Y-DNA E Didn't Back Migrate To Africa

One theory held by Dienekes and a few others (and a minority of archaegeneticists) is that Y-DNA haplogroup E, the predominant Y-DNA type in Africa (with most of the remainign men having Y-DNA haplogroups A, B, T or R1b-V88) is a back migration to Africa from Eurasia. The points below, originally composed and somewhat edited since, as a comment to a post on his blog that reiterates this claim, sets out the arguments against this conclusion.

Significant Back Migration To Africa Did Happen

There is no real dispute that there has been some back migration from Eurasia to Africa. Y-DNA haplogroups T and R1b-V88 and mtDNA haplogroups M1 and U6, as well as autosomal data for North Africa, Ethiopia and Somolia, for example, all make a very strong case for back migration having taken place. North Africans, generally speaking have automsomal DNA which 80%-90% or so derived from West Eurasia. Most Ethiopian and Somolians have autosomal DNA which is 40%-50% derived from West Eurasia. The Maasai have a lower but significant share of West Eurasian autosomal genetic contributions. Most other sub-Saharan Africans have almost no West Eurasian source autosomal DNA.

The vast majority of Africans with significant back migrating uniparental and autosomal genetic traces (apart from the people of Madagascar who have quite recent Asian genetic origins in addition to their African roots and cases of recent European admixture such as colored South Africans), speak an Afro-Asiatic language, and most of the rest speak a Nilo-Saharan language. Prehistoric Eurasian admixture is almost completely absent in Paleo-African populations (Khoisan, Pygmies, San, etc.), and populations that speak Niger-Congo languages, apart from some small communities near the outer boundaries of the Niger-Congo linguistic region like the Fulani of Northern Cameroon.

There is also good evidence to suggest that there was at least, a relatively recent back migration that gave rise to the Ethio-semitic languages, and a probably also a more ancient back migration. Ancient DNA and the mtDNA mutation rate data (for mtDNA haplogroupds M1 and U6 which have roughly the same age) suggest that this back migration probably sometime in the late Upper Paleolithic or early Neolithic. A good guess might be around 12,000 years ago when North Africa was repopulated in the Mesolithic at the start of a "wet" Sahara period.

A recent study of autosomal genetics in Ethiopia, however, estimates the Eurasian contribution at ca. 3000 years ago, around the time of the Ethiosemitic language subfamily's arrival there as estimated from linguistic evidence.

But, this most recent study but does not show much difference in Eurasian genetic contributions between Ethiosemitic language speakers and populations such as Cushitic language speakers who are known to have preceded Ethiosemitic language speakers in the region, despite an expectation that a three thousand year old contibution of 40%-50% of Ethiopian autosomal genes would show greater percentage contributions in Ethiosemitic language speakers than in earlier Cushitic language family speakers as earlier uniparental genetic studies had suggested might be the case, particularly on the Y-DNA side (suggesting a male dominanted Ethiosemitic colonization event).

When Did The Out of Africa Migration Happen?

There is archaeological evidence for an Out of Africa and into the Near East migration from somewhere around Sudan to Arabia via the Sinai Pennisula around 104,000 years ago. Modern humans aren't found in Eurasia outside the Middle East (at that time, in South Asia) before about 74,000 years ago, however, and there is a gap in the record of modern humans in the Levant from about 75,000 years ago until about 50,000 years ago. At that time modern humans could have retreated to non-Levantine refugia in Eurasia, or could have simply failed only to be demically replaced by a sceond wave of modern human migration.

The case for a Eurasian refugia during modern human's absence from the Levant is the stronger one.

The flat phylogeny (i.e. with many contemporaneous branches dispersed over a very great geographic area) of mtDNA haplogroup M suggests a rapid Southern route expansion when modern humans finally burst into East Eurasia. The youngest archaic hominin remains on the mainland side of the Wallace line in East Eurasia appear to date from around 100,000 years ago, and the well dated appearance of modern humans in Australia and Melanesia around 45,000 years ago bounds the modern human expansion into East Eurasia from South Asia to sometime before that date. An expansion of modern humans into East Eurasia from South Asia around the time of the Toba explosion, 74,000 years ago, would be a fairly good match to the available genetic based dates from the Y-DNA CF branch, and from the mtDNA M and N derived haplogroups found in East Eurasia, and would also be consistent with the rather thin early archaeological record in East Eurasia from that time period.

