Not long ago I predicted that Ust'-Ishim belonged to a basal form of Y-chromosome haplogroup P (see here). As it turns out, the 45,000 year-old western Siberian genome belongs to K(xLT) or K-M526, which is actually pretty close to my guess. The Ust'-Ishim paper was published today and is sitting behind a paywall here, but the extensive supp info is free.
Here's a map to help visualize the information, featuring Ust'-Ishim as well as Mal'ta boy, another North Eurasian Upper Paleolithic genome published recently.
The Ust'-Ishim genome was sequenced from the fossil of a femur bone found on the right bank of the Irtysh River. This area is very close to the Urals, and almost in the middle of the former Mammoth steppe that once stretched across North Eurasia from Iberia to Alaska. Interestingly, M526 is an ancestral mutation to the markers that define Y-chromosome haplogroups N, Q and R, which possibly dominated North Eurasia since the Upper Paleolithic (note that the 24,000 year-old Mal'ta boy belongs to a basal form of R).
Moreover, R1a and R1b are the most frequent haplogroups in Europe today. Thus, it would seem that most European males derive their paternal ancestry from North Eurasian hunter-gatherers whose ancestors spread out across Eurasia from the Middle East over 45,000 years ago.
I know that a lot of people have been arguing recently that K-M526 and the derived P-M45 originated and diversified in Southeast Asia, and then migrated north well within the last 45,000 years (for instance, see here). However, considering that K-M526 was already in reindeer country 45,000 years ago, as well as the Denisovan (ancient Siberian hominin) admixture among Southeast Asians, that might well turn out to be the equivalent of arguing that up is down and down is up.
By the way, Ust'-Ishim also belongs to pan-Eurasian mitochondrial (mtDNA) haplogroup R*, and in terms of genome-wide genetic structure appears roughly intermediate between West and East Eurasians. These outcomes fit very nicely with its Y-haplogroup.
However, it's slightly closer to Mesolithic Iberian genome La Brana-1, Upper Paleolithic Siberian MA-1 (or Mal'ta boy), and present-day East Asians, than to present-day West Eurasians, including Europeans. That's because it lacks "ancestry from a population that did not participate in the initial dispersals of modern humans into Europe and Asia". This is obviously the so called Basal Eurasian admixture discussed in Lazaridis et al. (see here), which is probably associated with early Neolithic farmers.
Also worth mentioning is that Ust'-Ishim harbors longer stretches of Neanderthal chromosomal segments than present-day Eurasians, which suggests that admixture between modern humans and Neanderthals took place in the Middle East not long before the ancestors of Ust-Ishim moved into Siberia (50-60,000 years ago). But this was already covered months ago, and you'll find lots of links on the topic on Google.
Citation...
Qiaomei Fu et al., Genome sequence of a 45,000-year-old modern human from western Siberia, Nature 514, 445–449 (23 October 2014) doi:10.1038/nature13810
See also...
Y-haplogroup P1 in Pleistocene Siberia (Sikora et al. 2018 preprint)
Sweet! Maybe this throws out the K2b developing in SE Asia, mess. He could likely be an ancestor to all of us with R and Q. Any chance someone on here wants to look for anything K2b and under?
ReplyDeleteFrom this genome, they can say that admixture of the ancestors of the major Eurasian population with Neanderthals occured 50-60k years ago (and perhaps some more recently as well). It also supports that there was a Basal population in the middle-east or North Africa which split from the rest of the out-of-Africans at least 70k years ago. In this case, they Basal group would not have any Neanderthal DNA.
ReplyDeleteSo, it should be pretty easy to tell from early farmer genomes whether or not the Basal group had also mixed with Neanderthals. Do EEF genomes have only half the Neanderthal admixture?
What's the f4 formula for testing for Neanderthal admix in West Eurasians? I can try it later today with the Human Origins samples.
ReplyDeleteDavidski, Hinxton5 gnome is available!
ReplyDeleteYes, I know, I'm running it now. I'll post the results tomorrow.
ReplyDeleteContinuation of previous comments thread...
ReplyDelete@ Ryukendo.
It seems like Ust-Ishim was a basal member of the WHG-ANE-ENA clade, or a Basal crown Eurasian, in other words.
I won't be surprised if there were further radiations from this clade that did not survive, and we are only seeing the tips of the three branches that did.
Yeah, Ust Ishim Looks “Basal Eurasian” (more or less equally related to present day Eurasians, lower separated from Africa as ascertained on modern axes) its with a slight tilt towards South East Eurasians (East Asian and Oceanian), specifically slightly away from Native American. Same story on ADMIXTURE.
But apparently this is non significant under most models, with only a few placing Ust Ishim with Eastern Non-Africans (after MA-1 splits from them), and direct statistics indicate approximately equal relatedness to all modern day non-West Eurasians.
Ust Ishim is more heterozygous than modern Eurasians, including Europeans who seem to be BE admixed . Makes you wonder about that pattern of collapsing heterozygosity from Africa, and low heterozygosity in other ancient hunter gatherer samples – clearly there has to be something else going on.
As a population which seems to fit a position more basal than the WHG-ANE-ENA divergence and less basal than "Basal Eurasian", Ust-Ishim seems kind of interesting to me in the context of South Asian admixture, in case there is some ancestry there which diverges earlier than ENA-(WHG-ANE) but after BE-(ENA-(WHG-ANE)). Maybe there could be something interesting in statistics for MA1-Ust_Ishim in South Asia.
@ Michael It also supports that there was a Basal population in the middle-east or North Africa which split from the rest of the out-of-Africans at least 70k years ago. In this case, they Basal group would not have any Neanderthal DNA.
At least in the Lazaridis paper the levels of sharing between the ancient samples (Stuttgart, Loschbour, MA1) and present day Eurasians was "statistically indistinguishable".
Lazaridis tested this for non-African populations with
f4 (Altai, Denisova;Test, Yoruba)/f4(Altai, Denisova, Vindija, Yoruba).
This paper used f4(Denisova, Altai; Africa, Test)/f4(Denisova, Altai;Africa, Mezmaiskaya Neanderthal).
@Matt
ReplyDeleteI was wondering exactly the same thing! He seems to score very high in kalash/South Indian (there is only one component for South Asia, peaks in kalash and extends over to Cambodia) and papuan. This is a weird result were he to be equally related to everyone in Eurasia. This is very little to go by in any case.
Agree that f Stats with him and many more pops would be interesting. Why formal tests were conducted on so few pops, and why papuan were excluded I have no idea.
@All
I can see all papers on all major publications for free. If you want info from any paper just holler.
I am wondering. What are the distances from MA-1, to Loscbour, and an ENA population. How close is he to being equally distant?
ReplyDeleteIs there a possibility that ANE is nothing more than Ust-Ishim, with some WHG/Gravettian?
It would be interesting to test, on treemix, if it cannot be ruled out.
ReplyDeleteMA-1 and Loschbour are definitely equidistant from ENA (Onge and Han). That was shown in Lazaridis et al.
ReplyDeleteThis paper shows that MA-1 and La Brana-1 are equidistant from Ust-Ishim. That's covered on page 67 of the supp info.
It appears that Ust-Ishim could easily be the ancestor of both or either MA-1 and La Brana-1, because...
"Specifically, we find that none of these samples shares significantly more alleles with Ust-Ishim than any other (all |Z|-scores in Table S11.1 and Table S11.2 have values of <2). This finding is consistent with Ust-Ishim having separated from other Eurasians around the time of the ancient divergence of West and North Eurasian hunter-gatherers (represented by La Braña and MA1), and east Eurasians (represented by Onge, Karitiana and Han). The absence of evidence for shared genetic drift with any of these groups suggests that Ust-Ishim is likely to be close to the ancestor of most Eurasians, and is consistent with the old radiocarbon date."
Nice. Is there any way to predict the date of separation of ANE and WHG? Obviously it's between 24-45kya.
ReplyDelete1. I knew that Ust-Ishim had R mtdna '' Ust Ishim had mtdna hg R of some sort and that is verified. '' June 10th 2014.
ReplyDelete2. I predicted that Ust'Ishim y-dna haplogroup will end of being one of the following.
a. We will never know he was to old to get DNA from.(PROBABLY THE CASE, BUT BELOW ARE THE HYPOS)
b. F(XGHIJK )
c. PRE C(XC2' AND C1)
d. PRE D
5. something else and push y-dna phylogeny back.
I was correct about one thing it is now clear that early population with haplogroup K were dominated by mtdna haplogroup R or at least N, and not M. Ust-Ishim also is strong evidence against a southeast asian origin of K-M526, and P as a whole rather it seems an extinction event took place around 47,000 years ago and Southeast Asia was not effected as bad as most areas.
ReplyDeleteryukendo kendow,
ReplyDeleteCan you get acces to this paper and show the highlights/important information? Thanks in advance!
http://www.nature.com/ng/journal/v46/n8/full/ng.3021.html?WT.ec_id=NG-201408
RK This is a weird result were he to be equally related to everyone in Eurasia. This is very little to go by in any case.
ReplyDeleteI've had a read through Dienekes, John Hawks and Reich's comments and its quite an interesting picture.
(http://phys.org/news/2014-10-evidence-neanderthal.html
http://johnhawks.net/weblog/reviews/ancient-genomes/ust-ishim-fu-2014.html
http://dienekes.blogspot.co.uk/2014/10/high-coverage-genome-from-45000-year.html)
It seems like a lot of population the structure between Africans and Eurasians already existed at this time, and physical cranial and post-cranial modern form had been reached for quite some time. Although these people may not have been phenotypically very divergent.
(Not just Africa - there are even some old cranially modern fossils in China long before Ust-Ishim).
But Ust-Ishim apparently converges with a period of an expansion of toolmaking took place. It may have been at the same point as divergence of what you term “crown Eurasians” - actually a check on that is applying the Ust-Ishim derived mutation rate to divergence of WHG, ENA.
