search this blog

Thursday, October 8, 2015

Ancient Ethiopian genome reveals most Africans have recent Eurasian ancestry (or not)


By recent I mean post-Neolithic, because this genome is only dated to ~4,500 BP. Admittedly though, I am skeptical that all of the Eurasian admixture arrived so late, and await more data points from prehistoric Africa. By the way, this individual belongs to Y-haplogroup E1b1 and mtDNA haplogroup L3. The main text is behind a pay wall, but the supp info is free:

Characterizing genetic diversity in Africa is a crucial step for most analyses reconstructing the evolutionary history of anatomically modern humans. However, historic migrations from Eurasia into Africa have affected many contemporary populations, confounding inferences. Here, we present a 12.5x coverage ancient genome of an Ethiopian male (‘Mota’) who lived approximately 4,500 years ago. We use this genome to demonstrate that the Eurasian backflow into Africa came from a population closely related to Early Neolithic farmers, who had colonized Europe 4,000 years earlier. The extent of this backflow was much greater than previously reported, reaching all the way to Central, West and Southern Africa, affecting even populations such as Yoruba and Mbuti, previously thought to be relatively unadmixed, who harbor 6-7% Eurasian ancestry.

Llorente et al., Ancient Ethiopian genome reveals extensive Eurasian admixture throughout the African continent, Science, Published Online October 8 2015, DOI: 10.1126/science.aad2879

Update 26/01/2016:

Erratum to Gallego Llorente et al. 2015

The results presented in the Report “Ancient Ethiopian genome reveals extensive Eurasian admixture throughout the African continent“ were affected by a bioinformatics error. A script necessary to convert the input produced by samtools v0.1.19 to be compatible with PLINK was not run when merging the ancient genome, Mota, with the contemporary populations SNP panel, leading to homozygote positions to the human reference genome being dropped as missing data (the analysis of admixture with Neanderthals and Denisovans was not affected). When those positions were included, 255,922 SNP out of 256,540 from the contemporary reference panel could be called in Mota. The conclusion of a large migration into East Africa from Western Eurasia, and more precisely from a source genetically close to the early Neolithic farmers, is not affected. However, the geographic extent of the genetic impact of this migration was overestimated: the Western Eurasian backflow mostly affected East Africa and only a few Sub-Saharan populations; the Yoruba and Mbuti do not show higher levels of Western Eurasian ancestry compared to Mota. We thank Pontus Skoglund and David Reich for letting us know about this problem.

87 comments:

Lank said...

Would you be able to get ahold of the Mota genome? Would be interesting to see a few things.

1) How it relates to other Africans
2) How it relates to "Basal Eurasian"
3) How it fits into modern variation overall (it shows a 'Eurasian' pull in the paper's PCA, but nothing in the formal stats)

Davidski said...

I'll see what I can do.

Lank said...

That's great to hear! :)

Tobus said...

@Lank: it shows a 'Eurasian' pull in the paper's PCA

I thought the main thrust of the paper was that Mota *lacks* a Eurasian pull that is seen in modern Africans, hence the conclusion that there has been Eurasian admixture into Africa since Mota's time.

truth said...

Well, looking at global PCAs one could see that West-Africans some pull towards West-Euriasians in regards to the more Paleo-Africans, so that didn't come as a surprise, however I didn't expect that for Mbutis.

apostateimpressions said...

Do they give estimates of Eurasian admixture in the rest of sub s Africa?

German Dziebel said...

Looks like "Basal Eurasian" in fact entered Africa, not came out of it. No surprise it was detected outside of Africa as early as K14.

Rob said...

great to see aDNA from Africa.,

Roy King said...

This is a cool study! If, indeed, a major Eurasian back migration into Africa occurred circa 1000 BCE, then the Neolithic source population must have avoided the ANE component that we now see in most Near Eastern populations. Could the Bedouins/Yemeni populations be the source? If so, it was likely pre-Arabic--probably South Arabian back migration. This paper also points to the importance of complete African sequences. The Illumina chip data have just too much ascertainment bias to readily detect the Near Eastern component in Mandenka or Yoruba.

andrew said...

This is technologically stunning and paradigm shifting.

Despite Mota's relatively young and reliably dated age (2500 BCE), it appears that he lived in a stone age hunter-gatherer population and is an individual with no Neolithic era Eurasian admixture, something found using him as a baseline to be present at 6% or more levels in all other Africans (with the local Ari people to whom Mota has great genetic affinity having about 15% Eurasian admixture), which is comparable, for example, to Turkish admixture in Anatolia.

Presumably, though, this individual could not be used to distinguish ur-African genes from those that arrived from Eurasia via back migration in the Upper Paleolithic era, for example (e.g. waves that are presumed to have brought mtDNA M1 and U6 to Africa).

His isolation at this late date is less surprising when one considers that he lived in the highlands of an endorheic basin well inland in Africa, and not a natural destination for early pastoralists and farmers working their way up river basins. Even today the region is a 90% rural enclave of the past. The timing is also a few hundred years before archaeological and genetic evidence suggests that food producing migrants arrived in Ethiopia from South Arabia (giving rise to the Ethio-Semitic language) and from the Blue Nile.

All of the Eurasian admixture identified in relation to the baseline of unadmixed Mota in sub-Saharan Africa is basically EEF and carries proportionate levels of Neanderthal admixture (which Mota lacks).

As an aside, the paper also suggests rather strongly that the Sandawe people of Tanzania (who speak a click language isolate) and the linguistically Omotic Ari people of the Omo Valley (the most diverged of the Afro-Asiatic languages) share a strong genetic affinity with each other that is distinct from the click language speaking Hadza people, the Nilotic people, and other Afro-Asiatic populations.

German Dziebel said...

@Roy King

"If, indeed, a major Eurasian back migration into Africa occurred circa 1000 BCE, then the Neolithic source population must have avoided the ANE component that we now see in most Near Eastern populations."

All of the SSAfrican populations, including San, are shifted toward Ma'ta (prototypical ANE in Eurasia) vs. K14.
http://anthropogenesis.kinshipstudies.org/2014/11/ancient-kostenki-14-markina-gora-dna-a-glimpse-into-a-population-on-its-way-from-america-to-africa

Karl_K said...

"Admittedly though, I am skeptical that all of the Eurasian admixture arrived so late, and await more data points from prehistoric Africa."

Definitely. The Berbers probably arrived much earlier than this. I'm sure a 10,000 year old African genome will show additional Eurasian back-flow into modern African populations.

Nirjhar007 said...

Cool, African aDNA, yay!! :D...

Tobus said...

@German:
Looks like "Basal Eurasian" in fact entered Africa

This isn't "Basal Eurasian", it's post-Neolithic West Eurasian.

