Following with the previous post on SE Asia, I just made what is my best hunch of what could have been the migration routes of early Eurasian peoples:
 |
| click to expand - apt for wallpaper |
No arrows are placed as the routes may have been used biderectionally at times (though the general direction of the migration was West to East, naturally). Flows between adjacent routes were of course most likely to have happened (capillarity not represented) Also I'm a bit unsure about which could be the major routes in Mid and NE Asia specially, so I followed coast and rivers but, if you have any ideas, please let me know.
1. Interesting and a good start. I really suck at creating these kinds of graphics (the last art or drafting class I took was when it was last required in school when I was twelve years old) and also don't have any good software for doing it, so I really respect and appreciate the efforts of people like you and Razib who can produce maps like these and other graphics.
ReplyDelete2. There was probably at least one thin "Northern route" across Siberia from the Near East that would have carried mtDNA hg X2 (but seemingly no other West Euraisan Y-DNA or mtDNA haplogroups) to Northeast Asia and the New World, merging into the Southern route groups that make up the balance of the New World migration. The absence of mtDNA hg X2 anywhere on the coastal route seems to make this separate Northern route connection presumably linked to Paleosiberian populations like the Ket, necessary sometime in the Paleolithic.
3. I suspect that a lot of the early migration was coastal, which for key parts of the relevant time periods would have been just off the coasts of the land masses with current coastlines shown on the map. Shallow water archaeology in the Indian Ocean and Persian Gulf remains a potential goldmine for paleolithic era finds.
4. It isn't entirely clear from your map in the vicinity of Pakistan, if it reflects the ancient boundaries of the Sarasvati River, or the boundaries of the current river system that arose ca. 3000-5000 years ago.
5. I would extend the Western most Zumia line into Tibet proper (given the strong affinities between the oldest lineages in the Andamans and Tibet), but I'm not convinced that you have the direction of flow right for Eastern Zumia. I think that there is case to be made that this flow from from Tibet to the coast instead of the other way around. I'm also inclined based on the genetic studies of Tibet to think that the only other route of demic migration into Tibet was via the Yellow River.
6. While the "who" got there first can be a hard question when there is evidence of multiple waves of pre-historic migration, the archaeology provides some pretty good guidance regarding "when" modern humans first arrived in particular places in any meaningful numbers, i.e. where the frontier of human migration was a particular times. One way to show frontiers, that wouldn't require revision every time a new paper comes out would be to put "flags" or "dots" with a date and/or name for some of the oldest archaeological finds in each region (e.g. oldest known modern human habitation in Tibet, Australia, China, India, Japan, etc.).
ReplyDelete7. Given the discoveries of 2010, it seems as if there ought to be some representation of the range of the Denisovians and their encounters with early modern humans, as should the Hobbits of Flores. This might be best captured by a multiple sequential frame representation as well, with the Denisovians admixing with modern humans in the first frame and then vanishing. This may also shed light on the Y-DNA C v. Y-DNA D sequencing. Despite its scattered distribution, which would normally suggest an older strata, the apparent lack of Denisovian admixture in populations rich in Y-DNA D (Andamanese, Japanese, Tibetans, Northeast Asians), argue for this not being a wave that came into contact with the Denisovians. Perhaps some colored or hash marked blob to indicate the likely region of possible admixture in the Denisovian range would be in order.
8. Incidentally, the direction of migrations into Tibet is one of the reasons that I agree with you that the Sino-Caucasion and Sino-Basque linguistic connections just don't make any sense at all, except to the extent that they come through South Asia.
The only East Eurasian/New World languages with any plausible link to West Eurasian languages after a very early West Eurasian-East Eurasian split would be the Ket/Na-Dene language family, since the Na-Dene speakers are the subset of New Worlders with mtDNA hg X. But, since "DNA studies show [that the Ket have] affinities to the peoples of Southeast Asia (Tibetans, Burmese, and others) not shared by other Siberian peoples.", it seems plausible that the mtDNA X may have arrived via low level admixture with the Ket's West Eurasian origin neighbors in Siberia to become a minor Ket genetic component (perhaps lost now) with no linguistic impact.
9. I'm inclined to think that there were multiple pulses of migration, and the uniparental phylogeny literature supports this view. Ideally, one would want a slideshow with sequential frames, or multiple colors with dates at key points along them to demonstrate this concept. Some of the recent literature on Y-DNA hgs C and D that I've seen has explored this somewhat. But, of course, figuring out this sequence is tricky, is clouded be subsequent mass migrations of peoples, and requires considerable faith in the accuracy of at least relative mutation rate dating in different uniparental hgs.