So it does make sense that the population could be equally related to many divergent non-African modern day people.
It also seems like most of the Neanderthal admixture had already happened, just relatively recently before Ust-Ishim (around 10,000 years before) and that subsequent events were more like 1/1000 than the 3/100 that seems to have seems approximately shared between BE and other Eurasians. A slow accumulation, which we would might see in East Asians and WHG with large enough sample sizes, but the breeding events were so rare they did not change the genome too much in terms of admixture levels between populations.
I think it might be the case that Ust Ishim's population ultimately led to nothing (as John Hawks suggests is very possible), as Ust Ishim man could just be part of a very early colonisation of climates that Homo Sapiens were not yet well adapted to, compared to other more physically adapted hominids (if even them).
Either way I think that Ust-Ishim might be able to act as a proxy for a basal group following the Middle Eastern basal group, even if its cousins left descendents and it didn't, and may give us useful statistics.
The problem might be that Ust-Ishim might tend to give statistics with a lot of populations, simply because it is so basal it can act close to a proxy for many populations. But then maybe you could trade off where its signal is particularly strong?
RK Agree that f Stats with him and many more pops would be interesting. Why formal tests were conducted on so few pops, and why papuan were excluded I have no idea.
I get the impression the reasons tend to be “complex admixture history”, but it still would've been interesting to see.
Ust-Ishim man is K2*, and not on the branch leading to either P or NO. The paper dates the breakup of K2 at 50 (47-55) kya, so about 5000 years before Ust-Ishim man. Plenty of time to get there from SE Asia - or from wherever K itself originally broke up (South Asia?), heading the other way.
ReplyDeleteEither way, his lineage no longer exists, unless as some obscure K* somewhere. So this doesn't get K2 out of Southeast Asia.
@ Matt
ReplyDeleteI'll check these out.
@ All
Could you guys tell me why this strikes a blow against the emergence of K in Southeast Asia?
The subclade within which PQR are nested, K2b, has all its most basal members confined to SEA, and PQR's only major sisters, M and S, are confined to--Papua New Guinea. P* are found in Aeta and other negritos in the Phillipines.
N and O, sister to K2, are both diversified and expanded from Southern China, while LT, also called K1, are found in India.
The evidence for an eastern origin of K, with so many basal K2b clades in Southeast Asia, is very strong; unless we want to propose a parallel migration of all groups eastwards followed by a clean geographical segregation of each subclade(is this what you are suggesting Chad?), the furthest west K could have originated should be around India. The current distribution of K and its descendants, MNOPQRS, also match the distribution of the ENA-WHG-ANE clade almost perfectly, with the addition of I and prob J, but definitely not G, F, E and etc.
It seems more likely that a pop-dense area in the southern tier of eurasia, incl. India and SEA, periodically radiated into northern Eurasia, with pops following undulating glacier rims/climactic zones, with subparts of the tier, e.g. India and SEA, drifting apart over time producing ANE-WHG vs ENA, until the most recent radiation produced us.
Furthermore, ANE is old in India, high even amongst Pulliyar, and the sister of WHG is ANE in Siberia. This makes a cent.Asian-Siberian route for the most recent population of Europe more likely, and the Middle Eastern route less likely, because WHG would likely have encountered Basal in the latter and this is simply not the case.
@ Tesmos
ReplyDeleteI assume you are interested in the pop gen history?
Anyway, the analysis they did on pop gen hist was not very extensive, mostly IBD. Dutch from every province shared more IBD with north Dutch than was expected, making the authors conclude that the depop+exmigration from of the north with rising sea levels from 5000BC-50CE drove this dynamic.
Secondly, Dutch from every province shared more with Utrecht and central Dutch than expected, for obvious reasons.
capra internetensis,
ReplyDeleteAt some point you might understand the implications of finding K2* so far from southeast Asia at such an early date, in a genome that has recent Neanderthal admixture (presumably from the Middle East) but lacks Denisovan admixture.
@ Davidski
ReplyDeleteI've been wondering the same. It seems like the ENA-WHG-ANE pop divided early on into various ENA clades with high Denisova/a trace of Denisova, or those with no Denisova. Razib placed K origin North of Bay of Bengal for that reason, as K there would have split in that manner.
Either way the mixture of Neanderthal into Ust-ishim dates back, according to the authors, to the same pulse that introduced Neanderthal to everyone else.
Apparently the Neanderthal admixture took place in the Middle East, and then very soon after Ust-Ishim's ancestors moved to Siberia.
ReplyDeleteBut somehow many layers of K ended up in Southeast Asia, along with Denisovan (Siberian hominin) admixture.
That looks like a case for multiple migrations from Siberia to Southeast Asia, not the other way around.
Ryukendo kendow,
ReplyDeleteYes, i am interested in the pop gen history and thank you very much for checking it out and the info.
@ Davidski
ReplyDeleteThe authors did all their work 'under the simplifying assumption that the gene flow occurred as a single event.'
In any case, for the pool of K chromosomes to be distributed so widely that NO, sister of K2b emerged in South China and LT/K1 in India, and then P to originate in a pool of K2b chromosomes in Central Asia/Siberia, and for P* (and no other P), M+S then to end up in Phillipine negritos and Papuans respectively, and then NO splits the distribution of K2b in half later is a singularly unparsimonious theory.
Far more parsimonious that K was confined to a small area and expanded in a star-like fashion, resulting in the chaotic jumble we see near the epicenter today, while descendants of K occupy other continent-sized areas with great homogeneity.
Davidski,
ReplyDeletegoing by Reich 2011, Denisovan admixture seems to occur almost entirely east of Wallace's line. Western Indonesians and even Malaysian negritos don't have it. So the furthest east branches of K2 picked it up, the rest did not.
It is not much to ask for one branch of K2 to head north and die out, while most headed south. All evidence still has P originating in SE Asia.
But you're talking about modern DNA, which is usually difficult to read.
ReplyDeleteAlso, simple models are the best, at least to start off with, and the fact that we now have two genomes from the Mammoth Steppe with K-M526 (Ust-Ishim and Mal'ta boy), that lack Basal Eurasian and Denisovan admixture, suggests genetic continuity on the Mammoth Steppe since its colonization very shortly after the admixture event between humans and Neanderthals, presumably in the Middle East.
So the case for Ust-Ishim and Mal'ta boy being the descendants of migrants from southeast Asia doesn't look persuasive to me.
capra internetensis,
ReplyDeleteNew work shows Denisovan admixture is pervasive across south and east Asia.
Which Malaysian negritos are you talking about, because the Onge do have it.
Page 97 from the supp info...
ReplyDelete"Empirically, we observe that the ratio is higher in East Asian and Native populations than in Europeans (Table S17.1). A Wilcoxon signed-rank test comparing the ratios for European populations and Ust’-Ishim to the ratios for the East Asian and Native American populations shows that the ratios for Europeans and Ust’-Ishim are significantly smaller than the ratios for East Asian + Native American populations (P = 0.0028). This likely reflects the small amount of admixture from Denisovans that has been reported for eastern non-Africans like East Asians and Native Americans 1,2. The ratio for Ust’-Ishim is within that for Europeans, indicating that the Ust’-Ishim individual may not share the additional admixture with Denisovans that occurred in the population ancestral to East Asians and Native Americans."
Ust-Ishim is similar to modern Europeans in terms of Denisovan ancestry, despite the fact that Europeans have Basal Eurasian ancestry, which would certainly dilute any input from the east.
"Maybe this throws out the K2b developing in SE Asia, mess".
ReplyDeleteI don't think so. All it does is push the dating for K2 in SE Asia back to before 45,000 years ago.
"M526 is an ancestral mutation to the markers that define Y-chromosome haplogroups N, Q and R, which today dominate North Eurasia and the Americas".
It is also an ancestral mutation to the markers that define Y-chromosome haplogroups SE Asian K2b, K2c and K2d. As you are no doubt aware N and O can be considered K2a. And when it comes down to it Q and R are can be considered K2b2, downstream of K2b-P331. K2b1-P397 is virtually confined to east of Wallace's Line. Of course that doesn't mean at all that Ust-Ishim has any significant SE Asian genetic component overall. By the time K2b2 entered the steppe it would have become considerably admixed with populations it met along the way. Not to mention populations already present on the steppe margin.
"It seems like a lot of population the structure between Africans and Eurasians already existed at this time, and physical cranial and post-cranial modern form had been reached for quite some time".
Exactly.
"rather it seems an extinction event took place around 47,000 years ago and Southeast Asia was not effected as bad as most areas".
That doesn't really fit unless the phylogeny is drastically overhauled in the future.
"The evidence for an eastern origin of K, with so many basal K2b clades in Southeast Asia, is very strong; unless we want to propose a parallel migration of all groups eastwards followed by a clean geographical segregation of each subclade(is this what you are suggesting Chad?)"
Quite.
"Far more parsimonious that K was confined to a small area and expanded in a star-like fashion, resulting in the chaotic jumble we see near the epicenter today, while descendants of K occupy other continent-sized areas with great homogeneity".
I agree that is what seems most likely to have happened and is the best explanation of the evidence.
@ Davidski
ReplyDeletemt Hap R's descendants share an extremely similar distr as K's descendants, and similarly has an eastern center of gravity, this time in India. This is the hap of Ust Ishim as well, in this case a rather basal variant.
While one can make the case that continuity existed in Siberia, the fact is that Ust-ishim is autosomally marginally more similar to Karitiana than to even ancient West Eurasian, more similar to Han than to Karitiana, and most similar to Onge. Qiaomei fu even postulated an early migration in the supp due to this. If we are to envision a case of continual genetic drift to finally produce the ANE-WHG clade in West Eurasia, we have to ask why we get such a weird result here.
Unfortunately, sensible interpretations of modern DNA often lead to conclusions that have nothing to do with reality.