As I explain at greater depth below, the distribution of Y-DNA haplogroup D suggests that it was a maritime secondary Out of Africa wave, rather than being part of the initial Southern route expansion Out of Africa, or the migration of modern humans into Europe.

Y-DNA Haplogroup E is not derived from a back migration to Africa

The case for a back migration of Y-DNA haplogroup E, the African branch of Y-DNA clade DE, is not well supported, however.

Y-DNA Haplogroup E as a clearly African rooted distribution.

Y-DNA haplogroup E is the predominant Y-DNA haplogroup in most of Africa. Y-DNA haplogroup E is common in three of the four major linguistic groups of Africa (Afro-Asiatic, Nilo-Saharan, Niger-Congo) and is present in all of them, unlike back migrating Y-DNA haplogroups T and R1b-V88, and back migrating mtDNA haplogroups M1 and U6, all of which are principally found in Afro-Asiatic language speakers. Auotosomal genetic components from West Eurasia in Africa likewise are predominantly found in Afro-Asiatic speakers, while they are nearly absent in West African Niger-Congo speakers.

Furthermore, Neanderthal DNA, which is found at about the same percentage frequency in all Eurasians, is absent from most African populations who are overwhelming Y-DNA haplogroup E bearers. Thus, back migrating early Y-DNA haplogroup DE* or haplogroup E bearers, if they existed, would have had to do so prior to Neanderthal admixture which appears to have happened before or at around the same time as West Eurasian and East Eurasian populations became distinct (and certainly prior to 45,000 years ago, as Australian Aborigines and Papuans have the same levels of Neanderthal admixture as other Eurasians). Indeed, even if the YAP mutation that defined Y-DNA macrohaplogroup DE actually did take place in Eurasia, the circumstantial evidence would suggest that this would have had to have taken place before these people developed most of the genetic features that make Eurasians genetically distinct from Africans - call them "pre-Eurasians" in Eurasia, perhaps.

Y-DNA haplogroup E subhaplogroups that do not appear to have a source in Bantu introgression are found in African pygmies, whose population genetic split from other Africans is one of the most basal dating to perhaps 70,000 years ago, and who were believed to have been linquistically distinct prior to Bantu introgression. This points to its African antiquity.

Africa also has basal branches of Y-DNA hg E not found anywhere else (e.g. E2). The most basal lineages, paragroup E*, have been found in a single Bantu-speaking male from South Africa, amongst pygmies and Bantus from the Cameroon/Gabon region, and in two individuals from Saudi Arabia. The distribution of Y-DNA haplogroup E overall in Africa is continuous throughout Northern and Subsaharan Africa, and is terrestrial.

All of the basal branches of Y-DNA haplogroup E are found in Africa, including Y-DNA subhaplogroup E2, some major lineages within subhaplogroup E1, and parahaplogroup E* that are found only in Africa. In contrast, all of the subhaplogroups of Y-DNA haplogroup E which are found in West Eurasia are from phylogenies rooted in Africa. Moreover, the SW European and SE European Y-DNA haplogroup E subhaplogroups don't have a common European source - their phylogeny reveals separate West Asian and Northwest African origins for Y-DNA haplogroup E in Europe with European specific lineages within these African rooted clades that are comparatively recent.

Inconsistently with a back migrating Y-DNA haplogroup, Y-DNA haplogroup E is less common in East Africa, where Y-DNA haplogrups A and B (found only in Africa and basal to both Y-DNA macrohaplogroups DE and CF) are more common, than Y-DNA haplogroup E is in most of the rest of Africa. All other back migrating uniparental markers and automomal genetic clines display the opposite trend in their distributions.