ReplyDelete10. One virtue of a multiple sequential frame approach would be to show changing coastlines and changing freeze lines that made parts of Northern Siberia and Northern Tibet uninhabitable at certain points. If one wanted to be cute one could also put little icons of mammoths to symbolize pre-extinction megafauna in some of the earlier frames, and show them absent in some of the later frames.
11. In addition to open water, there have been some key land barriers that have kept regional paleolithic populations pretty distinct from each other (but, given semi-nomadic living, in a less fine grained way that the Neolithic where clines can pinpoint people to with ten km or so of their ancesteral homes).
For example, the India-Burma division has had remarkably little gene flow. Also, the Tibet-India division has been for long time periods almost a valve allowing people out but not in from some directions after its earliest colonization. The Tarim Basin-Lake Baikal-Altai region also seems to be one of the ancient boundaries between West and East Eurasians until the Turkish expansion in the CE, with a component of Proto-Uralic migration in the circumpolar region as the other main exception.
A migration model that makes breaks Eurasian into a dozen or two geographic regions in the Paleolithic, quantifies the small amount of interregional admixture, and models intraregional migrations and social strata separately after the first cut of regional separation, seems like a pretty good fit for the genetic data.
One could then have arrows in both direction for adjacent regions proportionate to the indications from autosomal DNA and known historical movements of the extent of admixture between adjacent regions in each direction.
Color coding and labeling different pre-historic population regions, as you did for Zumia and easly good for Sahul, Australia, Tazmania, Melanesia, Ancesteral South India, Ancesteral North India, Tibet, lowland Southeast Asia, South China, North China, Japan, and Northeast Asia in territories that match the genetic clusters and reflect these boundaries make it easier to grok the migration arrows. Indeed, if one had a few circumpolar regions (Canadian circumpolar, Beringian circumpolar, Eurasian circumpolar), then inter-regional migration arrows could show that as well.
Andrew Oh-Willeke, it seems when a New Asians like Southern Chinese and a Southeast Asian Men Y Hg O*-M175, also a descendant of Y Hg K2*-M526, a descendant of majority a European and a Middle Eastern Men with a Y Hg R1a-M17, Hg R1b-M343, or even an Old Western Eurasians Men Y Hg I*-M170, Hg J*-M172 and Y Hg T*-M70?
Delete"For example, the India-Burma division has had remarkably little gene flow".
ReplyDeleteYou'll have trouble getting Maju to agree with that. I've been trying to persuade him of it for some time now. Mitochondrial haplogroup M appears to have suffered a geographic bottleneck as it passed across that division.
"There was probably at least one thin 'Northern route' across Siberia from the Near East that would have carried mtDNA hg X2 (but seemingly no other West Euraisan Y-DNA or mtDNA haplogroups)"
I'd certainly add haplogroups A and Y/N9 to that. In fact I think it's quite possible that all the eastern N haplogroups, including the Australian ones, crossed Eurasia via that route. R then moved back west from SE Asia.
"I suspect that a lot of the early migration was coastal, which for key parts of the relevant time periods would have been just off the coasts of the land masses with current coastlines shown on the map".
Possibly, but just as likely not. For example I would not have arrows showing the early arrival in South India that Maju has on his map. It seems to me now that the South India M haplogroups are relatively recent arrivals from further north, not a product of a coastal migration along the southern coastline. Consequently I am also sceptical of the route Maju has marked along the Hadramawt and Makran coastlines.
"I'm not convinced that you have the direction of flow right for Eastern Zumia. I think that there is case to be made that this flow from from Tibet to the coast instead of the other way around".
Agreed.
"I'm inclined to think that there were multiple pulses of migration, and the uniparental phylogeny literature supports this view".
Very much in agreement here.
"If one wanted to be cute one could also put little icons of mammoths to symbolize pre-extinction megafauna in some of the earlier frames, and show them absent in some of the later frames".
And I would lay money on the pattern of extinction following the pattern of human expansion.
1. Get GIMP: it's free and useful. You may need to train a bit but it's great.
ReplyDeleteYou can still do decent maps with Windows Paint anyhow (of course using a base map from elsewhere always).
I'm a horrible artist (drawer) myself but I have a good grasp of geography and use that to my advantage.
(2) "There was probably at least one thin "Northern route" across Siberia from the Near East that would have carried mtDNA hg X2"
This would belong to a second phase, related to the colonization of West Eurasia, not depicted in this map.