All of the SSAfrican populations, including San, are shifted toward Ma'ta (prototypical ANE in Eurasia) vs. K14.

Check the Z-scores in Tables S10/S11 (of the original Seguin-Orlando K14 paper), the San/Mbuti scores aren't significant indicating these these pops are outgroups to the MA-1/Kostenki branch. The instances where the SSA score approaches significance are the populations identified as having post-Neolithic Eurasian admixture here (Yoruba, Mandenka etc.), and all of those show higher scores using Loschbour/Stuttgart/Otzi etc. in place of MA-1... it's not ANE that's causing this.

andrew said...

@Roy King

Pickrell (2014) http://www.pnas.org/content/111/7/2632.abstract shows as admixture of a Levatine-like (not South Arabian-like) population ca. 1300 BCE with an Omotic-like population as the source of Eurasian admixture in the Khoisan and many non-Ethio-Semitic Ethiopians. I am inclined to think that a realistic date is more like 1700 BCE-2000 BCE, but the big picture is pretty consistent.

This is also an area where agriculture arrives rather late. http://dispatchesfromturtleisland.blogspot.com/2014/02/sources-of-west-eurasian-ancestry-in.html

At long as the EEF types start their journey from W. Anatolia towards Ethiopia by ca. 5000 BCE- 4500 BCE, a roughly 1200 mile trip, it isn't unrealistic for them to arrive in the right place at the right time. The fact that the Omo Valley is an endorheic basin also probably contributes to its isolation and late contact with the Neolithic since early farmers and herders who migrate up river basins will never end up there on purpose. But, if you are going to make it from the Sahel range to points south, without having to brave of jungle, it is the best route to take.

Krefter said...

mtDNA/Y DNA doesn't connect East Africans and Neolithic Europeans(EEF) at all. They have no connections besides sharing generic ancient Near Eastern lineages. East African's Near Eastern Y DNA/mtDNA connects them to SW Asians. So, I'm very skeptical of their Near Eastern ancestors being like EEF.

German Dziebel said...

@Tobus

"This isn't "Basal Eurasian", it's post-Neolithic West Eurasian."

In SI, Sardinian keeps coming up as the likeliest source for West Eurasian admixture in SSAfrica (including Mbuti), hence my earlier hunch. Basal Eurasian is attested in an ancient Eurasian sample (K14), which is still closer to Lithuanians than a later but BE-richer Stuttgart. So I'm just connecting the dots. BE must be a subset of West Eurasian diversity exported into Africa, not a subset of African diversity imported into Neolithic Europe.

"Check the Z-scores in Tables S10/S11 (of the original Seguin-Orlando K14 paper), the San/Mbuti scores aren't significant indicating these these pops are outgroups to the MA-1/Kostenki branch."

Look at the plot I included in the referenced post. When K14 and MA-1 are directly compared, all SSA, including San, shift toward MA-1.

Tobus said...

@German:
BE must be a subset of West Eurasian diversity exported into Africa, not a subset of African diversity imported into Neolithic Europe

Yes, that's why it's called Basal *Eurasian* in the first place no? My understanding is it branched off after OOA (in Eurasia!) but before the West/East Eurasian divergence.

Look at the plot I included in the referenced post.

I did, and immediately noticed that the SSA SE ranges cross the 0 mark, so I checked the Z-scores to see how significant they were... and they weren't.

Moreover they SSA scores increase when using Loschbour instead of MA-1, so I don't think it's ANE that's causing the "shift".

Karl_K said...

Clearly Tobus and German both agree that Basal Eurasian was imported into Africa from Eurasia and is not just a slow leak of alleles out of Africa, as some people still believe.

Right?

Aram said...

The haplogroup T can be involved into this back migration to Africa.

Karl_K said...

@Krefter

"I'm very skeptical of their Near Eastern ancestors being like EEF."

And yet from the data, they clearly were very like EEF (Sardinians and LBK) So this must be a case where the haplogroups are missing the real story.

Krefter said...

@Karl_L

Maybe they were Bedouin without African ancestry.

Karl_K said...

@Krefter

Maybe, but what haplogroups do Bedouin have, and when did they get them? Does that make the situation more clear?

Tobus said...

@Karl_K:Clearly Tobus and German both agree that Basal Eurasian was imported into Africa from Eurasia and is not just a slow leak of alleles out of Africa

I think the data clearly points in that direction, so yes, I'm assuming that's the case until someone can prove it wrong.

mickeydodds1 said...

So, just to provoke another long standing bone of contention involving y haploytpes, and seeing that German Dziebel is on board, did y haplotype 'E' originate in Africa or Eurasia?

Karl_K said...

@Tobus

Agreed. Several additional ancient African genomes, or much better, an actual Basal Eurasian genome would make the case 100%.

I think the Neanderthal % in modern people makes a very strong case anyway.

If Basal Eurasian resulted from a generic Eurasian (like Ust Ishim) with eventual 20% African admixture, then this Basal Eurasian would have ~20% less Neanderthal admixture than other Eurasians. But this is not the case.

Shaikorth said...

This sample's total lack of Neanderthal/Denisovan even in comparison to modern Africans makes a strong case for him being unmixed. Will likely be a good reference for estimating African affinities in Eurasia by f4 or d-stats.

Krefter said...

@Karl_K,
"Maybe, but what haplogroups do Bedouin have, and when did they get them? Does that make the situation more clear?"

See my blog....

http://mtdnaatlas.blogspot.com/

I don't have Bedouin mtDNA but I do have Arabian mtDNA. J1b, J1d, T1a, T2c1, T2e, N1b1, N1a3a, N1a1a, U3, H, and R0a take up some 80% of West Eurasian mtDNA in Arabia. The only stuff they share with EEF is T2e, T2c1, N1a1a(often of differnt variety than EEF N1a1a), and U3. I'd say 80% of Arabian West Eurasian mtDNA split from EEF mtDNA in West Asia over 8,000 years ago.

AFAIK, East Africans share with Arabians not EEF mtDNA wise.

Karl_K said...

@Shaikorth

"This sample's total lack of Neanderthal/Denisovan even in comparison to modern Africans makes a strong case for him being unmixed. Will likely be a good reference for estimating African affinities in Eurasia by f4 or d-stats."

Yes. But the final level of Neanderthal was determined by using this genome, so... of course he had a total lack of Neanderthal as a result.

It is probable that earlier back-migrations from Eurasia occurred. Without additional ancient genomes, it might be impossible to tell to what extent this impacted what we now consider "unmixed" people, such as this ancient Ethiopian.

This is, of course, heavily intertwined with the constant discussion about the location of the origin of Y-haplogroup E.