ReplyDeleteI think this will also turn out to be such a case.
I can't really see a way you can have multiple waves of migration from Siberia to Southeast Asia, these Siberians being ANE like, yet Southeast Asia is symmetrical in relatedness to MA-1 and La Brana/Loschbour (and all of East Asia really, except what seem like apparently low levels of similarity in populations which border Siberia).
ReplyDeleteIt's hard to see how you can have any population flow from Siberia in any large quantity to where the Dai live if the Dai are symmetrical relationships to MA-1 and La Brana/Loschbour.
Certainly any such dynamic was reversed (at least from Out of China if it's not mainland Southeast Asia) before Native Americans were on the scene as a population. Whether or not there was any population continuity in North/Inner Asia, it certainly doesn't seem to act as a source to East Eurasia. The ANE people pretty seem like they were kept in / stayed in their box for the most part, when it comes to East Eurasia.
There's not much evidence linking Ust Ishim to any later groups in North / Central Eurasia or Southeast Asia. A single y dna haplogroup seems not very substantial, even if you accept that its origin was where it was found.
If anyone has access to wiki link
ReplyDeletePrüfer, K.; Racimo, F.; Patterson, N.; Jay, F.; Sankararaman, S.; Sawyer, S.; et al. (2014) [Online 2013]. "The complete genome sequence of a Neanderthal from the Altai Mountains". Nature 505 (7481): 43–49. doi:10.1038/nature12886.
"They also found that the Neanderthal component in non-African modern humans was more related to the Mezmaiskaya Neanderthal (Caucasus) than to the Altai Neanderthal (Siberia) or the Vindija Neanderthals (Croatia)"
Mezmaiskaya Cave/Maykop region.
n Davidski,Onge doesn't have any Denisovan admixture.
ReplyDelete" Here, we quantify Denisova admixture in 33 additional populations from Asia and Oceania. Aboriginal Australians, Near Oceanians, Polynesians, Fijians, east Indonesians, and Mamanwa (a "Negrito" group from the Philippines) have all inherited genetic material from Denisovans, but mainland East Asians, western Indonesians, Jehai (a Negrito group from Malaysia), and Onge (a Negrito group from the Andaman Islands) have not".
http://www.ncbi.nlm.nih.gov/pubmed/21944045
That might be due to the fact that Onge represents only one Y haplogroup, the haplogroup D. A
That's an old study based on genotypes. Onge do have Denisovan admix which can be seen when looking at full genome sequences.
ReplyDeletehttp://www.ashg.org/2014meeting/abstracts/fulltext/f140121559.htm
So, widespread Denisovan admixture in south and southeast Asia, and into the Americas.
Therefore, I ask again, why is Ust-Ishim on a par with modern Europeans in this regard, and where did his recent, long chunks of Neanderthal ancestry come from if he was a Southeast Asian?
"Therefore, I ask again, why is Ust-Ishim on a par with modern Europeans in this regard, and where did his recent, long chunks of Neanderthal ancestry come from if he was a Southeast Asian?"
ReplyDeleteThe simple answer to that is that only his Y-DNA (and possibly his mt-DNA) come from SE Asia. The bulk of his genes are more ancient Central Asian.
@ Davidski
ReplyDeleteThanks for bringing this up. Well, Denisovan in South Asia kinda changes things.
But this does not make your theory any more improbable. This just makes both theories equally bad at explaining everything in a parsimonious manner.
Only a series of very unlikely events could result in the distr we see today if K or K2 originated in Siberia. The sister of K2, NO, must migrate to South China, then three daughters of K2: P*, M and S, arrive in PNG and segregate themselves cleanly into Negritos and Papuans, while somehow avoiding East Asia and India and without bringing East Asian NO or Indian groups into Papua. Secondly, it would have to have all this diversity wiped completely clean from every area except Wallacea. Lastly, a giant pop expansion from... Siberia? Everyone is a HG at this point, so there is simply no ecological or technological reason for this; demic diffusion from a subarctic environment seems just unlikely.
Also, modern ENA-WHG-ANE are knotted at the south and distinct at the north. In siberia east and west eurasians did not meet until the formation of native americans as a pop, while in South Eurasia all major eurasian groups are represented and grade into each other via less derived forms. East Asians especially seem to get their start as an ASI-like pop that then expanded into China+Siberia from MSEA, before the neolithic exp of Mongoloids limited these basal East Eurasian to Negrito pops in Sundaland. The same seems to be the case for ANE in South Asia, which seems very old.
Perhaps Denisovan Admixture occurred in Eastern India, and West Eurasian groups came out from West India, with Y hap K, or maybe IJK, being the signature for this ur-Indian population.
I no longer have any faith in interpretations of modern Y-haplogroup data after the R1a1a from India debacle. That was apparently a done deal, and look what happened.
ReplyDeleteBased on the ancient DNA evidence I've seen to date, this issue looks rather straightforward to me at the moment:
- Ust-Ishim is at the base of the Eurasian phylogenetic tree and closely related to the (non-Basal Eurasian nor Denisovan) ancestors of all Eurasians.
- It lacks Denisovan admixture which is found all over South and East Asia and the Americas.
- It shows very recent Neanderthal admixture, probably from the northern Middle East, which also shows up in all South and East Asians.
- Therefore Ust-Ishim's people migrated south and east from Siberia and/or Central Asia to populate other parts of Eurasia.
Let's see how this stands up when we get more Upper Paleolithic genomes from across Eurasia.
So, probably still not understanding it right but...
ReplyDeleteif Usty has the same neanderthal admix as everyone but Basal Eurasian might that not mean a second out of Africa along the Atlantic coast which missed out on the neanderthal sex party in Egypt/Arabia?
"Therefore Ust-Ishim's people migrated south and east from Siberia and/or Central Asia to populate other parts of Eurasia."
ReplyDeleteThis basically leaves two options.
The first option: these haplotypes originated in Siberia, and people closely related to "Usty" moved from Siberia to Southeast Asia.
The second option: The original haplotype defining mutations may have originated near the middle-east. Then, there was some kind of super autosomal mutation, or cultural advance, or climatic change, or some other unknown circumstance that arose and rapidly allowed these people to take over Eurasia. They spread in every direction, with bottlenecks along every path. Usty was a dead end that shared a recent common ancestor with a group that settled successfully in Southeast Asia. Some of these people mixed with Denisovans as they rapidly migrated. There were numerous founder effects and maybe Denisovans even inhabited the entire path to Australia and New Guinea.
Probably there were numbers of waves of migration not from within Africa, but from just a bit out of Africa, for a limited amount of time.
The Basal Farmer people were from a different group that became isolated soon after the initial mixing with Neanderthals. They lived in the Levant and only expanded because they invented agriculture. Otherwise, they would have disappeared along with many other branches of the human family tree.
"
ReplyDeleteBlogger Grey said...
So, probably still not understanding it right but..."
Davidski is pointing to a very powerful observation that the current migration does not match [if at all]very well with incomming data Ust-Ishim and Mal'ta (MA-1)
Trying to explain archaic human migration in a linear and parsimonious manner is not working out quite as expected. Mother nature has a few surprises. Introgressive hybridization between our ancestors,Neaderthals/ Denisova hominins and perhaps some possible undiscovered unkowns that we cannot imagine until we get more ancient data ?
@ Davidski
ReplyDeleteThis is my last comment on the subject. Apologies for hogging airtime.
The issue is hardly 'simple'. The argument you make is analogous to saying that 'The only ancient DNA tells us that Sima de los huesos is ancestral to Denisova. Therefore there must have been a movement from Iberia to East and SE Asia by Denisova ancestors. The modern distribution in Asia, in introgression and whatnot, cannot be trusted.' A 'simple' and hardly plausible statement.
The fact is, the authors stated that they did all their work 'under the simplifying assumption that the (Neanderthal) gene flow occurred as a single event.' If their assumption is true, then it means that even remains from Australia at this time would share this phenomenon, because the dating would be from the initial OOA introgression. If their assumption is false, which is very likely as John Hawks suggests, then it means they're merely averaging out old introgression with introgression that occurred recently, prob in Siberia. In neither case does this necessarily preclude any region of Eurasia from being a source for Ust-Ishim, because this Neanderthal heritage is shared by everyone.
For there to be 'population continuity in Siberia' to draw a direct line from K2* to R, there would need to be 1) migration from Siberia to SEA, and 2) the West-East divergence would have to begin in Siberia. But this is a priori impossible, because 1) All the descendants of K2*, MNOPS, have their highest diversity south of Siberia or North Eurasia, incl. N, O, P and R. 2) ENA and ANE met each other as fully-diverged branches of East and West Eurasians in Siberia, with no contact prior, so they must have diverged elsewhere. East Asian, Onge, Papuan, and ANE+WHG all meet in the south, while only East Asian and ANE met as diverged tips of two branches in the North. 2) East Asian appears to originate in the South. Its closest group is ASI/Onge, and not, say, WHG/ANE, which would be the case were Siberia the focus for the branching of new groups. Haplogroup NO, another descendant of K2*, originates in Southern China. 3) East Eurasians share different cold- and pigmentation- adaptations from West Eurasians, while Papuans and Onge share the same hair and skin alleles as Africans. Thus there is strong evidence for two colonisation events of temperate regions from tropical ones in Eurasia, and thus of a tropical evo hist of East Asians and Europeans, but no evidence that there ever was a temperate phase in the evo hist of Papuans and Onge.
The Eastern and Western branches of humanity must have diverged in the South, and pursued separate paths north, probably divided by the Himalayas. ENA went East and South from North India, with papuan-like pops going first and mixing with Denisova, and Onge and Dai-like pops entering later, diluting the fraction, while ANE-WHG went to the North, before spreading East and West in Siberia to reach Europe.
