilotes like the Dinka and Nuer didn't directly participate in cross-"green-saharan" migrations -- I associated their West-African affinity with a more archaic less-obvious connection with W.African pops. Though nowadays I'm coming across more findings demonstrating upper-nilotes (more "full-blown" pastoralists) occupying a wider, more westerly pastoral range of the southern-sahara, surrounding the far peripheral of lake mega chad -- the ancient saharan cattle cults and spread of earlier-LGM pottery traditions could make this plausible.

Another possible indicator, though non-archaeological: I'm aware there's other factors to skin-tone but a more northerly latitude within the southern-sahara might explain why presentday upper-nilotes are among the most melanin rich pops in the world.


Global Solar Radiation Map
https://i.imgur.com/J5mPQzd.png
This is an interesting topic, here is my very nonexpert opinion. :)

Based on the available genetic data, I doubt that proto-Nilo-Saharans had much if any Mota-related ancestry. The more diverged NS languages, with inconsistent classifications within or outside of the NS family, tend to show signs of divergent groups, potentially speaking languages from unknown language families, coming into contact with Nilo-Saharans. This may be what obscures their classification with other NS languages. They are either spoken by populations living close to or in Ethiopia, probably (confirmed in the case of Gumuz, likely to be true of Koman speakers) carrying significant Mota-related ancestry, or hunter-gatherer groups, whose original languages may have had little to do with Nilo-Saharan.

The Chabu are hunter-gatherers from a forest in the southwestern Ethiopian highlands, with evident local Paleolithic ancestry. But they neighbor the Majangir, who form part of a Nilo-Saharan belt stretching into South Sudan. Most of the loanwords (https://en.wikipedia.org/wiki/Shabo_language) into Chabu are Majangir. Within Sudan and South Sudan, the Mota-related component is clearly linked to proximity to Ethiopia:
https://i.postimg.cc/L8Gpjfvq/chabu.png
The Kuliak are also hunter-gatherers and supposedly have a core of non-NS vocabulary suggesting language shift, according to Wikipedia (https://en.wikipedia.org/wiki/Kuliak_languages) citing Blench (not sure which text). They also have Cushitic and Nilotic influences in their languages, and currently neighbor the very fearsome Karamojong Nilotes, who tend to get into conflicts with all of their neighbors :P. Seriously, the Karamojong are not to be messed with, as you're probably aware:


The Karamojong are in constant conflict with their neighbors in Uganda, Sudan and Kenya due to frequent cattle raids. This could be partly due to a traditional belief that the Karamojong own all the cattle by a divine right, but also because cattle are also an important element in the negotiations for a bride and young men use the raids as a rite of passage and way of increasing their herds to gain status.

The Ongota (https://en.wikipedia.org/wiki/Ongota_language) Ethiopians are not considered Nilo-Saharan, but they are another example of a group of hunter-gatherers that are difficult to classify, ranging from divergent Cushitic/Afroasiatic (possibly with Nilo-Saharan influence), an isolate, or a creole. Genetically, they are quite distinct as well, and are predominantly mtDNA L0a which is unheard of elsewhere, with a high frequency (16%) of the rare mtDNA L6.

But I digress. Basically, I don't think the distribution of the divergent languages spoken by hunter-gatherers or close to the Ethiopian border are all that informative for the origin of Nilo-Saharans. But I am still open to the idea proto-NS may have been spoken in the eastern part of the Saharo-Sahelian area.

pgbk87
01-29-2019, 09:56 PM
Yeah, that's what's giving me pause. I thought that the prevalence of J (https://en.wikipedia.org/wiki/Y-DNA_haplogroups_in_populations_of_Sub-Saharan_Africa) in the Nubians (43.6%) and Sudanese Arabs (47.1%) might suggest Islamic era Arab gene flow, as well, but the Beja also have a high amount (38.1%). Is it reasonable to assume major Arab gene flow into the Beja, as well? Considering their proximity, it seems reasonable to assume that all of these groups got their J from the same source populations (Arabs or otherwise), unless the subclades tell a different story.

North Sudanese seem to have an extra layer of Anatolian Farmer-like ancestry compared to Somalis (for example), that may help to explain such high Y-DNA J, on top of recent Arabization and islamization(?)

As for the Beja, they seem to be an essentially Horn African population, with minor Omotic-like affinity and all. They don't differ much from Ethio-Semites or Oromos. There are actually Omotic speakers who have unusually high Y-DNA J like the Shekkacho, and that situation is more intriguing to me, especially since they seem quite genetically similar to Cushitic and Semitic Ethiopians.

Megalophias
01-30-2019, 01:48 AM
Yeah it's definitely all conjecture at this point. Honestly I don't have much faith in the reality of Nilo-Saharan as a whole, but at least East Sudanic seems promising. A link with the 'Aqualithic' and the later cattle cult pastoralists does seem likely, and maybe also the early sorghum farmers of Kassala.