Y-DNA Haplogroup D has a sporadic East Eurasian distribution.

Y-DNA D is a rare haplogroup that has a sporadic Asian distribution with a deep rooted divide between D2 a Japanese subhaplogroup, and D*, D1 and D3 found in a network centered on the Andaman Islands and Tibet, with low frequencies in Eastern India and Southeast Asia, and seeps out from Tibet at low frequencies into Central Asia. Y-DNA haplogroup D is absent from Africa (including Madagascar), Oceania (which was colonized first by Austronesians), Europe, other parts of West Eurasia, the Americans, Melanesia and Australia.

It appears from the mtDNA make up of populations where Y-DNA haplogroup D is found that the earliest Y-DNA haplogroup D populations in East Eurasia probably had female counterparts who were part of mtDNA M (excluding later separately named mtDNA haplogroups) which is found only in East Eurasia, with no counterparts who were from mtDNA haplogroup N and its decendants that are found in both West and East Eurasia.

The geography of D is consistent with long range maritime migration to successive uninhabited islands or unoccupied areas in the wake of migrations occupying most territory by CF clade members. It may represent a secondary Out of Africa migration.

Modern humans did not arrive in Japan until about 30,000 years ago, the likely root of the Y-DNA D2 subhaplogroup which is restricted almost entirely to Japan or in much lower frequencies adjacent areas traceable to Japanese migration, so Y-DNA haplogroup D in East Eurasia can be no younger. The Andaman islands may not have had modern human occupants until about 20,000 years ago. Tibet had modern human occupants prior to 20,000 years ago, but possibly not much before 30,000 years ago.

Despite the fact that modern humans were in South Asia at the time of the Tobu erruption, 74,000 years ago, Y-DNA D is not distributed widely throughout South Asia. It is instead geographically limited to areas with high levels of Ancestral South Indian (ASI) admixture (which has its closest extant match in the Andaman islanders), but the distribution of Y-DNA D in India is more localized than the ASI autosomal genetic component in India and Y-DNA haplogroup D is uncommon even in mainland India.

Modern humans arrived in Europe around 40,000 years ago and it lacks Y-DNA haplogroup D. Australia and New Guinea were first occupied by modern humans around 45,000 years ago and lack Y-DNA haplogroup D. Denisovian autosomal DNA does not overlap with the populations where Y-DNA haplogroup D is found (basically Melanesians, Aborginal Australians and to a lesser extent Philippino Negritos), suggesting that Y-DNA haplogroup D populations probably didn't co-exist with the archaic hominin Denisovians for any prolonged period of time, if at all. Thus, there is strongly suggestive circumstantial evidence to suggest that Y-DNA haplogroup D dispersed after 40,000 years ago, but before 30,000 years ago, long after the initial Out of Africa migration, and close in time to when the Neanderthals began the downward spiral into extinction in West Eurasia.

The Americas, which lack Y-DNA haplogroup D, were mostly first occupied by modern humans starting around 14,000 years ago via the Bering Strait from Siberia (in mainland East Eurasia), although that founding population could have arrived in Beringia as soon as about 22,000 years ago. The Beringian first Native American founding population, among other things, supported itself by hunting mammoths. And, the earliest modern human migrants to the Americas from Beringia probably didn't have boating capabilities on a part with Y-DNA haplogroup D dominanted peoples who first settled the Andaman Islands and Japan, since if they did, the North American glaciers of the last glacial maximum would not have presented a serious barrier to their migration into the rest of the Americas. Later circumpolar populations in the Americas who did have solid maritime abilities arrived in the Americas many thousands of years later than 14,000 years ago.

Of course, the population bottleneck associated with the founding population of the Americas, which may have numbered in the hundreds on a census basis, could easily have excluded many Y-DNA lineages present on the East Eurasian mainland and could even have omitted many Y-DNA lineages that were present at low frequences in the founding population of Berginia before it expanded into the rest of the Americas as these lineages could have been lost due to genetic drift. Still, it is safe to say that the founding population of the Americas was probably not derived from a population in which Y-DNA haplogroup D was predominant, and Native American population genetics show more affinity to Siberia than Japan or Tibet, where Y-DNA haplogroup D is common.