The only barrier for the Southern origin to work is that Ust Ishim has no Denisova. But it only takes a single improbable event for this to occur, which is that the Denisova-free population moving west contains a single person with Haplogroup K2*, and P* if it existed, each. That's it. And we know this is possible... by Sima de Los Huesos. He carried mtDna ancestral to Denisova amongst a Neanderthal population that moved West, and that genotype survived for a long time. This should be a better interpretation of his scenario IMO.
If MNOPS as a whole emerged in Siberia, the series of improbable events would come by the trainloads, and more importantly it would disagree with the 1) pop gen hist and 2) branching order of the tree of Eurasians we have now. The Middle East is even worse, as a start there means we have a chaotic jumble of all haplogroups, including G, E, I, J, and all the descendants of K, in the Ur-Eurasian population, and every single case of Y-haps found in any population is due to chance in a chaotic 'starburst'--none is due to population splits consistent with Y-Chrom cladistics.
ReplyDeleteIn fact, I am willing to bet that Ust-Ishim scores >60% South Asian+Papuan components in ADMIXTURE, because the basal ancestry he represents in the WHG-ANE-ENA clade has retained itself best in South Asia and Oceania, while the later, highly drifted offshoots in Europe and Asia are poor fits.
Last of all, the only thing Underhill suggested was that the root of R1a lay outside both a European-like population and an Indian-like one, and in this they were right weren't they?
My two cents.
Underhill was so wrong in his two papers on R1a that I could write a book on the topic.
ReplyDeleteBut anyway, if R1a originated in an Eastern European Hunter-Gatherer (EHG) population, which looks plausible at this stage, then I don't see why we can't call that a European-like population?
@a
ReplyDelete"and perhaps some possible undiscovered unkowns that we cannot imagine until we get more ancient data"
cool not just me confused then :)
@ Davidski
ReplyDeletePlease do in fact post an article on it.
R1a, and R in general, likely originated in a population that was ancestral to ANE or an ANE pop itself, such as EHG. It seems likely that the most ANE pop in Eurasia are the Kalash and Pashtun, followed by South Indians and North Caucasians, and only then by North Europeans.
You posted an article a whiles ago, 'who's the most European of them all', where you basically identified a WHG-like component ADMIXTURE produces as the essential feature distinguishing Europeans from others. I agree that the label 'European', if it is to have any meaning, should correspond to WHG, the only portion of ancestry that corresponds somewhat to a geogaphical boundary.
@ Grey
Just to clarify, there was only one OOA and the neanderthal sex party only happened once, so Basal and Crown Eurasian likely have similar Neanderthal admix.
Probably a bad experience there lol.
My sources tell me that the most EHG-like people today are Northeast Europeans.
ReplyDelete@ Davidski
ReplyDeleteWhat do your sources say about EHG and ANE vs WHG? Are they all as tight-lipped as Barak says on this issue?
Well that at least tells us that ANE and EHG are not the exact same thing.
Grey said...
ReplyDelete"cool not just me confused then :)"
I think some might fall into the trap of trying to extrapolate current data to fit in what has been found with regards archaic human samples. Add to this labeling issues, and you have perfect cocktail for confusion.
If by some chance 10,000 years from now someone were to find and R1a Bedouin grave [elevated to 100% EEF]or R1b Sardinia[elevated to 100% WHG] one could come to an erroneous assumption on ancestral origins. Today, at this moment in time with all available evidence, we know they are not from these regions. The same case could be made for ANE R1a ie Southern India has high levels of ANE therefore we should look for ancestral R1a in this region. However, the variance of R1a is really nothing when compared to Eastern Europe where some experts are placing PIE; yet Eastern Europe has lower levels of ANE than some Indian populations.
One interesting note on the possible dilution R-M207 R1a and R1b in the above examples happened in regions Sardinia and Middle East, known to be elevated in IJ marker M429.
Here is what Qiaomei Fu et al. say about the Onge (p. 65 of their Supplemental Information), "We include the Onge as they have no evidence of the Denisovan admixture that has affected indigenous Papuans, Australians, and Philippine populations, and may therefore be a useful population for understanding the early dispersal of modern humans out of Africa."
ReplyDeleteOnge must have very little Denisovan admixture. The ASHG paper that Davidski referred to tested four populations and found the least Denisovan admixture in Onge. The highsest admixture was in the Jamatia tribe of Eastern India which is clearly of South-East Asian origin. Most South Asians have very little or zero Denisovan admixture.
I believe that it is entirely possible that the ancestors of Ust-Ishim migrated north to Western Siberia from the Indian Subcontient.
Figure S10.3 (admisture analysis) shows that Ust-Ishim has some of all the various compoenents at K=10. For what it is worth, the South Asian component is the largest.
@Davidski
ReplyDeleteI have to agree that the sensible interpretation is often wrong, humans being active little buggers who don't flow around gradually in a parsimonious manner. With any luck we will get more aDNA in the future and find out.
@ ryukendo kendow
ADMIXTURE analysis (K=10) is in the supp info, and yes Kalash, then Papuan, then She are the largest components. There is no evidence I can see that he is related to ANE - Amerindian and Siberian components are minimal, he is on the Oceanian side of a PCA opposed to American, and TreeMix puts him on a separate branch from Mal'ta boy. However, it probably doesn't mean anything except that he is too early to have differentiated much (TreeMix flip-flops on whether he belongs with ENA or basal to both ENA and WE depending on the reference population).
The nearest sites archaeological sites of the right time period are unfortunately 1000 km away in the Altai (as are the Denisovans). That said he is presumed to be Siberian Initial Upper Paleolithic. This is later replaced by Early Upper Paleolithic microblade makers, who could possibly have developed locally, but are generally thought to have come in from elsewhere. There is no connection I have been able to find between SE Asian and Siberian material culture in the Paleolithic - apparently even Indian and SE Asian remain quite distinct for a very long time.
@ Ryukendo The Middle East is even worse, as a start there means we have a chaotic jumble of all haplogroups, including G, E, I, J, and all the descendants of K, in the Ur-Eurasian population, and every single case of Y-haps found in any population is due to chance in a chaotic 'starburst'--none is due to population splits consistent with Y-Chrom cladistics.
ReplyDeleteI am not knowledgeable on the y-dna tree or mt-dna tree and frequencies, so apologies if I get something wrong but -
Would it not make more sense to use y-dna divergence times against the autsomal rate both estimated from mutational rates rather than hope that the distribution of a particular group fits clades?
K(xLT) or K-M526 does seem by consensus old enough to have diverged prior to divergence date of the derived / “crown” Eurasians from one another.
Ust-Ishim's 45,000 year BP, and is either outside or simultaneous with the clade grouping for WHG-ANE and ENA. K(xLT) or K-M526 is dated to around 47,000 to 55,000.
Where R and P probably aren't old enough to proceed population divergence / structuring – R and P mutation dates was probably after the population divergence of at least the WHG-ANE and ENA clades. P is uncertain at 27,000 to 45,000, while R should be a definite at around 24,000 to 34,300.
It seems that for at a time after a split from Africa there was a fairly unstructured Eurasian population, then for a shorter time a fairly undifferentiated “except basal” xBasal / derived /crown population.
So any splits from around these times might well be totally uninformative about population split times, and where they or their descendents have unequal representation between groups, that might just be due to tendencies towards drift (magnified by polygyny / other male reproductive variance), which are uninformative about split dates.
That's how I'd understand it. Or is there are reason to think that the y chromosomal phylogeny's split times should “lead” or “match” the autosome when comes to population splits?
ReplyDelete@ Matt
There is indeed no reason why split times for y-chrom should match autosomal splits, the same way there is no reason to expect the TRMCA of a particular locus to be exactly at the split between two populations having the two alleles we are comparing.
At a particular locus, one person might have an allele that diverged from another person's millions of years ago, because both alleles were preserved in the population's gene pool, even if they share a parent say 3 generations ago. Mongolians have hap C and N, and these are extremely far apart in the branching.
However, splits should correlate with the geography of gene flow, and the cladistics of tree shape. It makes sense that mongolians can retain an ancestral marker like C, because new mutations do not imply that the old allele is lost; it might stay in the gene pool. Similarly for Hap F in Hmong. On the other hand if Hap N occurred in a Khoisan then something weird happened. We can infer that that individual is prob cape malay or something.
"It seems that for at a time after a split from Africa there was a fairly unstructured Eurasian population, then for a shorter time a fairly undifferentiated “except basal” xBasal / derived /crown population."
Precisely. This means that the geographical origin of that 'undifferentiated population' should see a high occurrence of basal variants from many branches in the tree, jumbled together in a chaotic geographical pastiche, while the areas that received gene flow should see a homogenous mix of descendant branches, segregated according to the cladistics of the split, with sub-pops diverging precisely from the 'drift' and 'polygyny' you mentioned, leading to more fixation events.
This is what we see in the ANE-WHG region, and also amongst E.Asia and Siberia; each is dominated by its own subset of descendant branches. The only area with the 'jumbled pastiche' is Southeast Asia and India, with 'Western Branches' like P* appearing amongst Aeta, M appearing in Papuan, Basal R1 and R2 in India, and whatnot. Not to mention Co-occurrence of mtDna supergroups M, N and R in India, the only West eurasian pop with more than only N and R.
This means that the geographical origin of that 'undifferentiated population' should see a high occurrence of basal variants from many branches in the tree, jumbled together in a chaotic geographical pastiche, while the areas that received gene flow should see a homogenous mix of descendant branches, segregated according to the cladistics of the split, with sub-pops diverging precisely from the 'drift' and 'polygyny' you mentioned, leading to more fixation events.
ReplyDeleteDon’t tendencies for fixation events / loss of mutations (at least for these locii) depend strongly on population size, where regions with a long term large population size (even a structured long term large population size) could maintain many rare variants for longer, even without being an actual origin?