I understand that mtDNA N and R coalesced only in SE Asia (or South Asia?) and then back-migrated westwards. But the part affecting West Eurasia is not mentioned except for the thin blue line in Palestine (Skhul/Qahfez) and the routes via Arabia/Iran to South Asia.
X2 is derived from X which is a West Asian clade but X is derived from N, which I locate in SE Asia. So X2 belongs to a later stage, after some N and R backmigrated westwards with, probably, Y-DNA P.
(4) The Sarasvati is mythical only as far as I can tell. I undertand that the modern rivers existed more or less as today in the Paleolithic. I have never read anything against this. Actually for SA I follow essentially Fields et al, as well what I know of Indian pehistory. Nowhere a different fluvial structure is suggested.
(5) "I would extend the Western most Zumia line into Tibet proper".
Zomia. Actually some authors do extend it. But for Paleolithic purposes like these it's less important because the outskirts of Tibet were only colonized since c. 30 Ka ago and the plateau proper surely only in the last 10 Ka or so, as it was no doubt a wasteland in the Ice Age.
Tibet proper is a refuge and not any population source. However the Eastern outskirts (Yunnan, Sichuan) may have been inhabited since old.
I guess it's possible to consider a route somewhat as the one you suggest for Y-DNA D towards Japan. But a coastal route is at least as likely (as Japanese D1 is not particularly related to Tibetan D subclades but links with them via Indochina).
(6)... "the archaeology provides some pretty good guidance regarding "when" modern humans first arrived in particular places in any meaningful numbers"...
Not really. We know very little of the Paleolithic of East Asia, specially SE Asia. We have to wait till the area is properly surveyed and that may take many decades, maybe a century.
All we know is that dates are generally being pushed backwards in time. In some cases, datings are controversial, what does not help at all.
(7) Sure but if one thing I've learned from making maps and designing games is: "keep it simple", at least simple enough.
I dealt with those matters here.
(8) Absolutely.
"... the Na-Dene speakers are the subset of New Worlders with mtDNA hg X".
I'd forget about mtDNA X. It's presence in America is best explained by Y-DNA Q in fact, which has the same West Asian origin (most basal diversity is between Yemen and India) and is much more common (up to the point of annoying uniformity). It's clear that Amerindians and also Na-Dene-Yeniseans and Inuits are your usual Siberian peoples "type A" with lots of Western Y-DNA and Eastern mtDNA. The overall flow of this super-population is SW-NE, instead Y-DNA N peoples have a flow from East to NW (possibly a bit more recently and more strictly attached to the northernmost latitudes, rather than the steppary corridor of Q).
(9, 10, 11) Sure, but I won't work on that unless I'm making some money out of it. Too much work!
"For example, the India-Burma division has had remarkably little gene flow".
ReplyDeleteYou'll have trouble getting Maju to agree with that. I've been trying to persuade him of it for some time now.
I'm not sure anymore of what you are "trying to persuade" me of. In a different conversation you were just insisting on the various M sublineages that criss-cross that border area.
IMO there is a buffer zone in NE India (not any rigid barrier, just a lower density area) but this one did not work until a second phase in the Great Eurasian Expansion. In the first phase it is clear that haplogroups cross that "border" once and again (at least M, N and R do, as well as some of thir sublineages.
But then the flow stops almost totally (except for some lesser stuff into NE India's buffer zone from both regions).
I'd certainly add haplogroups A and Y/N9 to that.
First you'd have to demonstrate a West Asian origin for macro-haplogroup N, what cannot be done AFAIK.
I would not have arrows showing the early arrival in South India that Maju has on his map.
I did not put any arrow at all.
It seems to me now that the South India M haplogroups are relatively recent arrivals from further north, not a product of a coastal migration along the southern coastline.
It seems to me that you are just a fanatic of your own ideas with little substance behind your beliefs.
@ Maju : "as Japanese D1 is not particularly related to Tibetan D subclades"
ReplyDeleteThe Japanese y-dna D is D2.
http://www.isogg.org/tree/ISOGG_HapgrpD.html
http://www.biomedcentral.com/1741-7007/6/45
http://www.biomedcentral.com/1741-7007/6/45/figure/F3
"The Japanese y-dna D is D2".
ReplyDeleteSure: my bad, it's C1, right?
The Sarasvati is mythical only as far as I can tell.