The Kenyan A-M13 isolate on Yfull (A-YP4751* (https://www.yfull.com/tree/A-YP4751*/)) might be indicative of this; the subclade's TMRCA relative to all other A-M13s is as old as the hg itself (11,000 ybp) which would've originated in and around the Lake Turkana basin bordering South-Sudan and Ethiopia.
He is A-M118, which was found in Uganda and Kenya (and not in Ethiopia or Central Africa) in the Green Sahara paper - that would fit your theory, but has been 12 000 years and people have moved around a lot. We need a bunch of Sudanese A-M13 samples though, there's nothing from there AFAIK.


It's unfortunate we don't have any of their Y-DNA samples but I'd wager the Gumuz, who also speak a language controversially placed within Nilo-Saharan, would fit neatly in this "left-behind" East-African LGM category -- the Gumuz lack of L2 mtdna and negligible West-African affinity makes them a good candidate
We have 21 Gumuz from Pagani et al (2015) (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4457944/) : 57% A-M13, 33% B2b1a-M8495, 5% B2a-M150, and 5% E-Z1902 (presumably V12). Seems broadly similar to South Sudanese Nilotes. B2b must have once dominated a large swath of Africa, it's still very common in the forager groups like Pygmies, San, and Hadza. B2a apart from its role as a minority Bantu haplogroup seems hardly studied.

Lank
01-30-2019, 03:49 PM
B2a-M152/M109 seems to be the subclade associated with the Bantu expansion. B2a(xM152) is frequent in Kenyan hunter-gatherers, but also occasionally found in other East Africans, Central Africans (usually Pygmies), and a few West Africans. Perhaps M152 was picked up by early Bantu mixing with Central African natives, or more ancient Niger-Congo people mixing with Central Africans?

pgbk87
01-30-2019, 07:56 PM
I am curious as to what uniparental markers are prevelant amongst the Chabu. I'd guess a lot of E-M329 and Horn African specific mtDNA L3.

Brandon S. Pilcher
02-14-2019, 12:12 AM
I really can't wait for ancient DNA from Nubia. I intuit that the West Eurasian admixture in the region predates the Arabization of NE Africa, and that the inhabitants of Nubia and North Sudan were at one point similar to Horn Africans, just with some extra "blue" Nilotic admixture (from hypothetical Nubian language bearers) and probably "green" Egyptian-related admixture, though I suppose some of the latter could have indeed come from the Arabs.
My prediction is that most ancient Nubian ancestry is going to most closely resemble the yellow "Northeast African" component, with some "Nilotic" blue and maybe a bit of "West Eurasian" turquoise thrown into the mix. The skeletal research I have seen indicates a particularly close affinity between ancient Nubians and predynastic southern Egyptians---more so, in fact, than the latter have to younger and more northerly Egyptians. Here is a summary by two anthropologists who have studied the matter in depth:


Studies of crania from southern predynastic Egypt, from the formative period (4000-3100 B.C.), show them usually to be more similar to the crania of ancient Nubians, Kushites, Saharans, or modern groups from the Horn of Africa than to those of dynastic northern Egyptians or ancient or modern southern Europeans.
---"The Geographical Origins and Population Relationships of Early Ancient Egyptians", Egypt in Africa (1996), pp. 23-24

I agree that ancient Nubian aDNA would be exciting to have in large quantities. Given the above information, it might shed some light on how their Egyptian neighbors further down the Nile would have looked at the genetic level in ancient times as well.

beyoku
02-14-2019, 07:55 PM
B2a-M152/M109 seems to be the subclade associated with the Bantu expansion. B2a(xM152) is frequent in Kenyan hunter-gatherers, but also occasionally found in other East Africans, Central Africans (usually Pygmies), and a few West Africans. Perhaps M152 was picked up by early Bantu mixing with Central African natives, or more ancient Niger-Congo people mixing with Central Africans?

IMO B2a and many of its subclades (https://www.ncbi.nlm.nih.gov/pubmed/20213473) in the Nile Basin and Rift Valley look strongly Nilotic. Looks to picked up by Bantu from NS speakers.

Lank
02-14-2019, 08:42 PM
IMO B2a and many of its subclades (https://www.ncbi.nlm.nih.gov/pubmed/20213473) in the Nile Basin and Rift Valley look strongly Nilotic. Looks to picked up by Bantu from NS speakers.
Oops, I totally forgot about Nilo-Saharan B2a. My bad, please disregard my post. :P

Megalophias
02-14-2019, 08:55 PM
I don't really buy that the Kenyan foragers got it from Nilotes, though.

NetNomad
02-15-2019, 09:10 PM
I don't really buy that the Kenyan foragers got it from Nilotes, though.

B2a is 38,000 years old. Enough time for one or two branches to have entered Kenya pre-Nilo-Saharan (~15K YBP).