So the fact that Southeast Asia and India carry a lot of early divergent* y-dna groups for pan Eurasian haplogroups doesn’t necessarily mean that that was where the pre-divergence population lived (just about plausible for India, certainly seems out of the question for SE Asia). But would seem more likely to mean that these regions have better preserved the pre-divergence maximum of variation, i.e. larger long term population size. For many Southeast Asian groups this is not true as individuals (Aeta, Papuans, etc.) but may be true when pooled.
*We say early divergent which I guess they are, but I think it is the case that this does not mean frozen after the early divergence - but they’re probably as mutated from the ancient consensus form if you look at them long enough, just these populations are understudied.
Isn’t it more likely that there were several K(xLT) migrations in different directions (perhaps starting from South Asia, along with mtDNA R), one a “dead-end” in the north (Ust-Ishim) that eventually vanished with climate deterioration and another one the more successful “southern” (SE Asian) K(xLT) that only later replaced the first wave in the north?
ReplyDelete@ Matt
ReplyDeleteActually, the main determinant of diversity--not diversity of unique haplotypes (which is meaningless if they are all, say, variants of Hap O), but diversity in terms of number of branches in the tree represented in the population, is the number of bottlenecks. This, when placed in sequence, give us the tree structure. Otherwise the usage of Y- and mt- DNA to root mankind in Africa would not work. This has to do with the cladistic position of the founding population, and not really on actual population size.
"For many Southeast Asian groups this is not true as individuals (Aeta, Papuans, etc.) but may be true when pooled."
Aeta, Papuan, and other SEAsian diverged from each other an extremely long time ago, with gene flow after that time insufficient to qualify them as a panmictic population, so I am not sure they can be 'pooled'. In fact they split so long ago and admixture is so insignificant that Reich was able to model phillippine negrito as diverged from Austronesian fillipinos, at the same time as they diverged from Papuan. The population size of the HG aeta, compared to their fillippino neighbours, would be really low for a really long time. But it is the fillippinos who have less upstream clade diversity than Aeta, not the other way round, because they can be reduced to a more homogenous and cladistically younger founding population than Aeta can. If all Aeta had P* and fillippinos a diversity of O, yes, Aeta lost all other alleles, but that does't change their position on the tree.
So that basal branches are preserved in SEA is highly indicative IMO.
Btw, I agree that the basal population is more likely to have occurred in East/central India than Southeast Asia. An aeta-like/Papuan-like gene pool was probably more widespread in SEA before overprinting during the East Asian Neolithic, transferring the center of gravity of the distr of clades to India.
@ Matt
ReplyDeleteJust to clarify, what you're proposing would work as a way to differentiate the effective population size of different populations. But we're not interested in that here; we're interested in the branch order.
I have enjoyed reading the discussion on this thread tremendously, and I think ryukendo kendow and Matt in particular have made some very good points.
ReplyDeleteI would like to add, however, that all published analyses (at least all that I have seen) have indicated a very short branch length of the common ancestor of all K2 after its MRCA with K1 (and the same for K vis-à-vis IJ, IJK vis-à-vis H, or HIJK vis-à-vis G). The clarification of the bifurcations in the phylogeny of the early descendants of F-M89 is surely an important step in the genetic investigation of human origins, but these bifurcations do not seem to be particularly significant in an evolutionary sense; the gap between the TMRCA of G and HIJK and the TMRCA of NO and P should be at most about 5000 years according to published estimates, even after adjusting upwards for the discovery of K2* in the Ust-Ishim specimen. Any significant "racial" distinction among early Eurasian populations of anatomically modern humans should be much deeper, like at the DE-YAP vs. C-M130 vs. F-M89 level. There is a huge gap between the TMRCA of {F + C} (or {F + DE}) and the TMRCA of all extant F-M89 ({F* + GHIJK}).
@ Ebizur
ReplyDelete"the gap between the TMRCA of G and HIJK and the TMRCA of NO and P should be at most about 5000 years according to published estimates, even after adjusting upwards for the discovery of K2* in the Ust-Ishim specimen. Any significant "racial" distinction among early Eurasian populations of anatomically modern humans should be much deeper."
Hmmm.. but doesn't the discovery of Ust-Ishim precisely show that the divergence within crown Eurasian dates to later than this unexpectedly shallow period, and that the parent population was in the midst of this unstructured 'big bang'?
I think we can put this with Jarve's paper from ASHG: Rapid radiation of common Eurasian Y chromosome haplogroups occurred significantly later than the out of Africa migration.
"Compared to the length of the branches that separate African and non-African diversity, the internal branches distinguishing continental and sub-continental differences outside Africa are generally short, consistent with the model of a rapid initial colonization of Eurasia and Oceania. (This would spread long stretches of Neanderthal everywhere.--RK) Following the split between African and non-African haplogroups [90 KYA (95% CI: 87-94 KYA)], the differentiation of South and Southeast Asian haplogroups H, S, M, and C did not begin until around 43 KYA, and haplogroups N and R, widely spread among Northeast and Northwest Eurasian populations, started to diversify significantly later [17 KYA (95% CI: 16-19 KYA) and 26 KYA (95% CI: 25-28 KYA), respectively]. Many major phylogenetic groups in different geographic regions seem to originate from the period around 50 KYA."
This places Ust-Ishim in the K (X MNOPS) period, before the diversification of K's descendants, nicely. It also suggests that populations in tropical Asia expanded before populations in North Eurasia did.
I am just pointing out the fact that the divergence between an early proto-K2 population and an early proto-G population, for example, should have been about the same degree as (or even less than) the divergence among modern Celtic, Germanic, and Italic Indo-Europeans, with some minor differences being attributable to founder effects, but no real significant difference except where due to mixing with a deeply divergent outgroup (which is probably a good explanation in the case of either {Italics} or {Celtics and Germanics}).
ReplyDeleteI agree with Rukendo that "TRMCA of a particular locus to be exactly at the split between two populations having the two alleles we are comparing " and that South Asia today harbours much basal diversity around K, P, R2, R1, etc.
ReplyDeleteHowever, as Im sure others are aware, given that there are less geographic barriers, and recently smaller population density, in central –northern Asia, past structure could have been swept away .
Whatever the case, I think in reality that finding where a haplogroup ‘arose” really tells us nothing. Contrary to what most people here think, new haplogroups arising by way of stochastic mutation tells us little about actual population events. As we have already seen from the ancient data, groups were diverse with both ancestral and derived groups from as early as the Neolithic, meaning that early AMHs and First farmers alike were genetically diverse from the outset.
Why should a random genetic mutation, ie 'evolution' from Hg K* to P* to R1* mean anything from a population perspective ? It doesn't. They didn't necessarily migrate, split, etc at this stage
ReplyDeleteIt's very unlikely that early Neolithic farmers were diverse. They obviously became increasingly diverse as they pushed further out of the Fertile Crescent, because they incorporated indigenous hunter-gatherers into their ranks.
ReplyDeleteBut even so, in Europe Neolithic genomes mostly cluster together, and with more sampling we'll just see them sliding along the same north-south cline depending on how much WHG they have. East-west differentiation in Europe is a phenomenon associated with later population movements, mainly from Eastern Europe, bringing ANE or EHG.
To be clear, I should have explained why my previous point is relevant to the present discussion.
ReplyDeleteIf we assume that the Ust-Ishim specimen was autosomally equidistant to ENA and {ANE+WHG} because he predated the split between the ancestors of these groups (as opposed to being a 50%/50% mix of ENA and {ANE+WHG}), then at least one of the following three statements must be true:
1. the occurrence of subclades of C-M130, D-M174, and E-M96 in modern Eurasian populations whose Y-DNA otherwise belongs to F-M89 is due to retention of ancestral diversity from a polymorphic ancestral population (which has been lost in some descendant branches due to founder effects and/or genetic drift)
2. the occurrence of subclades of C-M130, D-M174, and E-M96 in modern Eurasian populations whose Y-DNA otherwise belongs to F-M89 is due to very minor (not autosomally significant overall) introgression from deeply divergent outgroups
3. the occurrence of subclades of C-M130, D-M174, and E-M96 in modern Eurasian populations whose Y-DNA otherwise belongs to F-M89 is due to autosomally significant admixture from deeply divergent outgroups ("Basal Eurasians"), the only one such event that has been described so far being the postulated "Basal Eurasian" gene flow from Southwest Asia associated with the demic diffusion of Neolithic populations in Western Eurasia.
Which one of these three is appropriate may depend on the population in question. If we assume that Eurasians (in contrast to Palaeo-Africans) have expanded from somewhere near Japan, for example, then the high total frequency of {D-M174 + C1a1-M8 + C2-M217} Y-DNA in that country may be explained by retention of ancestral diversity that has been lost elsewhere due to founder effects and drift, whereas the occurrence of C1a2-V20 in Mesolithic, Neolithic, and modern Europeans may be explained by introgression as stated above. The presence of subclades of E1b1b1-M35 in modern Western Eurasians may perhaps be attributable to insignificant introgression, but it is more likely attributable to significant gene flow from a deeply divergent outgroup ("Basal Eurasian" or African or whatever) IMHO.
Davidski,
ReplyDeleteQuite the opposite is more likely. A careful study of the Neolithic actually reveals that it wasn’t the simple demic diffusion originally described, but movements of small groups from diverse parts of the Near East, and not one single ‘core’ area. So there is no reason for them to be genetically homogeneous, and were thus likely associated with a variety of Hgs - G2, probably R1b, even residual I* hanging around in Near east, possibly J2. In addition to that, even within the individual haplogroup families, you’d have had both ancestral and derived subclades. All this makes for quite a bit of diversity.
This is not contradicted by their overall clustering from a genome -wide perspective.
"If MNOPS as a whole emerged in Siberia, the series of improbable events would come by the trainloads"
ReplyDeleteMy thoughts exactly.