If that haplogroup actually originated in an out-of-Africa population, then a back-migration must have had a massive impact on the genetics of Africa. This could have "smoothed out" a lot of the variation between all modern human populations.

In this case, most people would have more Neanderthal DNA than we think at this point, and it would push our most recent common ancestor further back in time.

There is really no way to know at this point. So I really hope they get to sequencing some more ancient Africans.

MfA said...

Karl_K said...
I'm sure a 10,000 year old African genome will show additional Eurasian back-flow into modern African populations.


Indeed, I think there is still some room for additional Neanderthal ancestry for contemporary populations, with using older African ancient DNA.

Chris Davies said...

"Admittedly though, I am skeptical that all of the Eurasian admixture arrived so late"

I agree. I think Maju is correct. Many of the F,G,H,I,J,K Y DNA haplogroups found in Sudan could be the result of an Eurasian back-migration as early as the Palaeolithic. And then spread to other African populations. Even the Mbuti have haplogroup J.

Grey said...

Is it really "Out of Africa" or "Out of the Tropics" or "Out of SSA"?

If the latter then couldn't "Eurasia" in this context include north Africa?

Grey said...

should be

"Out of the Tropics" / "Out of SSA"

not or.

Shaikorth said...

"Many of the F,G,H,I,J,K Y DNA haplogroups found in Sudan could be the result of an Eurasian back-migration as early as the Palaeolithic."

That question will be solved by next gen Y-sequencing soon enough, as will the issue of whether E is more archaic in Africa or Near East.

Kurti said...

Another strong argument for the Back to Africa migration of Haplogroup E?

German Dziebel said...

@Tobus

"Yes, that's why it's called Basal *Eurasian* in the first place no? My understanding is it branched off after OOA (in Eurasia!) but before the West/East Eurasian divergence."

I wrote "a subset of WEST Eurasian" diversity (K14 has BE, MA-1 doesn't), not an early split from an ancestral African population that predates the split into East and West Eurasians.

@Karl K

"Clearly Tobus and German both agree that Basal Eurasian was imported into Africa from Eurasia and is not just a slow leak of alleles out of Africa, as some people still believe.

Right?"

yes, absolutely. So, "Basal Eurasian" is a misnomer.

@Tobus

"Moreover they SSA scores increase when using Loschbour instead of MA-1, so I don't think it's ANE that's causing the "shift"."

Note a telling polarity elsewhere in the plots. Loschbour "lifts" Khoisans as well as Sardinians/French (against K14), while MA-1 "lifts" Amerindians. It means there must have been a continuous West Eurasian impact on SSAfrica (most ancient MA-1-like, then Loschbour-like, then Sardinian/Stuttgart-like, according to the Mota study). Sometimes they compounded, hence the greater leftward pull for Khoisan (compared to MA-1) when Loschbour is in the picture. But when two Paleolithjic genomes (MA-1 and K-14) are directly compared, Khoisans are closer to MA-1. MA-1 is really "Basal Human." I agree that I wish those trends were all perfectly statistically significant, but, as everywhere in human origins research, a little bit of data is better than no data.

Kurti said...

AH not I see this individual was already yDNA E.

But than I don't think all back to Africa migration started 2500 BC.

Alone the Afro_Asiatic expansion in North Africa dates further back. I still suspect with pöder samples we will find more Eurasian admixture in Africa. And E being in reality a back migration.

Karl_K said...

@Kurti

Definitely. Tunisian Berbers have the highest "North African" genetics and almost no Sub-Saharan-African admixture. They are almost all haplogroup E, clearly Eurasian in ancestry, and have been in Africa for 12,000 years or so.

I think that's a pretty clear indication of haplogroup E arriving in Africa with pre-Neolithic back-migrations.

And those people must have brought some great new technology to have expanded so widely on the continent.

German Dziebel said...

@Shaikorth

"This sample's total lack of Neanderthal/Denisovan even in comparison to modern Africans makes a strong case for him being unmixed."

Or admixed with "archaic Africans" instead.

Tobus said...

@German:
I wrote "a subset of WEST Eurasian" diversity (K14 has BE, MA-1 doesn't), not an early split from an ancestral African population that predates the split into East and West Eurasians.

I'm not sure what you're trying to get at German, the two aren't mutually exclusive. BE predates the East/West split *and* is a subset of West Eurasian diversity via admixture into EEF. Do you need to review the Laz. (2014) model?

Loschbour "lifts" Khoisans as well as Sardinians/French (against K14), while MA-1 "lifts" Amerindians. It means there must have been a continuous West Eurasian impact on SSAfrica (most ancient MA-1-like, ...)

Surely it just means that something more Loschbour-like had an impact on SSA, Sardinians and French while something more MA1-like had an impact on Amerindians.

But when two Paleolithjic genomes (MA-1 and K-14) are directly compared, Khoisans are closer to MA-1

No, and please make an effort to understand this, the Khoisan SE surrounds the 0-mark and has a non-significant Z-score (-0.8). This means that when MA-1 and K-14 are directly compared, Khoisans are *NOT CLOSER* to either. This is basic D-stats 101, go back and confirm it with Patterson (2012) if you need to.

I wish those trends were all perfectly statistically significant

The lack of significance is an important data point in and of itself, and wish as you may, you can't just act like it does have significance when it suits you. Any model you propose needs to work with Mbuti/Khoisan forming a clade with respect to MA1/K14.

German Dziebel said...

@Tobus

"I'm not sure what you're trying to get at German, the two aren't mutually exclusive. BE predates the East/West split *and* is a subset of West Eurasian diversity via admixture into EEF."

No, BE is a subset of West Eurasian diversity only. Let's go past Lazaridis et al. K14 has BE and it's still closer to Lithuanians than it is to Stuttgart.

"No, and please make an effort to understand this, the Khoisan SE surrounds the 0-mark and has a non-significant Z-score (-0.8). This means that when MA-1 and K-14 are directly compared, Khoisans are *NOT CLOSER* to either. This is basic D-stats 101, go back and confirm it with Patterson (2012) if you need to."

Your usual nonsense. The signal is just weaker for MA-1 in San than it is for Loschbour but it forms a single, continuous 0+ pattern. That's the objective reality. If part of this pattern is "significant", then the other part of is significant, too. If Patterson's stats can't capture this significance from the strongest signal to a weaker one, it's the problem of Patterson's stats, not of the data. They just need to lower their threshold, maybe.

"Any model you propose needs to work with Mbuti/Khoisan forming a clade with respect to MA1/K14."

It works just fine. The Mbuti/Khoisan clade comes from extinct archaic Africans. The MA-1 pull comes from modern humans who absorbed those archaic Africans.

Maju said...