ReplyDeleteOne can dispute that the Ghaggar-Hakra river system was actually the Sarasvati, but there is no reasonable dispute given satellite imagery, paleo-climate date, and archaeolgy showing many settlements along what are now dry riverbeds that its course changed pretty dramatically sometime in the time period ca. 2500 BCE to 1500 BCE, and that the Hakra once flowed through the Thar Desert. This goes back at at least 10000 years BP, but may have had a different form yet before then.
The main implications of this are to shift to the East the populations in the Indus River Valley area and to weaken the East-West nexus.
IDK, in truth the map lacks a reference for some West Thar Desert sites, that are a bit off where expected.
ReplyDeleteA lesser route through "nowhere" (Rajastan) between Gujarat and Haryana or Uttar Pradesh makes some sense. But this may be not different from general porosity, right? The Thar desert should have existed for all the Paleolithic anyhow.
it's C1, right?
ReplyDeleteYep.
"Haplogroup C1, an ancient but at present extremely rare lineage, is specific to the Japanese and Ryukyuan populations of Japan, among whom it occurs at a frequency of about 5%"
-> http://en.wikipedia.org/wiki/Haplogroup_C_(Y-DNA)#Distribution
"In a different conversation you were just insisting on the various M sublineages that criss-cross that border area".
ReplyDeleteBut you have consistently disagreed with my comments on the region, claiming no barrier at all. In fact you've argued that the haplogroups in the region have all come either from the west ofr the east. None are indigenous to the region, which is pretty unbelievable.
"First you'd have to demonstrate a West Asian origin for macro-haplogroup N, what cannot be done AFAIK".
Surely N is an L3 lineage, so it must have come from Africa. I find it unbelievable that it could have traveled all the way to SE Asia leaving not the slightest trace of its passing.
"I did not put any arrow at all".
OK. I wouldn't have put the swirling lines in Southern India that you have placed there.
"It seems to me that you are just a fanatic of your own ideas with little substance behind your beliefs".
Maju, don't be ridiculous. Have a look at the haplogroups instead of insisting on your own ideas with little substance behind your beliefs.
"there is no reasonable dispute given satellite imagery, paleo-climate date, and archaeolgy showing many settlements along what are now dry riverbeds that its course changed pretty dramatically sometime in the time period ca. 2500 BCE to 1500 BCE, and that the Hakra once flowed through the Thar Desert".
I remember something about the Ganges capturing the eastern headwaters of the Indus at some time. Presumably that is the phenomenon you're mentioning.
"Haplogroup C1, an ancient but at present extremely rare lineage, is specific to the Japanese and Ryukyuan populations of Japan, among whom it occurs at a frequency of about 5%"
ReplyDeleteAn ancient lineage. So does anyone really believe it arrived there from SE Asia? Via the Ryukyus perhaps?
"I find it unbelievable that [N] it could have traveled all the way to SE Asia leaving not the slightest trace of its passing".
ReplyDeleteWell, I find it quite believable because that's what the data tells. Similarly you could say that it's unbelievable that M could have traveled all the way to South Asia not leaving the slightest trace... but that's what happened.
We see no pre-M nor pre-N anywhere except at the L3 node. That's what we know, believe it or not.
"I wouldn't have put the swirling lines in Southern India that you have placed there".
Seems demonstrated by archaeology and simulations.
"So does anyone really believe [C1] arrived there from SE Asia?"
I do.
Maju, a moderator of a Chinese DNA FTDNA Forum named Owen Lu told me when my mtDNA Hg B4 (Geno 1st) - B4c - B4c2 (Geno 2,0 Revision) are more related with a Northern Chinese people who speak an Altaian and Mongolian, or not? If i see my results from Geno 2,0 ; My mtDNA Maternal line Hg B4c2 are born aroud 8000 Years ago BP in Asia Continent and then a carrier this mtDNA went and stayed in ISEA like Indonesia Archipelago.
DeleteI wouldn't be able to be so precise but in general terms it seems very apparent to me that mtDNA R and Y-DNA K2 expanded together. This does not necessarily mean 100% together, together in every corner and occasion: people do get mixed every single generation (that's what sexual reproduction, including the taboo of incest, is about) and now and then this admixture necessarily happened across "paleo-ethnic" borders.
DeleteWhat we have to understand is that our ancestors 500 years ago are around one million (some are probably the same person reaching to us via different ancestral lines but still) and that, when go back to these extremely ancient dates (I estimate this expansion to be 60 or 70 Ka old) the figure of ancestors is mind-boggling, many many times the human population of Earth back in the day (but again they are repeated at many different root positions, and that's what make the various populations distinct: repeated "cousin marriage", sort of: more likely to be 3rd or 4th degree cousins than 1st degree ones, safe enough for biology).