"I no longer have any faith in interpretations of modern Y-haplogroup data"
It seems to work extremely well everywhere else.
"Well, Denisovan in South Asia kinda changes things".
Perhaps it is closely associated with mt-DNA M. That would certainly explain its higher presence in Papua than in Australia. I have always had difficulty understanding that situation if the Denisovan element was associated with Y-DNA C or mt-DNA N.
"So the fact that Southeast Asia and India carry a lot of early divergent* y-dna groups for pan Eurasian haplogroups doesn’t necessarily mean that that was where the pre-divergence population lived"
It does, unless you're going to postulate haplotype extinction on a massive scale. And don't forget we're not just dealing with M and S here. S is just one of four branches within K2b1a-P405, all found only east of Wallace's Line. And M is just one of four branches within K2b1-P397, all found scattered around Wallacea. The 'interpretations of modern Y-haplogroup data' fit that region exceptionally well.
"Rapid radiation of common Eurasian Y chromosome haplogroups occurred significantly later than the out of Africa migration".
That is what I believe the evidence shows. The most likely explanation I can think of for that second rapid expansion is the spread of the efficient boating technology required to cross Wallace's Line. We can be certain the technology did not move only in one direction. That would also explain a rapid movement through South Asia along the Ganges and Indus. The resulting mixture was than able to move into the, at most sparsely settled, steppe region.
"1. the occurrence of subclades of C-M130, D-M174, and E-M96 in modern Eurasian populations whose Y-DNA otherwise belongs to F-M89 is due to retention of ancestral diversity from a polymorphic ancestral population (which has been lost in some descendant branches due to founder effects and/or genetic drift)"
ReplyDeleteOf course, this case would require the presence of C-M130, D-M174, or E-M96 in either the Ust-Ishim population or some other very closely related population from which both ENA and {ANE+WHG} are descended, and we should expect to find such Y-DNA if we dig more and find other ancient specimens closely related to "Usty." (In other words, "Usty" merely happened to belong to K2*, although there were other, deeply divergent Y-DNA haplogroups also present among members of his population or close relatives.)
Diversity in uniparental markers doesn't necessarily translate to diversity in genome-wide genetic structure.
ReplyDeleteEuropean Neolithic farmers are a great example of this, because they show a very diverse mtDNA gene pool, and yet they just slide along the WHG/EEF cline, and usually land somewhere around present-day Sardinians.
'' I have always had difficulty understanding that situation if the Denisovan element was associated with Y-DNA C or mt-DNA N. ''
ReplyDeleteIf this find proves anything it proves that haplogroup K2 is associated with mtdna R the main subclade of N,and not with mtdna M. He was at the root of both R and K for crying out loud.
In Australia 28.1% of Aborigines have
R, 56.1 have N(xR) 15.1% have M.
In PNG 65.7 have R the only type of N in PNG while the other 35.3% have M.
While in PNG around 82% have K2b while around 12% have C2 (CxC3) in Australia around 45% have C2 (CxC3 while around 50% have K2b with the rest having K2a (O) . K2 is obviously associated with R, you could be correct that C is associated with N (xR) but Ust-Ishim strongly supports the theory that K2 is associated with R.
http://www.pnas.org/content/suppl/2007/05/10/0702928104.DC1
ReplyDeletehttp://en.wikipedia.org/wiki/Haplogroup_K2b-P331
"Diversity in uniparental markers doesn't necessarily translate to diversity in genome-wide genetic structure.
ReplyDeleteEuropean Neolithic farmers are a great example of this, because they show a very diverse mtDNA gene pool, and yet they just slide along the WHG/EEF cline, and usually land somewhere around present-day Sardinians"
Noone is disputing that. Of course they were broadly all quite similar. That's what autosomes tell us, they certainly weren't "Oceanic", Siberian or sub-Saharan.
But under the facade they were Of homogeneity there clearly was diversity - genetic and geographic; just like Modern Europeans today are very similar (excepting a few outliers), who still otherwise harbour some rather diverse prehistory.
My main point is that individuals did not live in homogeneous groups/ populations. As someone stated above, ‘Utsy’ might have been a chance relict of a person (Y-DNA speaking) , in a population that otherwise already had R1a, N, etc....
That's not correct.
ReplyDeleteEarly European Farmers were more homogenous than modern Europeans, and easily modeled as a mixture of only two components, usually with very similar levels of these components.
@ Davidski
ReplyDeleteAgree.
@ Mike Thomas
"However, as Im sure others are aware, given that there are less geographic barriers, and recently smaller population density, in central –northern Asia, past structure could have been swept away . "
This cannot lend support to the northern origin. There is no reason why drift and replacement in small populations in N.Eurasia should selectively target basal branches of haplogroups, leaving all the most derived branches in N.Eurasia and Basal+Derived branches in SEA and India. If P*, K* and such is old in N.Eurasia but N.Eurasian pops were small, there is no reason why drift should result in, say, a pool of R1a and N1c, than say a pool containing K*/K2 clades and R1a.
If in fact basal branching began in Siberia and them movements brought genes south, then Southern pops should have more homogenous and derived pops than northern ones, and southern haplogroups should be more cleanly segregated geographically than northern, and this is simply false. If migration then resulted in homogenisation in the north, one should still expect to see at least a single instance of Basal branches surviving, given that P* has been able to survive in such a small population as Aeta, or F in Lahu for so long. In fact, every single expansion in northern Eurasia centers on a derived clade that seem to have southern roots; not one centers on a basal clade that could have been plausibly autochthonous.
Secondly we simply have to throw out the Siberian expansion idea. It conflicts with everything we know about the pop gen hist of East Asians, WEurasians, Oceanians, branching order, climactic adaptation, everything. Remember that Australia was quite possibly populated before Europe was. The northern origin of East Asians ceased to be mentioned as a plausible theory years ago.
"Why should a random genetic mutation, ie 'evolution' from Hg K* to P* to R1* mean anything from a population perspective ?... They didn't necessarily migrate, split, etc at this stage."
The mutation does not tell us anything 'at that stage'. But which branches are present within a population later does, because it tells us which population are bottlenecked descendants of which other populations after the mutation happened (a 'subset' of the other population), and thus allows us to construct a tree. The available evidence suggests that East Asian, Oceanian and non-Basal west Eurasian are non-overlapping subsets of Some south/Southeast Asian population. The opposite statement doesn't even really make sense, because it doesn't even agree with a cladistic model of genetics.
@ John Smith
Those are very sharp observations. They seem to imply an even stronger connection between Ust-Ishim and the southeastern rim of Eurasia.
@ Ebizur
I think all three are significant, but we probably underestimate the significance of 1). Agree that Hap E and G in W.Eurasian probably represents an example of case 3).
@Terryt @Ebizur
Probably the OOA population spread extremely rapidly en route to Australia, leaving a 'string of pearls' around the Indian Ocean Rim, most of which were reproductively isolated. The rapid emergence of the first defining mutations of the haploid trees probably dates from this period.
Its funny to think that the huge divergences between Eurasian pops today had their roots, perhaps, in an otherwise homogenous population divided by a single ford or mountain range.
Rukendo – what you’re saying makes sense . I was just playing devil’s advocate for discussion’s sake.
ReplyDeleteDavid – you’re still misunderstanding me. I am referring to Y haplogroup heterogeneity, not that the Neolithic farmer groups were entirely unrelated to each other. Remember, that autosomes show more similiarities between individuals/ groups whilst Y DNA brings out their differences.
Well, I dunno about the Y haplogroup heterogeneity.
ReplyDeleteThe Neolithic Middle Easterners who pushed into Europe mostly carried G2a and occasionally something else like E, and the native hunter-gatherers who sometimes hooked up with them mostly carried I2 and occasionally C6.
Possibly R1b and J2 also though not yet sampled.
ReplyDeleteDavid – what do you think about Rukendos idea of a southern origin for K and P* ?
I'm very skeptical of a more southerly origin than Central Asia for both.
ReplyDeleteryukendo kendow,
ReplyDelete"Could you guys tell me why this strikes a blow against the emergence of K in Southeast Asia? . . .
The evidence for an eastern origin of K, with so many basal K2b clades in Southeast Asia, is very strong; unless we want to propose a parallel migration of all groups eastwards followed by a clean geographical segregation of each subclade"
The overwhelming majority of Y lineages from the dates we're talking about no longer exist. You and others apparently fail to appreciate either the very large random component in which lineages survive and where, or the fact that demographic and historical factors likely to differ between the regions in question probably also have had distorting effects. You're entirely misinformed on how "strong" inferences from statistics-free, eyeballed phylogeography are in cases like this. Alan Templeton's rule of thumb is that "the evolutionary tree of genetic variation for a particular DNA region is informative only for the second half of the tree's time depth", and K(xLT) coalesces well over halfway back in the Eurasian Y chromosome tree.
My guess is that the preservation of early K-M526 diversity in SE Asia is attributable to the relatively subdivided population and large total effective population size.
In the absence of any other evidence (including ignoring the Y phylogeny immediately above K-M526), the present-day distribution of M526 subclades would make SE Asia a reasonable guess for the location where K-M526 originated. But even in that case, it would just be a guess, subject to high levels of uncertainty. And, since we do have other evidence, even prior to this publication the terryt fantasy version of the peopling of Eurasia was self-evidently absurd. The confirmed presence of K-M526 this early and this far north, in a Denisovan-free individual, just reinforces the point.
"This cannot lend support to the northern origin. There is no reason why drift and replacement in small populations in N.Eurasia should selectively target basal branches of haplogroups, leaving all the most derived branches in N.Eurasia and Basal+Derived branches in SEA and India. If P*, K* and such is old in N.Eurasia but N.Eurasian pops were small, there is no reason why drift should result in, say, a pool of R1a and N1c, than say a pool containing K*/K2 clades and R1a."