The confined distribution of Y-DNA haplogroup D bearers to places that were probably either completely or nearly completely unpopulated by modern humans when they arrived, whille they are absent from places were Y-DNA clade CF bearers were already present in any great numbers, also disfavors the notion that an offshoot of their particular culture took Africa by storm in a way that made Y-DNA haplogroup E the predominant patriline genetic source for all of Africa, displacing all of its predecessors and being found in all of Africa's major language family populations.

The split of Y-DNA D and E took place long before the other back migrations to Africa sometime after men with Y-DNA in pre-split haplogroup DE left Africa.

There are just eight known cases of individuals from paragroup DE*, two in Tibet, and six from West Africa, although they may be some in the Andaman Islands. This implies that the dispersal of bearers of pre-split y-DNA haplogroup DE split between Asian and African populations, wherever it happened, took place at some point shortly before, rather than after the DE split.

The YAP mutation that defines DE and associated hinted at research from the 1000 Genomes project suggest that DE's divide is vary ancient, dating to roughly the Out of Africa era, or at the latest, early Upper Paloelithic. Mutation rate evidence that Y-DNA haplogroup D is older than Y-DNA haplogroup E has been contradicted by later research by some of the same investigators.

While mutation dating is problematic, it isn't so irrelevant and inaccurate that one cannot say that the DE divide is much more ancient than the mtDNA M1/mtDNA U6/Y-DNA T/Y-DNA R1b-V88 backmigrations or events giving rise to Ethiosemitic languages.

The timing of the Y-DNA DE split is at an age roughly contemporaneous with the split of the Khoisan and Pygmy populations from other Africans, the most basal population split in African population genetics.

The timing of the haplogroup D expansion suggested by the geographic distribution of Y-DNA haplogroup D is much younger than the date suggested by mutatioon rates which indicate that Y-DNA haplogroup D is only a bit younger than the root of Y-DNA macrohaplogroup CF, which is the dominant non-African Y-DNA clade and the only one found at any great distance from Africa in East Eurasia other than much more rare Y-DNA haplogroup D. The date implied by the mutation rates is on the order of twenty thousand to thirty thousand years more recent than the best estimate of the date at which Y-DNA macrohaplogroup DE split into Y-DNA haplogroups D and E.

A back migration of people with Y-DNA haplogroup E to Africa, in addition to preceding Neanderthal admixture with non-Africans, would also have to have been made up exclusively of men and of women with mtDNA haplogroup L3, the only African mtDNA lineage that is a source for non-African mtDNA lineages, because mtDNA haplogroup M1, the only potentially East Eurasian origin mtDNA lineage in Africa, arrived in Africa scores of millennia later. But, there are no indigeneous populations of East Eurasia who had mtDNA haplogroup L3 prior historic or almost historic era migrations (i.e. prior to 6,000-7,000 years ago).

A more plausible possibility that could explain this gap is that a small group of pre-split haplogroup DE bearing men from Africa either migrated to a refugium (perhaps in somewhere Arabia that subsequently became too arid to leave a remnant population in Arabia, or perhaps in South India) where they were part of a small endogameous community until their maritime expansion into East Eurasia that wasn't swept up in the initial Southern route expansion out of Africa, much later.

Alternately, this small endogamous community of pre-split haplogroup DE men could have existed for tens of thousands of years somewhere in Africa, only to be totally replaced within Africa during the more successful expansion of haplogroup E within Africa sometimes between then and the Bantu expansion, leaving only a handful of of descendents who ultimately ended up in West Africa behind, and leaving no surviving members who had been part of a D lineage that emerged in Africa.

Conclusion

After considering all of the evidence, it becomes clear that a scenario in which Y-DNA haplogroup E back migrates from Eurasia to Africa is grossly implausible, even though it is possible to imagine highly contrived and unprecedented set of events. It might not be proof beyond a reasonable doubt, but there is surely clear and convicing evidence that Y-DNA haplogroup E has its origins entirely in Africa.