But a common case, it seems, is that two originally distinct populations (A and B) become neighbors and begin intermarriage. So you begin with AA and BB pops. but you end up with AB and BA (or maybe a mix of BA and BB, if B is significantly larger than A), where the first letter represents the patrilineages and the second one the matrilineages. Patrilocality is assumed but that's what we find in most cases in fact, especially towards the North. Some anthropological studies have suggested that brothers' cooperation is slightly more stable than sisters' one but one can also appeal to the reality of small nomadic groups that you can still find in some parts of North Eurasia (or Arctic America as well) because the environment does not really allow for larger groups, which were probably more common in the more benevolent southern parts. For these groups patrilocality probably worked very well. Matrilocality (or uxorilocality) happen more commonly in larger agricultural societies, for example the Iroquois, but usually it is in conflict with the patrilocal model. Ambilocality also exists but I suspect that there was a tendency for patrilocality to resurface once and again, as I said: notably in the northern areas, where the signatures are very apparent and can be dated to the very beginnings of the Upper Paleolithic (the Native American case: with "Western" Y-DNA Q but mostly "Eastern" mtDNA such as B).
The most widely cited figures regarding Japanese Y-DNA seem to contain overestimates of the frequencies of haplogroups C and O3 because of oversampling of males from sparsely populated rural areas. IMHO, the Japanese Y-DNA data published by Dr. Hammer's team in 2005/2006 are cited far too often to the exclusion of other sources of data regarding Japanese Y-DNA. The "approximately 5% C1-M8" figure for Japan is based on Hammer's data set, which includes a sample of 70 Japanese males from Tokushima, Japan, 7 of whom (7/70 = 10%) have been determined to belong to haplogroup C1-M8. However, Tokushima Prefecture, which is located on the eastern side of the island of Shikoku, contains only 809,950 persons (0.63% of the total population of Japan) according to the 2005 Population Census of Japan.
ReplyDeleteIn contrast, Nonaka et al. (2007) have found Y-DNA haplogroup C1 in only 1.5% (2/137) of their sample of Japanese males from the Kanto region, which contains approximately 42 million persons (about one third of Japan's total population), including 1/13 males from Saitama Prefecture and 1/45 males from Chiba Prefecture:
Kanto region total (Nonaka et al. 2007)
2/137 = 1.5% C1-M105
3/137 = 2.2% C3-M217(xC3a-M93, C3b-P39, C3c-M48/M77/M86)
(5/137 = 3.6% C-M130 total)
11/137 = 8.0% D2-M55(xD2a-M116a)
18/137 = 13.1% D2a-M116a(xD2a1-M125, D2a2-M151)
2/137 = 1.5% D2a1-M125(xD2a1b-JST022457)
35/137 = 25.5% D2a1b-022457
(66/137 = 48.2% D2-M55 total)
3/137 = 2.2% O1a-M119(xO1a1a-M101, O1a2-M50)
9/137 = 6.6% O2b-SRY465(xO2b1-47z)
33/137 = 24.1% O2b1-47z
(42/137 = 30.7% O2b-SRY465 total)
2/137 = 1.5% O3-M122(xO3a1-M121, O3a2-M164, O3a3-JST021354, O3a4-JST002611)
3/137 = 2.2% O3a4-JST002611
5/137 = 3.6% O3a3-JST021354(xO3a3b-M7, O3a3c-M134)
4/137 = 2.9% O3a3c-M134(xO3a3c1-M117)
6/137 = 4.4% O3a3c1-M117(xO3a3c1a-M162)
(20/137 = 14.6% O3-M122 total)
1/137 = 0.7% Q1a1-M120
It seems likely that Hammer's team has overestimated the frequency of haplogroup C1 in the Japanese population because of their poor sampling strategy. However, that does not change the fact that haplogroup C1 seems to possess about the same degree of internal variance as its sister haplogroup, C3-M217.
According to the data I have at my disposal, Y-DNA haplogroup D2 is found with greater-than-average frequency in densely populated, highly urbanized, comparatively wealthy parts of Japan, whereas Y-DNA haplogroups C (both C1 and C3) and O3 are found with greater-than-average frequency in sparsely populated, mostly rural, comparatively poor parts of Japan. Haplogroup O2b seems to be distributed rather evenly throughout Japan.