I'm going through all the comments because, admittedly, I'm quite perplex and suspect confounding factors being at play, for example outgroup false pull caused merely by the sheer ancient diversity of the African genetic landscape, which is hard to simplify into any single population.

I find that Andrew has made interesting comments that I tentatively subscribe, for instance:

"Presumably, though, this individual could not be used to distinguish ur-African genes from those that arrived from Eurasia via back migration in the Upper Paleolithic era, for example (e.g. waves that are presumed to have brought mtDNA M1 and U6 to Africa)".

"This is also an area where agriculture arrives rather late".

Another issue that causes my perplexity is that, right after Sardinians, the most likely candidate for this alleged Eurasian admixture are Belorussians and Lithuanians, who typically are at the opposite polarity of the European diversity, while Basques and Russians score very low, in spite of their similitudes to either population. Ancient European samples are rather bad matches both (Stuttgart less than Lochsbour but neither is good enough), at least when the outgroup is Mbuti (Yoruba does seem to have at least minor Eurasian admixture and hence is worse outgroup than the Mbuti).

A key detail may be the Neanderthal admixture table: while the Yoruba estimates are slightly larger than the error margin, the Mbuti ones are clearly much smaller and hence statistically insignificant. This probably indicates that the alleged Eurasian admixture in Mbuti is an artifact while the one in Yoruba may well be largely an artifact but still partly real.

More research is needed.

Jared Knows said...

@Kurti

Eh, I don't see how this would be much evidence on the side of E being Eurasian. This genome is fairly recent, but it also lacks (/lacks/) the Eurasian-archaic signature in modern Africans. This at the same time as being E1b1 and L3x2a. Northwest Africans are also hardly hardly SSA. E in Berbers shows a clear founder effect, the majority of the diversity of E still lies with people that are predominately autosomally SSA. Some of E in North Africa is probably due in part to Afro-Asiatic as well as wetter Sahara.

German Dziebel said...

@Karl_K

I'm now not sure Tobus is on the same page with me regarding BE. I think he believes BE is both an early OOA offshoot predating any Eurasian splits AND a West Eurasian component that back migrated into Africa. From my perspective, this is just bad logic.

Karl_K said...

@Maju

This is not an artifact. It was already strongly shown from previous studies where they "masked" the Eurasian components. Only with those, they didn't have a genuine baseline, and assumed that Mbuti and Yoruba had 0% Eurasian admixture.

In this case we have an actual pre-admixture genome, and several modern post-admixture genomes from the exact same region that are the closest matches to it. So the extraction of the admixed alleles is trivial.

The things that perplex you are likely the result of not having an exact match to the Eurasian component that admixed. It was mostly like Sardinians, and they couldn't find a better sequenced population. It could be like Bedouin without SSA and with some extra something.

But in any case, the results are real because they see the same alleles in all kinds of Africans (and this was anticipated by independent research).

The low levels of confidence on the Neanderthal are because of the low levels of Eurasian. 3% of 6% and only a small number of Neanderthal specific SNPs anyway. So the statistics seem bad, but not for the EEF-like part (which would have brought the Neanderthal part that they are checking for).

Karl_K said...

@German

I agree with Tobus. Sorry.

I think I understand your logic and your point of view. Correct me if I am wrong.

You think that "modern" humans originated outside of Africa. And within Africa there was a related "archaic" group of humans. How divergent do you suspect they were? Did they carry the most deeply branching African haplogroups that still exist?

But then there were repeated into-Africa migrations that mixed with these archaics until they were finally brought up to "modern" human levels of allele frequency through admixture?

And you think we just can't see this because we don't have a genome from an African archaic?

Is that right?

I mainly don't buy it because the mtDNA and Y-haplogroups look to be a reasonable tree. And those agree with the out-of-Africa with repeated back migration model.

I can see a split at 300,000 years between Southern Africa and Northern Africa/'Eurasian' that then was smoothed over by later admixture between these major groups. But that is not a huge change really.

Karl_K said...

"I can see a split at 300,000 years between Southern Africa and Northern Africa/'Eurasian' that then was smoothed over by later admixture between these major groups. But that is not a huge change really."

I meant to say that I could hypothetically imagine this as being a remote possibility. Not that I believe that it is true at all.

Karl_K said...

@Jared

"Eh, I don't see how this would be much evidence on the side of E being Eurasian. This genome is fairly recent, but it also lacks (/lacks/) the Eurasian-archaic signature in modern Africans."

It only (/lacks/) that component because this genome is used as the zero admixture control in that analysis.

All of the alleles that this man had inherited from Eurasian backflow among his own ancestors were assumed in the study to be totally African and not Eurasian and not Neanderthal and not Denisovan.

The supplementary materials show this effect clearly.

This is the best we have right now. But it probably is not the whole story.

German Dziebel said...

@Karl_K

So BE is an ancient African component that stayed in isolation in Eurasia for tens of thousands of years, then got repackaged as part of a Sardinian-like population and migrated back into Africa? That's not sound reasoning.

But you basically got my model right. And yes, those "archaic" Africans carried the most divergent lineages found in Africa. The reason they are not found outside of Africa (but are easily found in America since 1492 times because that was a real out-of-Africa migration) is precisely because they were absorbed within Africa into an incoming Eurasian population.

"I mainly don't buy it because the mtDNA and Y-haplogroups look to be a reasonable tree."

You can always build a tree out of anything. But a simple-looking gene tree conceals a complex population history of a highly mobile new species that came on top several extinct ones. Y-DNA E is a subset of Eurasian CT and it's pan-African. A, A00, B are not found outside of Africa along the putative routes of out-of-Africa migration or in ancient remains outside of Africa. MtDNA shows a similar picture. Oase is an N-like lineage but not L3d or L3e or any of the African lineages. A simple out-of-Africa explanation for the trees requires an assumption of a bottleneck that spared recent Eurasian-only lineages but killed ancient African-specific ones. It's like saying that all the necessary evidence required to prove OOA got lost in a bottleneck. I can't buy that. I can poke holes in the phylogeny itself and the methodology behind it but the easiest solution is to explain divergent African lineages as archaic heritage and postulate an ancient migration out of Africa followed by a modern human speciation event outside of Africa (way outside of Africa) followed by back migrations of Eurasians into Africa from around 50-40,000 years ago down to Neolithic times.

BTW, if you re-read Cann et al. 1987 on mtDNA you'll see that the criterion they used to look for archaic admixture outside of Africa was "lineage divergence." They didn't find mighty divergent lineages outside of Africa, hence they concluded that there was an out of Africa migration without introgression. But the African genetic landscape contains precisely such divergent lineages and this fact favors an archaic introgression in Africa model on the basis of Cann et al.'s original criteria.

Karl_K said...