ReplyDeleteI'm having trouble wrapping my head around this level of incomprehension. "Basal branches" do not need to be "targeted". All but a few early branches of K-M526 died out everywhere. R1a and N1c are among the K-M526 branches that survived near where K-M526 likely originated. The "basal branches" found only in SE Asia and the branches leading to R1a and N1c all originated within a span of no more than about 1000 years.
"If in fact basal branching began in Siberia and them movements brought genes south, then Southern pops should have more homogenous and derived pops than northern ones, and southern haplogroups should be more cleanly segregated geographically than northern, and this is simply false. If migration then resulted in homogenisation in the north, one should still expect to see at least a single instance of Basal branches surviving, given that P* has been able to survive in such a small population as Aeta, or F in Lahu for so long."
Again, you're completely lacking any statistical intuition here. The population or populations that brought K-M526 into SE Asia would already have carried all the "basal branches" now found exclusively in SE Asia and Oceania, along with many other lineages which are now found nowhere. That a few more of these survived in SE Asia than elsewhere is not hugely informative, when nearly all of them died out everywhere. And the fact that some survived in small island populations becomes less surprising when you realize there are many different small, relatively isolated populations in SE Asia and Oceania. Again, more genetic diversity is preserved in a subdivided population than in a pan-mictic population of the same size.
You also seem to be confused about what exactly it means that "P*" has been found in the Aeta. This does not mean two different lineages that share no mutations below P have been found among the Aeta. It means that a Y lineage found among the Aeta falls on a different branch of P than is found in mainland Eurasians today. The Aeta lineage is not somehow older or "more basal" than R. It's just a lineage which had not previously been described, and does not have a top-level haplogroup name of its own. It's a brother of R, not the father of R.
"If this find proves anything it proves that haplogroup K2 is associated with mtdna R the main subclade of N,and not with mtdna M. He was at the root of both R and K for crying out loud".
ReplyDeleteK2 is almost certainly not associated with mt-DNA R when it comes to New Guinea/Melanesia. The great bulk of mt-DNAs in that region are M. However K2 certainly looks associated with R on any move back west from that region. It is certainly not unknown for male and female haplogroups to change partners at the origin of any expansion.
"In PNG 65.7 have R the only type of N in PNG"
Mostly in the form of B4a, whose arrival is associated with the Austronesian expansion. Any other R is P, most of whose branches are confined to Australia and so probably originated there. New Guinea has no basal N whereas Australia has four basal branches of that haplogroup.
"While in PNG around 82% have K2b while around 12% have C2 (CxC3)"
An example of a male/female switch. I'll explain. C2 (C1c now) in New Guinea is primarily C1c1, originating in Southern Wallacea. This lineage is the one that moved out into the Pacific with mt-DNA B4a1a1a, originating in Taiwan/Philippines.
"in Australia around 45% have C2"
Australia has almost none of the old C2. It has a completely different lineage from Wallacean/New Guinea. It used to be called C4, now C1d.
"while around 50% have K2b"
Specifically K2b1a-P60, an otherwise exclusively New Guinea/Melanesian haplogroup and even K2b1a-P60 is shared with New Guinea and presumably entered Australia from there. Any O in Australia is a product of the Austronesian expansion.
"but Ust-Ishim strongly supports the theory that K2 is associated with R".
By the time the haplogroups entered the steppe, and presumably for some perhaps limited period before then.
"The overwhelming majority of Y lineages from the dates we're talking about no longer exist".
We don't know that to be so although it must have happened to some extent. But once you start making up scenarios where haplogroups have undergone massive extinction you can 'prove' any particular hplogroup originated anywhere you wish to place them.
"In the absence of any other evidence (including ignoring the Y phylogeny immediately above K-M526)"
But we do have evidence of the Y phylogeny immediately above K-M526. South Asian LT (K1-P326) is immediately upstream and SW Asian IJ further upstream still. That fits perfectly a serial movement east from the Iranian Plateau, or somewhere very nearby.
"The population or populations that brought K-M526 into SE Asia would already have carried all the "basal branches" now found exclusively in SE Asia and Oceania"
That sounds suspiciously like a Garden of Eden scenario where all the sons were already present at the start of the exodus.
terryt,
ReplyDelete"The overwhelming majority of Y lineages from the dates we're talking about no longer exist".
We don't know that to be so although it must have happened to some extent. But once you start making up scenarios where haplogroups have undergone massive extinction you can 'prove' any particular hplogroup originated anywhere you wish to place them.
Unless you want to argue that the total number of human males was in the dozens 50,000 years ago, this point is not up for debate. Virtually none of the men living at that time left male-line descendants down to the present. We find in living humans only an infinitesimal sampling of Y diversity from that era (a half dozen or so lineages out of tens or hundreds of thousands).
"The population or populations that brought K-M526 into SE Asia would already have carried all the "basal branches" now found exclusively in SE Asia and Oceania"
That sounds suspiciously like a Garden of Eden scenario where all the sons were already present at the start of the exodus.
My point is that the "basal branches" under discussion here all originated within a relatively short time of K-M526. Assuming I'm correct that K-M526 originated outside of SE Asia (which I am) and at least a couple thousand years passed between the origin of K-M526 and its introduction into SE Asia, all of the major branches (those leading to NO, P, K2b1, K2c, and K2d) would already have been in existence prior to their arrival in SE Asia.
n/a - your points are well argued, but what about India - in lieu of its diversity of R linegaes (P* , R1a, R2) , which is hardly an island. Moreover, arguments about its social segregation are not valid because such ethnic, religious and social distinctions are relatively new constructs.
ReplyDeleteMike Thomas,
ReplyDeleteIndian "R*" turned out to branch off just slightly upstream of R2 as it was then defined (R2 may since have been redefined to include these lineages). And neither R1a nor R2 is likely to have originated in India. The overwhelming bulk of Indian R1a is definitely of relatively recent introduction from the steppe.
@ n/a
ReplyDeleteFirstly, I did not argue for an expansion from SEA. I argued that SEA was closer to the epicenter in the South of Eurasia, which was probably India.
Secondly, I did not make my conclusions off of 'eyeballing Y-haps'. If you realise, the majority of my inferences derive from the necessity to root the WHG-ANE-ENA subclade somewhere, and to place the common features of East and West Eurasian, incl. membership in Hap K, somewhere. Y chroms can lie, but autosomal DNA cannot, and we know the branch order for that already, which is ((WHG-ANE)(ENA)), not ((WHG)(ANE(ENA)). This points to the central role the Himalayas played in dividing the two migrations into temperate regions (not one migration). This, plus the fact that commonalities between the two groups, incl. K* and P*, all lie south of the himalayas, is what makes me think this is a compelling picture.
As for my 'lack of statistical understanding', let me rephrase. Yes, a lot of branches are dead. The thing to explain here is why SEA/Oceania has the descendants of more brothers close to the root, and West Eurasia has the descendants of only one brother close to the root. Why SEA/Oceania has the branches of multiple trees alive, while West Eurasia has the branches of only one tree alive.
A larger effective population size slows down the rate at which we lose the branches of trees, and smaller pops lose branches faster, but they do not therefore become 'homogenized' by losing all the branches of single trees/losing all the descendants of a single brother, until all their branches come from just one tree. With drift, you're gonna get thinner trees, and not entire trees killed off. The only time that happens is when a founding pop was absolutely tiny--in which case you get a bottleneck, which is what I said earlier.
Even the Ur-Asians carried Haps C, F, D, with the ancestors of whatever else they are carrying today. They thus carried live branches of multiple trees, and they were by no means a large pop at the time. On the other hand, the West Eurasians today have descendants from only a single brother in K2, PQR, and thus carry branches from only a single tree. They were always going to carry fewer branches, or lose branches faster; why did they not end up with branches from different trees, like the East Asians, though? The simplest explanation is that West Eurasians never had those other brothers in the first place; they are a bottlenecked subset of some more diverse population that had a larger pool of brothers. Similarly East Asians do not carry IJ, or E, or LT, they carry representatives of fewer trees than those further South and West, pointing to another pop division via bottleneck, allowing us to place them in the cladistics as deriving from those South and West.
West Eurasia has fewer branches than Oceania; given their pop history this was always a given. But West Eurasia has all its branches in a single tree, while the pops in SEA do not. This is what is informative, and why people look at Y chroms in the first place.
No, that was me doing the eyeballing, hence the map in the post above.
ReplyDeleteI think it can be a good way to cut through all the bullshit and focus on the bare facts.
Bullshit? Bare fact?
ReplyDeleteUsing n/a 's logic, if it's possible to say that 'all the diversity existed out of Africa already, during the green Sahara Eurasians endured Africa and introduced all the diversity, then they because of smaller pops in Eurasia later and larger eff pop size and difficult movement in Africa all the other diversity died out in Eurasia. Voila, into africa!'
This doesn't work because population genetics.
Thus we have come to the crux of the problem. Both sides make good points; I suspect expert population geneticists opinion night help away the arguments
ReplyDelete*might help sway
ReplyDeleteSo called expert population geneticists are usually about five years behind on this sort of stuff.
ReplyDeleteSome of them still think R1a is native to India.
No doubt, but I meant specifically about the utility of "basal diversity". Correct me if I am wrong, but you have used the same arguement about diversity of R1a in EE as Rukendo (sorry, I know I keep misspelling it) has used about K more broadly in central-southern Asia . If we are going to be consistent, we either have to reject or accept such an arguement for all cases, and not use it selectively (?)
ReplyDeleteThe argument regarding R1a1a is much less complex, since we're only going back about 6,000 years, and the phylogeny in question is very basic. On top of that we have ancient R1a from a variety of Eurasian remains that fit the modern phylogeny and the PIE steppe hypothesis.
ReplyDeleten/a has added some interesting points to the discussion. I hope r.k., n/a, and others will maintain civil decorum.