An Asians Y Chromosome Hg C3*-M217 (Old East Asian) and Hg O3*-M122 (New Asian) have a different an ancient ancestors!
DeleteOnly in Y-DNA, in autosomal DNA or mtDNA the differences are usually very minor, if they exist at all. This situation, when Y-DNA lineages show sharp differences while other ancestral markers do not, can be explained by long-lived patrilocality, I presume. An example can be the Finnish (or other West Uralics), whose Y-DNA (most of it) links them to East Asia but otherwise are almost pure Europeans.
DeleteDo you mean mostly an Uralic and Finnish (men) have an Asians Y Hg N*-M231 but the Uralic and Finnish Autosomal DNA and their mtDNA Haplogroups are typically an East European / Western Eurasians. It's sound plausible and make sense. Thanks, Maju for your information.
DeleteSpecifically N1 (subclades N1c in Finnish and other Western Uralics, N1b in Siberian ones most commonly). The reverse case is that of Native Americans, whose precursors carried a Western lineage (Q1, Q is most basally diverse towards Iran, although less common) prior to the LGM to NE Asia, mixing there with the "core" East Asian peoples that had arrived earlier from the south. This is well documented nowadays both archaeologically (they carried the "mode 4" or blade tech to East Asia) and genetically (the famous Mal'ta boy and his autosomal affinities to Native Americans but not mainline East Asians like Chinese). So what we see in North Eurasia is, first, a West→East process of migration and admixture, producing Native Americans primarily, and later (after the LGM), an East→West process of migration and admixture producing Uralic peoples in essence. Naturally both migrations left a more diluted legacy in other populations but there is still a sharp contrast between the Far North and the rest.
DeleteSee for example: http://forwhattheywereweare.blogspot.com/2013/07/a-review-of-haplogroup-n-y-dna.html
PS- See also: http://forwhattheywereweare.blogspot.com/2013/12/the-malta-adna-findings.html
DeleteDear Maju,
DeleteIn Far East Asians "case", even we know when a Modern East Asian descendant, more specifically a Northern Chinese today are a results of an intermarriage between a NE Asians and a SE Asians ethnic group, but if we don't use a "Specific Marker" like DNA Haplogroups who can't mix with another DNA like a Y Chromosome DNA and even a non Chromosome DNA like an mtDNA, they're can't trace our ancient ancestry geographical origins. Even though a Southeast Asians like a Western Indonesian, Vietnamese, Thai, etc share a similar Autosomal Regional DNA / Phenotype with a Northeast Asians like a Southern Siberian, Mongolian, Japanese and Korean people today but i can't deny when a Mongolian, "Altaian?", Tungusic and a Maternal (mtDNA) Japanese, Korean and Northern Chinese are quite different with a Southern Chinese, Austroasiatic and Austronesian DNA Haplogroups. Maybe a Human "Haplogroup DNA" marker are older than a Phenotype DNA markers who a Paleoanthropology Scientist classified a Human "Phenotype" Autosomal DNA as a "Human Race" like Caucasoid, Mongoloid, Negroid, etc....But, honesly i more like with my Chinese and Austroasiatic Autosomal DNA with a Southern Haplogroups Y Hg O*M175 and mtDNA Hg R, B and F rather than a little "European" looking like a Tatar Russian, Kazakhstan, "Altaian" and an Uyghur people with a Y Hg C*-M130, C2*-M217 + Y Hg D*-M174, D2*-M55 and mtDNA Hg CZ, C, Z, D and G. I think when I'm and mostly a Southern Chinese and a Southeast Asians don't like to have a "Genghis Khan" Y Hg C3-M217 / C2, "Nurhachi" Y Hg C3c-M48 and a Tibetan Y Hg D1, an Ainuid Y Hg D2-M55. Or a NE Asians mtDNA Hg M Type, However........
I'm not really sure what you're talking about (too long sentences with lots of compressed info in them) but Han Chinese (of all localities, N or S) and Japanese/Korean also tend to cluster both in terms genetic and phenotypic (skull measures) with SE Asians. NE Asians do not or only up to some point.
DeleteNE Asians are each one their in their own micro-cluster (in skull measurements) probably because they represent unique local solutions to the equations of admixture between various sources of genetic input, of which the East Asian one is important but just one. This is more apparent in the autosomal (general) genetic data, where Siberians clearly form a V-shaped cline between: (1) East Asians and ancient Mal'ta and (2) East Asians and Native Americans. Both the Mal'ta boy and modern Native Americans represent two extremes of the original Siberian UP, one more admixed with West Eurasians (Gravettian culture, which expanded via Europe) and the other heavily admixed with East Asians on their way to America.