@German

I respect your dedication to your hypothesis. But, I just do not see it as the most likely story.

I am curious, though. Where do you think the root of the "modern" humans lies? In which location? Which modern humans are best represented by this root without archaic admixture? When and where and how did the original split between non-africans and africans occur?

Chris Davies said...

How about Sardinians plus EEF/LBK *and* SSA all possess Sudanese-like admixture [itself being a mixture of African and Eurasian]. We know that Treemix shows a migration edge from Dinka into Sardinia and Stuttgart.

German Dziebel said...

@Karl_K

"I respect your dedication to your hypothesis. But, I just do not see it as the most likely story."

But it's the most logical one. Likelihood comes from people, logic from nature. :)

"I am curious, though. Where do you think the root of the "modern" humans lies? In which location?"

I follow the distribution of linguistic diversity here because language is a good proxy for behavioral modernity and it's not subject to the confounding effects of "archaic admixture." As measured by the number of phylolinguistic units (stocks, families), the greatest diversity is in the New World and in the Sahul. (Africa is rather homogenous in this regard, which is consistent with its being peopled rather late by modern humans.) These are also the areas devoid of "archaic hominids". And what's noteworthy is that it's precisely the New World and the Sahul that show the strongest Neandertal and Denisovan signals (both in terms of the actual allele matches and in terms of the high-Fst population structure). Since Neandertals and Denisovans were not present in the New World and the Sahul but Amerindians and Papuans show the strongest genetic signal related to those populations, it must be not admixture, but common descent coupled with strong genetic drift.
But
So my hypothesis is that we are of a East Eurasian hominid stock and that we re-colonized Africa some 50-40,000 years ago. We lost most (but not all!) of our ancestral East Eurasian hominid diversity and we replenished it (mostly in Africa) by mixing with local African hominins.

I mean it's a complex model but I don't care. Human migrations have always been a mess. But it's the most logical one. Plus it explains both linguistic and genetic data.

andrew said...

"Karl_K said...
I'm sure a 10,000 year old African genome will show additional Eurasian back-flow into modern African populations."

Doubtful. I think it is quite reasonable to infer that Mota, due to his geographically isolated hunter-gatherer context, didn't receive any Holocene era Eurasian admixture. There may have been other populations in Africa that received Eurasian admixture earlier than the Omo Valley in Ehtiopia, and we have evidence from prior studies that Eurasian admixture came to the Khoisan later than it did to the Omo Valley.

But, if we want to capture Eurasian backflow that Mota's ancestors experienced but ancient DNA from an African even earlier than Mota did not, I think you are going to need to find African ancient DNA from 20,000-35,000 years ago or more. With the breakthrough of sequencing Mota's ancient African DNA, and successes in sequencing ancient DNA from Denisovans, Neanderthals, Upper Paleolithic and modern humans Siberians, this no longer seems impossible. But, it would take a lot of luck.

The M1/U6 backmigration is estimated to have started ca. 30kya from Iberia to NW Africa (where M1 is now 3.2% of the mtDNA pool), and then to have made its was to East Africa at a later date. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1945034/ The frequency of M1 peaks at about 15% in Ethiopia and 10% in East Africa generally, and U6 adds something to the total, although M1 and U6 don't peak in the same geographic areas so it doesn't add much. Still, a back migration that had a 15%-20% impact on the maternal mtDNA pool of East Africa is huge. Surely, the back migration was not entirely secured through bride exchange between Iberia and NW Africa. But, our best evidence is that we'd expect European men from 30kya to be mostly Y-DNA I and F*, both of which are vanishing rare in places where mtDNA M1 and U6 are found, so we have a puzzle. One way to solve that would be that Upper Paleolithic men with Y-DNA E men advanced into Iberia for a time, married local mtDNA M1 and U6 women, and then mostly retreated back to Africa and spread from their, leaving some residual Y-DNA E in SW Europe.

""This sample's total lack of Neanderthal/Denisovan even in comparison to modern Africans makes a strong case for him being unmixed." Or admixed with "archaic Africans" instead."

Statistical analysis of modern whole genomes from Africa, which previously detected suspected archaic admixture in Eurasians including enhanced archaic admixture in populations we now know to have had Denisovan admixture, have found evidence for two such archaic African populations admixing with the ancestors of modern African populations possibly as recently as 10,000 years ago. IIRC, the stronger event was from a population near the southern edge of the jungle that the Congo River runs through. The weaker event was from a population near the West African Bantu homeland in Southern Nigeria. The weaker event, however, seems to be buttressed by a handful of West African men with a version of Y-DNA A so basal that it should predate ordinary maximally basal Y-DNA A by a hundred thousand years or more. There have been no hints, however, of archaic African admixture in East Africa.

German Dziebel said...

"There have been no hints, however, of archaic African admixture in East Africa."

So? The whole angle is rather new. Read (http://www.pnas.org/content/108/37/15123.full) and you'll see that the authors speak of "archaic admixture" in Africa in pan-African terms, although we obviously need more and more regional samples to test it. In the absence of very ancient African DNA, it will all remain speculative, though. All Africans (as compared to non-Africans) are pulled toward chimps in whole-genome PCA plots, which is unexpected if all humans stemmed from the same African pool but perfectly natural if admixture with archaic Africans occurred in Africa after Africa had been colonized from Eurasia. One other "hint" I can give you is that East Africa is rather rich in Y-DNA A1b1b-M32 (formerly A3) (up to 65% in some populations) and hg A's coalescence ages exceed by an order of magnitude any dates obtained by archaeologists from behaviorally modern toolkits globally.

Paradoxically, though, Mota shows hg E, which is the only African haplogroup with clear phylogenetic affinities outside of Africa, but autosomally it shows up as unmixed with Eurasians (or potentially archaic African admixed).

Bronze said...

@andrew

"Still, a back migration that had a 15%-20% impact on the maternal mtDNA pool of East Africa is huge. Surely, the back migration was not entirely secured through bride exchange between Iberia and NW Africa. But, our best evidence is that we'd expect European men from 30kya to be mostly Y-DNA I and F*, both of which are vanishing rare in places where mtDNA M1 and U6 are found, so we have a puzzle. One way to solve that would be that Upper Paleolithic men with Y-DNA E men advanced into Iberia for a time, married local mtDNA M1 and U6 women, and then mostly retreated back to Africa and spread from their, leaving some residual Y-DNA E in SW Europe."

This whole paragraph you wrote is definitely without question wrong, there is a lot more euroasian Y-dna/paternal dna in sub saharan Africa compared to euroasian mtdna. E is likely euroasian in origin, arriving very early in eastern and northern Africa, with the euroasian men being superior to the local SSA african men, the euroasian men outcompeted them and mixed extensively with local african women.

capra internetensis said...