Also, even if Y-DNA haplogroup E did back migrate from Eurasia to Africa, this would have had to have happened so early on in the Out of Africa migration process that its place of emergence would not different in practical observable implications in any meaningful way from an African origin, because it would have had to have predated the point at which the Out of Africa population had become very distinct genetically from Eurasians.

None of this is highly controversial. All of the facts, although not quite all of the connecting of the dots made in this post, is derived from peer reviewed published scholarly journal articles, either directly, or as citation support for my sources. I haven't taken the time to annotate this post (although I aspire to at some point), but have read almost all of those articles and blogged many of them.

The last published paper to argue seriously for a non-African origin of Y-DNA haplogroup E in Africa was published in 2007, and many of the sources I am relying upon in this post (including all of the Neaderthal DNA work and a lot of the detailed subtyping of Y-DNA haplogroup D and Y-DNA haplogroup E in Europe and Ethiopian DNA and much of the automsomal work, in general) has been published since then. The Asian origin for Y-DNA haplogroup E was a distinctly minority view in 2007, was disavowed by one of its original proponents based upon new evidence in 2008, and has almost surely lost informed supporters based upon the increasing weight of the evidence against this theory since then. I wouldn't yet call the theory pseudo-science, but it is starting to approach that threshold as our picture of humanity's genetic prehistory grows clearer.

We also know from recent research on Y-DNA haplogroup E in Europe, which I've previously blogged, that most of the Y-DNA haplogroup E in Europe has African sources in the Holocene era and in many cases much more recent African origins.

One important consequence of an African origin for Y-DNA haplogroup E (or at least a very early back migration that left no traces behind in Eurasia), is that it allows us, as we try to glean humanity's prehistory with archaeogenetics, to focus our attention entirely on Africa, with significant back migration from Europe taking place only fairly late in the game and probably in at most just a few main waves (Epipaleolithic, Neolithic, Ethiosemitic, and some historic era migration).

To the extent that particular Y-DNA haplogroup E lineages are associated with particular linguistic families or subfamilies in Africa, the populations that expanded and spread those languages expanded and spread those languages from a population that was mostly African in origin, even if there may have been outside thin superstate influences in the initial population from which it expanded.

While not nearly so decisive, this conclusion is also an important clue that tends to favor an African continental, rather than a Levatine origin of the Afro-Asiatic languages, because Semitic language speakers in the Levant have much more frequently bear Y-DNA haplogroup E than other West Eurasian populations. So does the fact that we can historically date the time period when Semitic languages first started to be spoken in the Eastern part of the Fertile Crescent (after about 2000 BCE), and their absence from Anatolia apart from small Akkadian trading colonies and influences from Arabic speakers in the last thirteen hundred years (after about 700 CE).

This conclusion isn't decisive, because it departs from the usual trend in which the first farmers exert cultural dominance over the people they encounter as they expand. But, an emerging picture of the North African Neolithic in which much of the demic influence of West Eurasia on North Africa precedes the Neolithic by a few thousand years as the area is repopulated during a "wet Sahara" climate phase, and North Africans become herders long before becoming farmers in a process that allows for a transition that has a stronger cultural component and a weaker demic component than parallel developments in Europe, weaken the strength of a first farmers paradigm for the linguistics of the Afro-Asiatic language family.

Instead, a scenario in which Afro-Asiatic languages are spread by indigenous hunting and gathering populations who adopted Fertile Crescent herding practices with animals domesticated there, while avoiding a profound introgression of peoples from the Fertile Crescent at that point, seems more plausible. The expansion was probably driven by food production technologies, and individual subfamilies within the Afro-Asiatic language family do show strong signs of demic expansion. But, some of the different linguistic families, at least, probably arose from language shift rather than population replacement, because the Afro-Asiatic language family, while made up of populations that have a number of genetic features distinct for Africa, are not geneticallly homogeneous and can't be linked in any simple phylogenetic tree at a population genetic level.