Y-DNA O3 was not found in North China in the Neolithic, it seems (see: http://forwhattheywereweare.blogspot.com/2013/12/ancient-east-asian-y-dna-maps.html ). Most of what was found (NE) was N, some C and some O(xO1,O2,O3). O3 dominated Central-South China (Neolithic core) and O1 dominated East China (Shanghai area). In the Metal Ages' data we see however that C and Q were found in the inland parts of North China, while C had become by then more important in NE China and O3 had apparently expanded northwards. "Finnish" N1c was found in Manchuria in this period too, while large frequencies of O2 were found in the SE (Canton area).
So to my eyes, O (incl. O3) seems mostly Southernly, while NE Asia was dominated originally by C, Q1 and N with at most some O* (that is now nowhere to be found AFAIK). In the Neolithic and post-Neolithic periods there was a S→N expansion primarily that altered the demographics of North China and the East Asian coastlands (W→E expansion within "China proper").
For sake of clarity, I should note that Hammer's sample of 70 males from Tokushima is the largest of all his regional Japanese samples, amounting to 27% of his pool of 259 males from Japan. The Tokushima sample is ridiculously large when considered in light of the demographic insignificance of the actual population of Tokushima, which accounts for less than two thirds of one percent of the total population of Japan.
ReplyDelete"Well, I find it quite believable because that's what the data tells".
ReplyDeleteOnly if you choose to interpret the data a particular way.
"Similarly you could say that it's unbelievable that M could have traveled all the way to South Asia not leaving the slightest trace... but that's what happened'.
But surely it did leave a trace. M1 basically spreads all the way between India and Africa.
"We see no pre-M nor pre-N anywhere except at the L3 node. That's what we know, believe it or not".
And that L3 node is firmly planted in Africa.
"Seems demonstrated by archaeology and simulations".
But not by the distribution of basal M haplogroups.
"It seems likely that Hammer's team has overestimated the frequency of haplogroup C1 in the Japanese population because of their poor sampling strategy".
That seems to be the case. By the way, good to hear from you again Ebizur.
"However, that does not change the fact that haplogroup C1 seems to possess about the same degree of internal variance as its sister haplogroup, C3-M217".
To me it seems obvious that C1 was in Japan long before C3 arrived there. Perhaps it came from SE Asia, but I think it is at least as likely that it didn't.
"Y-DNA haplogroup D2 is found with greater-than-average frequency in densely populated, highly urbanized, comparatively wealthy parts of Japan"
That's rather difficult to explain.
"Haplogroup O2b seems to be distributed rather evenly throughout Japan".
I strongly suspect that O2a's origin is somewhere on the mainland near Japan. And its expansion is a probably a product of the Chinese Neolithic.
@Terry:
ReplyDelete"But surely it did leave a trace. M1 basically spreads all the way between India and Africa".
No: this is a back-migrant lineage very clearly as it belongs to M1'51 which is half South Asian and because M1 is highly derived (long stem) in relation to her M and M1'51 maternal ancestors.
...
@Ebizur:
Would you have any suggestions about the possible main colonization routes in East Asia (or anywhere else)? I am uncertain about that coastal route as only main route and I wonder if any particular lineages look like having complemented it through other routes across the interior of today's China or even "broad fronts"... any idea?
"I am uncertain about that coastal route as only main route"
ReplyDeleteI'm uncertain about the coastal route being any route into SE or East Asia.
"I wonder if any particular lineages look like having complemented it through other routes across the interior of today's China or even 'broad fronts'... any idea?"
M-derived mtDNA haplogroups D and M8/CZ almost certainly moved overland into Southern China from Northeast India. The same goes for the other East Asian haplogroups M7, M9/E, M10 and M12/G.
The same for many of the SE Asian haplogroups. For example M21, M22 and M72. The others obviously crossed the sea to reach where they are now but almost certainly from the nearby mainland, not as a product of sustained movement along the coast.
So we finish up with the offshore haplogroups M14, M15, M17, M23'57, M27, M28, M29'Q, M31, M32'56, M42'74, M47, M71 and M73.
"this is a back-migrant lineage very clearly as it belongs to M1'51 which is half South Asian and because M1 is highly derived (long stem) in relation to her M and M1'51 maternal ancestors".