I really shouldn't feed the troll, but oh well

@German

"One other "hint" I can give you is that East Africa is rather rich in Y-DNA A1b1b-M32 (formerly A3) (up to 65% in some populations) and hg A's coalescence ages exceed by an order of magnitude any dates obtained by archaeologists from behaviorally modern toolkits globally."

You are conflating A1b with all A. The TMRCA of A1b1 and A1b2 (BCDEF) is much younger than that of all A.

"All Africans (as compared to non-Africans) are pulled toward chimps in whole-genome PCA plots, which is unexpected if all humans stemmed from the same African pool but perfectly natural if admixture with archaic Africans occurred in Africa after Africa had been colonized from Eurasia."

It's perfectly natural, is it? Which PCA? And by what mechanism would this occur?

German Dziebel said...

@Capra

Oh, another pseudoscientist's showing...

"You are conflating A1b with all A. The TMRCA of A1b1 and A1b2 (BCDEF) is much younger than that of all A."

Yes. And your point is?

"Which PCA?"

E..g, S9: http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1004353#pgen.1004353.s010

"And by what mechanism would this occur?"

By the same one as you think the Neandertal/Denisovan "admixtures" occurred.

capra internetensis said...

@German

Is the TMRCA of A1b an order of magnitude older than the oldest "behaviourally modern toolkits"?

On that PCA modern humans are distributed in a small patch in the centre of a triangle which has Neanderthal, Denisovan, and Chimp at its corners. Eurasians are pulled toward the Neanderthal point of the triangle, leaving Africans further toward the base formed by the line between Denisovan and Chimp. They aren't pulled toward Chimp at all, the shape of the PCA is simply due to Neanderthal admixture in Eurasians.

Man, I thought you'd at least find something tricky with ascertainment bias. Not up to your usual standards of sophistry.

German Dziebel said...

@Capra

"On that PCA modern humans are distributed in a small patch in the centre of a triangle which has Neanderthal, Denisovan, and Chimp at its corners. Eurasians are pulled toward the Neanderthal point of the triangle, leaving Africans further toward the base formed by the line between Denisovan and Chimp. They aren't pulled toward Chimp at all, the shape of the PCA is simply due to Neanderthal admixture in Eurasians."

Who's the sophist? On PC1, Europeans are pulled toward Neandertals, Oceanians toward Denisovans and Africans toward...chimps. Plain simple.

"Is the TMRCA of A1b an order of magnitude older than the oldest "behaviourally modern toolkits"?"

You tell me... What TMRCA did you assign to it? Under the assumption of common descent and not admixture in Africa...

capra internetensis said...

@German

You do realize that anyone can look at the PCA you've linked and see that you are lying?

How would "admixture in Africa" result in a different relationship of non-recombining Y chromosome lineages than "common descent"?

German Dziebel said...

@Capra

"You do realize that anyone can look at the PCA you've linked and see that you are lying?"

I have made an effort to actually show what I'm talking about. Unlike your baseless trolling of this string. And now you are saying that I'm lying?! Without actually showing where exactly my "lie" is? Instead, you are throwing around meaningless phrases such as "ascertainment bias," "Neandertals skew the plot", "base formed by a line separating Denisovans and chimps," etc.

In reality, the plot contrasts Neandertals and chimps on PC1 and Denisovans and Neandertals on PC2. Africans are shifted toward chimps (PC1), all of Eurasians and Near Easterners toward Neandertals (PC1) and Oceanians and Africans additionally (PC2) toward Denisovans. Africans' location in the top left corner of the enlarged plot shows precisely "admixture" with an African hominid relative of Denisovans that must have stayed in Africa and hence retained more of that primary pull toward chimps as well.

"How would "admixture in Africa" result in a different relationship of non-recombining Y chromosome lineages than "common descent"?"

Admixture with an extinct population implies that some lineages must have been lost. This would affect any estimates of divergence between what's left. The divergence of A00 from A, A from B and B from CT exceeds any divergence within CT, but pan-African hg E belongs with CT, so A00, A and B are likely introgressions, with large gaps between lineages caused by lineage extinction in Africa.

capra internetensis said...

@German

PC1 distinguishes both Neanderthals *and* Denisovans from chimps (with somewhat more loading on Neanderthals). *All* modern humans are well chimpward of both archaic humans. PC2 distinguishes Neanderthals from Denisovans. All Eurasians are shifted toward Neanderthals. Oceanians are shifted toward Denisovans as well (which puts them furthest out on the archaic human axis). Africans are simply less shifted toward Eurasian archaics.

Africans mixed with archaic relatives of Denisovans which pulls them toward both chimps and Denisovans relative to Eurasians in just such a way that it *looks* like Eurasians are admixed with Neanderthals, but really Eurasians have some sort of shared ancestry with Neanderthals which magically appears in larger linkage blocks in older genomes. AKA ad hoc bullshit.

Or... Eurasians mixed with Neanderthals, and this admixture follows the usual rules of linkage disequilibrium decay. Parsimony!

The estimated TMRCAs of NRY lineages are not affected by missing extinct lineages. Why would they be?

German Dziebel said...

@Capra

"*All* modern humans are well chimpward of both archaic humans."

This doesn't make sense. What instead is going on is that ALL humans (including Africans) are almost twice as close to Neandertals and Denisovans than they are to chimps. Africans are shifted further toward chimps in PC1.

"Oceanians are shifted toward Denisovans as well"

More accurately: Oceanians are shifted more toward BOTH Denisovans and Neandertals. Some Amerindians (less clear on the plot, but ascertained elsewhere), too.

" but really Eurasians have some sort of shared ancestry with Neanderthals which magically appears in larger linkage blocks in older genomes. AKA ad hoc bullshit. "

I can hardly follow your old genome bashing here. In reality, Amerindians and Oceanians share larger chunks with Neandertals than Eurasians. http://anthropogenesis.kinshipstudies.org/2014/10/ancient-ust-ishim-dna-as-seen-from-the-americas/.

"Eurasians mixed with Neanderthals, and this admixture follows the usual rules of linkage disequilibrium decay.'

Sure thing but the decay is from America and Oceania through Eurasia into Africa, and not the other way.

"The estimated TMRCAs of NRY lineages are not affected by missing extinct lineages. Why would they be?"

Duh. If A00 went extinct, our Y-DNA tree would have been younger by 100,000 years. Would it not?

capra internetensis said...

@German

So your interpretation of that PCA is that modern humans *literally* have chimp admixture that archaic humans lack?

If A00 went extinct that would make the tree younger indeed, but it would not affect "estimates of divergence between what's left".

Open Genomes said...

Mota Ethiopia 4524-4418 Cal BP is E1b1a2-M329

E1b1a2 is the earliest and most divergent branch of E1b1a-M2, which is modal in Sub-Saharan Africa.

Mota shares the most drift with the Ari Cultivators (one, in particular) who speak an Omotic language, and Omotic is the most divergent member of the Afroasiatic language family. It's clear that Mota is a native Ethiopian Highlander, because he has three SNPs that are high-altitude adaptations found among Ethiopian Highlanders today. Is this conclusive evidence that Afroasiatic languages - including Semitic, originated in the Ethiopian Highlands?

Afroasiatic Urheimat - Wikipedia

Open Genomes said...

If the Afroasiatic Urheimat was in the Ethiopian Highlands, and the Proto-Afroasiatic people did *not* have any Eurasian admixture, then it seems very likely that the Natufians of the Levant, 14,700 ybp - 12,800 ybp were *not* Afroasiatic (i.e. Proto-Semitic) speakers, as has been previously speculated.
Natufian Culture - Wikipedia

Virtually no branch of E1b1b1-M35, typically associated with Afroasiatic speakers, has a clade old enough to have expanded in the earliest Pre-Pottery Neolithic A (PPNA). There are a few isolated early E-M84 subclades in the Levant, found mostly among Palestinian Christians, which could be the remnants of the Natufians, but these didn't expand with the Early Pre-Pottery Neolithic like the main subclades of haplogroup G2a-P15 did.
YFull.com tMRCAs for E-M84

Check the dates in YFull for E-84. E-V22 is no older than 8200 ybp, and its sister clade in E-Z1919, E-V13, common in Europe, is no older than 4200 ybp. E-Z1919 cannot be Natufian either.

Matt said...

Looking at the last two posts' comments. Ah. Hopefully Eurogenes blog comments will not become as pointlessly "Germanized" as did those at Dienekes'.

Re; main topic I feel pretty skeptical of the recency here, as well. Would we really not have an ADMIXTURE component have emerged, at the levels of recency we are talking about? Nor a signal of sharing on derived EEF variants (e.g. SLC24A5 variants)? Mota may have been one of an isolated population in Africa, rather than postdating a 2000BC migration of West Eurasians to Africa (and anyway, were there actually EEF like pops around then to move into Africa? And per the supplement in tables S6 thorough S7 the superiority of EEF as an admixing source over say, Belarusians and Lithuanians actually seems quite thin.).

I'm not totally sure about the Neanderthal stuff, exactly. Do we have D(Neanderthal,Denisovan;Pop,Chimp) stats for Africans? Those should show clearly if there is an affinity to either Neanderthal or Denisovan in Africans and so indicate Neanderthal ancestry (in absence of Denisovan contribution), I think.

Following are some outgroup stats with African populations, ancient Eurasians and the Chimp outgroup, as previously calculated by Davidski:

http://i.imgur.com/a9jnY0O.png

They were pretty much all Z<3, so I think we didn't make much of them earlier, as not statistically significant, perhaps they need some degree of reinterpretation.

LBK is mildly preferred to the other ancient Eurasians, while it also seems that Loschbour is preferred to MA1 to a similar degree.

OTOH, EHG is apparently not preferred to Loschbour in all of the populations and most are neutral. That seems to be slightly unusual if EHG=MA1+Loschbour. These are low stats though. If correct though it would indicate a population with some EEF like qualities, also rather agnostic at the same time to EHG vs WHG.

http://i.imgur.com/Yi8URIc.png

This shows a D(Papuan,Chimp;LBK_EN,Pop) stat plotted against various of the stats from the above. The D(Papuan,Chimp;LBK_EN,Pop) sort of functions here as an index of generalised "Eurasianness" as defined by Papuans (most distant from Africans, phylogenically?). There's a lot of overlap in shift towards LBK vs WHG, or EHG vs WHG, even between populations who are quite distinct in their overall "Eurasianness" (least Eurasian are Ju_Hoan_North, of course).

It might be interesting to potentially see Mota on such a graph, to see whether it is less Eurasian, or just less shifted between particular modern day West Eurasians (which of course, if Mota is more "Eurasian" than modern Africans, but less "West Eurasian" might also mean Mota is potentially a relict of an older African+Eurasian fusion).

German Dziebel said...

@Capra

"So your interpretation of that PCA is that modern humans *literally* have chimp admixture that archaic humans lack?"

What are you talking about? Read above.

"If A00 went extinct that would make the tree younger indeed, but it would not affect "estimates of divergence between what's left"."

Fair. You were just getting annoying. But the point is that, in the case of admixture, a colonizing population absorbs a bunch of random old lineages from an autochthonous population, that may coalesce very long ago and show multiple gaps in lineage coverage. In the case of common descent there's a tight lineage cluster. We have a tight lineage cluster outside of Africa, and a subset of that cluster is found all over Africa.

@Matt

"Hopefully Eurogenes blog comments will not become as pointlessly "Germanized" as did those at Dienekes'."

Wait. I enjoy playing a fox in a chicken coop chasing the matts, the tobuses, the capras and the andrews of the web world.

capra internetensis said...

Sorry, Matt, I'll stop feeding him.

There are a ton of interesting analyses to be done with this genome, once it's available and people with adequate computers donate their time... (thanks people with adequate computers!)

Papuans are supposed to be descended in part from a *first* Out of Africa, so differences in affinity to Papuan might actually reflect some structure with respect to that.

Han might have a tiny bit of that, but I seriously doubt it would be enough to throw off the f4 ratios using Han as an outgroup. But who knows.

Maju said...

@Karl: Don't you think that it is highly suspect that consolidated farmer populations like the Yoruba or Luhya have barely more admixture than the Mbuti or the less admixed Khoisans? Don't you think that this almost total lack of clinality makes no sense whatsoever?

Also, don't you think that the effectively zero (statistically insignificant or negligible) Neanderthal admixture result of the Mbuti is very much contradictory with that alleged 6% Eurasian admixture?

As far as I can discern there are two possible hypothesis:

1. (Borrowed from someone, not sure who but makes some sense anyhow): the admixture is much older (allowing for greater homogeneous dilution) but Mota was for some reason (isolation most likely) preserved from it.

2. (My own): it's an artifact and further research will most likely disprove or at least redefine it into a more logical result.

And, to expand the puzzle with more near-impossible to explain questions: why are Belorussians and Lithuanians next in line after Sardinians as best candidates for the source of the alleged "Eurasian admixture", when they are so different in every West Eurasian and also Europe-only analyses? How do you make sense of that? Gravetto-Cardial?! O.O

Tobus said...