But M1 and M51 split after just one mutation and then both have a long tail. M1 a tail of four mutations and M51 a longer six mutation tail. So M51 may be an immigrant into India.
I should clarify. I don't believe for a moment that M1 is other than a back-migration to Africa. The options for its origin are not just restricted to either Africa or India though. From Wikipedia, everybody's favourite reference:
ReplyDelete"M1 is not restricted to Africa. It is relatively common in the Mediterranean, peaking in Iberia. M1 also enjoys a well-established presence in the Middle East, from the South of the Arabian Peninsula to Anatolia and from the Levant to Iran. In addition, M1 haplotypes have occasionally been observed in the Caucasus and the Trans Caucasus, and without any accompanying L lineages.[3][9] M1 has also been detected in Central Asia, seemingly reaching as far as Tibet".
So Anatolia, the Levant or Iran are the most likely possibilities. From where it would have gone both into India and into North and East Africa.
"So M51 may be an immigrant into India".
ReplyDeleteNot parsimonious.
"almost certainly moved overland into Southern China from Northeast India"...
I asked specifically Ebizur and not you for a reason, Terry.
You may find this link about Nepal interesting:
ReplyDeletehttp://www.biomedcentral.com/1471-2148/9/154
"I asked specifically Ebizur and not you for a reason, Terry".
ReplyDeleteAnd I'm sure he will be able to provide us with valuable information about Y haplogroup distribution. But we have to consider mitochondrial DNA haplogroups distribution also.
I've just noticed something else interesting about mtDNA haplogroup M. As you say, its diversity is starlike. As is that of two of its basal haplogroups: M3 and M47. Perhaps this indicates two separate expansion centres for M.
ReplyDeleteAs I see it the evidence very strongly argues in favour of mtDNA M having entered India north of the Thar Desert, via the Punjab not via the coast. From the Punjab M branched into two separate migrations. Several branches entered Madhya Pradesh and Maharashtra. There the first of the above haplogroups (M3) diversified immediately, and from there members of M spread east to Bihar. But that is not the end of the story.
The paper on Nepal claimed that the southern forested fringe of Nepal was probably uninhabited until the Tharu arrived. But further north the Siwalik Range, along the northern edge of the Gangetic Plain, has a rich and long human and pre-human presence. The hills stretch west, all the way to the Punjab, and east, perhaps as far as Zomia.
The southern face of the range provides an obvious route east for MtDNA M. My bet is that most, if not all, of what were to become the East and Southeast Asian M clades used this route east. In other words all the haplogroups on page 1 of my list and the Northeast Indian branches from page 2. That the movement along the narrow corridor was rapid is indicated by the fact that it included M51, and in Southeast Asia the second of the above basal haplogroups (M47) diversified immediately.
Next thing was, across the Lower Ganges somewhere, a mixing of the two M branches followed by a movement back west, which included several haplogroups that had coalesced around the Burma/Assam border region, having arrived there via the northern route through Nepal. It was at this stage that M entered Orissa and Andhra Pradesh, possibly as a result of a 'coastal migration'. The Andaman Islands were possibly also settled at this time.
A similar mixture was also able to enter South India and Gujarat over land. By this stage M4''64 had joined in from somewhere, probably Central India although M43 is found in Assam.
Check it all out. Fits all the evidence perfectly.
Per PhyloTree: M3 has only three basal sublineages (a, b and c), while M47 is not indicated a single sublineage (and stands as of now as merely "proposed" lineage, with all mutation sites in italics).
ReplyDeleteThe only clear star-like structure at the second level of M is M4"64, with 8 basal sublineages (up to 12 if we consider only coding region mutations).
By my criterion of 5 basal sublineages or more, there is no other star-like structure so high under M. So M exploded, M4"64 exploded less dramatically after it but all other sublineages show only gradual expansion patterns, not demographic explosions of any sort (at least not until you move a lot downstream, like D4, C or Z1'2'3'4'7, all them belonging to the NE Asian demographic scene).
"while M47 is not indicated a single sublineage"
ReplyDeleteMy mistake. I checked and found that to be so. Just no mutations from basal M. It is proposed as a Southeast Asian lineage though.
"M3 has only three basal sublineages (a, b and c)"
M3 seems quite confined in its spread though, Central and South India.
"all them belonging to the NE Asian demographic scene"
Very much associated with the expansion into the previously uninhabited regions of Northern Asia and America. The problem is what is the pattern of expansion into India. We certainly seem to see a Central Indian/Northeast Indian division.