@German:
I agree with Matt's wish not to have this blog devoted to your theory, so I've decided I'm not going to continue correcting your (blindingly obvious!) mistakes as well - you seem happy being wrong, so I'll let you enjoy it :)

German Dziebel said...

@Tobus

You know that in response to your Edenic nonsense I will always give you a piece of good science. So, keep out. :)

Tobus said...

@Maju:Also, don't you think that the effectively zero (statistically insignificant or negligible) Neanderthal admixture result of the Mbuti is very much contradictory with that alleged 6% Eurasian admixture

Q: If there's 3% Neanderthal in Eurasians and 6% Eurasian in Mbuti, how much Neanderthal is there in Mbuti?
A: 0.03 * 0.06 = 0.0018 = 0.18%

No?

Maju said...

@Tobus: Table S9:

% Neanderthal component haploid full data:
Yoruba: 0.62 (± 0.50)
Mbuti: 0.23 (± 0.45)
French: 2.92 (± 0.55)

So, if Yoruba have (table S5) 7-8% European admixture, then their expected Neanderthal genome would be 0.22%. So, yeah, you are right, the inconsistence is not in Mbuti but in the Yoruba, which appear to have almost 3x the expected Neanderthal admixture.

You may say: error margins! Sure, that "solves" it all but it may even make the Mbuti Neanderthal admixture "negative" (up to -0.22%).

Anyhow the method inflates the French figure from the 2.4% of previous studies to 2.9%, almost all the error margin but it may be a systematic error and is indeed within margins, so whatever.

Maju said...

Anyhow notice that the other column ("haploid, transversions") follows almost an identical pattern: Yorubas seem to have much more "Neanderthal" than expected.

capra internetensis said...

It would be interesting to see the same stats with WHG replacing LBK and see how well they track each other. If there is some Basal Eurasian thing going on they should diverge.

Gihanga Rwanda said...

D(Yoruba, Mota; Sardinian, Han), D=0.0098 Z=5.008
D(Yoruba, Mota; Loschbour, Han), D=0.0059 Z=3.366

They diverge some; don't really know what to make out of it.

Karl_K said...

@Maju

"Anyhow the method inflates the French figure from the 2.4% of previous studies to 2.9%, almost all the error margin but it may be a systematic error and is indeed within margins, so whatever."

That is the whole point. French have more. We have a new baseline using the new genome. The old values were too low.

Alberto said...

I think that the slight inconsistency in the stats about he possible Eurasian admixing population is due to they using the African populations with the least amount of admixture. If Mbuti only has 6% of it, it's normal that the stats can't have enough data to clearly make a distinction between Sardinian, LBK or Belarusian.

I understand they used Mbuti and Yoruba to avoid the populations that might have gotten much later admixture (like from the Arabic/Islamic expansions), but it's a double edge sword. Using populations with 20% admixture you could get more accurate results about the admixing population, but less accurate if that admixture was very recent.

In any case, I think it doesn't really matters. Most of these back migrations had to predate Mota's time (because crossing the Sahara once it became a huge desert seems difficult and meaningless for farmers/herders to do), and it had to be something Near Eastern in location. So I think something EEF-like is pretty much it, even if the stats are a bit erratic (and by the end of the day, looking at the bigger picture, Europeans in general -including northern ones- are the best matching modern populations for those Near Easter/Anatolian farmers anyway).

German Dziebel said...

Wow, guys, I see there's been a fountain of a conversation going on here since I left.

@Matt

"Mota is more "Eurasian" than modern Africans, but less "West Eurasian" might also mean Mota is potentially a relict of an older African+Eurasian fusion)."

That's a good thought. (It agrees with my contention that "Basal Eurasian" is a West Eurasian, not a "Eurasian" component.) It's definitely premature to consider Mota "unadmixed."

Nirjhar007 said...

Guys,
Interesting reading, don't know how to relate BTW.
https://en.wikipedia.org/wiki/Origins_of_Hutu_and_Tutsi
Modern-day genetic studies of the Y-chromosome (which is transmitted solely on the paternal line) suggest that the Tutsi, like the Hutu, are largely of Bantu extraction (80% E1b1a, 15% B, 4% E3). Paternal genetic influences associated with the Horn of Africa and North Africa are few (1% E1b1b), and are ascribed to much earlier inhabitants who were assimilated.

In comparison to the Hutu, the Tutsi have three times as much genetic influence from Nilo-Saharan populations (14.9% B) as the Hutu (4.3% B).


Trombetta et al. (2015) found 22.2% E1b1b in a Tutsi sample from Burundi, but 0% in the Hutu and Twa of Burundi

batman said...

Still some major "calibrations" to do:


"Here we present evidence from the newly excavated Fuyan Cave in Daoxian (southern China). This site has provided 47 human teeth dated to more than 80,000 years old, and with an inferred maximum age of 120,000 years."

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature15696.html

So - where does that place the Cro-Magnons - and the makrogroup CF?

Kelvin Walker said...

Non black people didn't exist until after 10,000 bc

Maju said...

@Kevin: we don't know what you seem so sure about. What is "black people" anyhow?

duge_buwembo said...

The north African Berbers are 20% black African admixed. You cannot use Haplogroups to prove that in Ancient times that the North African Berbers were in place and looked the same. You need to understand what Haplogroups are,they mean nothing because this debate is about racialism which is based on phenotypes. Haplogroups don't correspond to phenotypes. The Andamanese and other Australoids prove this. They don't have African Haplogroups but phenotypically they resemble modern Black Africans.

Maju said...

@Duge Buwembo: even if Andamanese may "resemble" Black Africans in shallow phenotype analysis, this only proves how shallow and largely useless anthropometry is, because Andamanese and other Negrito and Melanesian populations of that area are in fact a vast array of "end branches" of the Eurasian branch of Humankind, the one that migrate out of Africa. This is not only apparent in haplogroups but also in any sort of autosomal genetic analysis.

It basically seems to prove that the original human phenotype was "black", i.e. dark (highly melanized) skin, curly black hair, brown eyes... and that all the rest is either ex-novo adaptive evolution (lighter skin color, with two pathways: the West Eurasian and the East Asian one, genetically unrelated, also the Australo-Melanesian pathway to blond hair, distinct from the European one) or introgression from Neanderthals (probably straight hair, although also must have got some adaptive value, maybe against cold or rain, as it was selected for). Tropical Asian and Oceanian branches just retained better the original phenotype precisely because they lived in environments more similar to those of Tropical Africa and also they remained mostly isolated until recently.

But in any case, an Andamanese is more directly related to a European or a Chinese than to a Nigerian. Looks can be misleading.

Normandie Kent said...

Native Americans were non-black people, living in the America 25,000 years ago, they existed. So did Aboriginal people who were living In Australia for 50,000 years, they are not Black as in African.