ReplyDeleteI think r.k. is correct that a major bottleneck or founder effect would be necessary to prune a deeply diversified gene tree so severely that only one branch would remain.
This raises the issue of how we should model Palaeolithic hunter-gatherer populations. If they were all ever so small and endogamous that they might lose all but one deeply divergent branch of a gene tree, then they should never have been able to accumulate such deeply divergent branches in the first place. Every population should have been under a "constant bottleneck" (i.e. extreme genetic drift all the time), so each population should have possessed only one shallowly coalescent Y-DNA lineage, and the presence of that Y-DNA lineage should be almost perfectly correlated with an endemic and likewise undiverse autosomal profile produced by genetic drift within a tiny, inbred population.
This would rule out the possibility of my previously mentioned scenario (1). Any non-F Y-DNA in any modern Eurasian population would have to be explained either as (autosomally insignificant) introgression from a deeply divergent "Basal Eurasian" population or as (autosomally significant) admixture from a deeply divergent "Basal Eurasian" population. Many distinct "Basal Eurasian" populations may need to be posited in order to explain all the varieties of non-F Y-DNA that have been found in various present-day Eurasian populations.
In other words, I think that either r.k. is correct, Palaeolithic hunter-gatherer populations had the potential to maintain two or more deeply divergent lineages, and following a trail of reduction of cladistic diversity (indicating founder effects due to settlement of new territories) is the most accurate way of tracing the history of human dispersal, or n/a is correct, Palaeolithic hunter-gatherer population sizes were sufficient only to maintain a single shallowly coalescent lineage, and a single population could have (and must have, otherwise the diversity would not exist at present) pumped out any number of branches of its particular endemic clade (e.g. K2) over time to colonize empty territories or conquer new territories from their previous inhabitants.
I would also suggest that n/a might have one foot in the trap of accepting the ISOGG or YFull Y-DNA tree as a Bible and reading too much significance into it because of having lost sight of the arbitrary nature of the nomenclature. We have only 20 Y-DNA clades named with a particular letter of the alphabet, but these clades (and each of their subclades) are not equal in phylogenetic level or internal diversity. Modern representatives of the clade that has been labeled as N-M231 do exhibit a pattern that would be consistent with a low long-term effective population size (resulting in a bottleneck and low coalescence age) as you have suggested; however, modern representatives of that clade's nearest outgroup, O-M175, do not exhibit such a pattern on the whole (although some of O-M175's subclades, such as O1a-M119 or O2b-M176, may exhibit such a pattern individually).
To expand on my last comment a bit, I would note that the topology of the phylogenetic tree of human Y-DNA as we know it is actually quite complex. Taking the C-M130 lineage for example, it appears to have been bottlenecked at around the same time as F-M89, many thousands of years after the ancestral divergence between the C and F lineages. Then, after the appearance of the C1 and C2 lineages, C1 rapidly diversified into numerous subclades that are now scattered from Ireland to Japan and from Arabia to Rapa Nui, whereas C2 suffered another bottleneck after a length of time roughly similar to that which had passed between the MRCA of F and C and the MRCA of C1 and C2. After this second bottleneck, C2 finally split into two subclades, one which is now found mainly in East Asia, Central Asia, and southern Siberia, and another which is now found mainly in Central Asia, North Asia, and North America.
ReplyDeleteSimilarly, on the F side of the C vs. F split, G split from HIJK at around the same time as C1 split from C2. Then HIJK appears to have expanded rapidly all over Eurasia (and eventually beyond), sort of like C1, whereas (proto-)G remained limited to some small population and was "let out of the bottle" only much later, sort of like C2.
I think the rapid diversification of C1 and HIJK in the Palaeolithic period under your scenario of low effective population sizes of hunter-gatherers and ubiquitous and frequent bottlenecks can only be explained by the rapid (and coincidentally roughly simultaneous) expansions of two small ancestral populations over wide geographical areas resulting in population subdivision. Where was the common ancestral C population prior to the split between C1 and C2? To what region did C1 expand allowing it to diversify so rapidly? Did C2 stay behind in the ancestral homeland, or did it move to another area with limited opportunity for population expansion? Same for HIJK and G.
So, does the realization that modern Europeans have some Basal Eurasian ancestry make the conclusions of the Tianyuan genome paper completely wrong? Was he really an East Asian, or just in a position analogous to Ust-Ishim?
ReplyDeleteMore in response to n/a's comments of this sort: "All but a few early branches of K-M526 died out everywhere. R1a and N1c are among the K-M526 branches that survived near where K-M526 likely originated. The "basal branches" found only in SE Asia and the branches leading to R1a and N1c all originated within a span of no more than about 1000 years."
ReplyDeleteYou clearly agree with me in regard to the fact that some of the nodes in the ISOGG tree are more evolutionarily significant than others. For example, the C node, the F node, and the N node are evolutionarily significant because they have emerged from a bottleneck after a long interval of separation from the nearest outgroup, so each of these populations must have developed its own distinctive autosomal profile by the time it had passed through the bottleneck (assuming that Palaeolithic populations could not maintain two or more deeply divergent Y-lineages). In contrast, K2 is a much less evolutionarily significant node because it radiates into a plethora of subclades very soon after it has diverged from its nearest outgroup (K1). There should not have been any significant autosomal difference among the ancestral K2 population, the ancestral K2a population (or even the ancestral NO population), and the ancestral K2b population because these all emerged within an extremely short timespan.
However, looking both upstream and downstream from the K2 node, I see a lot of "evolutionarily insignificant" nodes that are not associated with any clear bottleneck. K, IJK, HIJK, GHIJK on the one side, and at least K2b2/P, NO, O2, and O3 on the other; I do not know the internal structure of each of the Australasian-specific subclades of K2 in sufficient detail to comment on whether they show evidence of significant bottlenecks. Considering the phylogenetic position of K2 within this sort of dense branching pattern, how can you claim that "all but a few early branches of K-M526 died out everywhere"? Your claim seems to be inconsistent.
On the other hand, the geographic location and age of the Ust-Ishim specimen do show that some early descendant of K2 ended up in Western Siberia. I think it is important to determine how his Y-chromosome is related to extant subclades of K2; if he belongs to an outgroup of K2a and K2b (i.e. K2a and K2b share a common ancestor more recent than their common ancestor with "Usty"), then it would be equally parsimonious to infer that Usty (or his recent patrilineal ancestor) had migrated northward (or northwestward, northeastward, etc.) from the homeland of the F-M89 people whereas the MRCA of K2a and K2b had remained in the homeland of the F-M89 people, and the descendants of K2a and K2b emerged from that latter (presumably more southerly) region. Claiming that Usty (or Y-DNA haplogroup K2) is indigenous to Siberia is tantamount to claiming that {WHG+ANE} and ENA (or Y-DNA haplogroup F) are indigenous to Siberia. Not impossible, but certainly quite controversial.
If, on the other hand, Usty's Y-DNA is more closely related to a particular subclade of K2 (e.g. K2b2 or even QR-M45), then it would suggest that that particular subclade of K2 had successfully colonized Siberia by the time in which Usty lived, but the other subclades of K2 could have traveled eastward by a more southerly route.
I have taken away quite a bit from this discussion as well. Thanks to everyone keeping up a high level of debate, as well as a trove of good points.
ReplyDelete@ Ebizur @ n/a
Thanks for your points.
I think a scenario, of slightly larger effective population sizes, with strong bottlenecks at each bifurcation, is a priori correct. Because Middle Easterners have descendants of (C)(DE)(GHIJK), Indians have (C)(DE)(HIJK(K1K2)), Europeans have (C)(DE)(IJK2) and without later input from Basal in Middle East they just have (C)(IK2), East Asians have (C)(DE)(K2), and Native Americans just have (C)(K2). So there is a consistent pattern of reductions in diversity as one moves North and East across Eurasia.
This raises the issue of how 'slightly larrger eff pop size' could have been maintained. At the same time, I'm not sure that this makes a problem for 'a small group spreading rapidly over great distances', because this just changes the focus of what 'subset' means from 'the number of founding persons in the population from a single group in a single migration/bottleneck event' to 'the number of small groups, each with a separate Y-Line (but by no means all the same Y-line) that contributed to the founding pop of a new geographical population'. Both these cause 'subsetting' of the population to occur with each peopling of a new region.
I hypothesize that 'slightly larger eff pop size' can in fact be maintained if the pop was not panmictic, with high isolation-by-distance and high subdivision, which fits with HG-style social organisation. In that case the issue boils down not to 'how many people contributed' but 'how many bands, each with own shallow Y-lineage, contributed'. The problem then becomes less the density of people, but the density of separate bands, that drives the retention of multiple deep clades within a particular geographical region, with subsetting occurring during the peopling of new regions.
''In PNG 65.7 have R the only type of N in PNG" ''
ReplyDelete56.3% of Papuans have P (3 subclades), while 2.4% have R14 both native to the region. Only 6.4% have austronesian R (B4a)
Around 15% of aborigines have P (2 subclades) the only type of R in Australia.
K2b is much more common in PNG than it is in Australia, and much more diverse likewise mtdna R is more common in PNG than Australia and more diverse
''K2 is almost certainly not associated with mt-DNA R when it comes to New Guinea/Melanesia. The great bulk of mt-DNAs in that region are M'' Other areas of Melanesia are dominated by what is now known as C1.
To me it looks like a K2 population with mtdna R was the first people to enter the area of Asia East of Iran and the Urals. They spread quickly like the native americans (who were also their descendants mostly) because they had no competition. I am confident that K2 was in places like China and India, before C or D (if D ever was in India which I doubt). Interestingly in Eurasia all K2 is either P-M45 or NO. To me it seems a disaster happened 40-50kya that did not damage Melanesia very badly at all but devastated mainland Eurasia leaving perhaps only 2 male survivors with K2, much like Toba makes it look like humans came from Africa in less than 100KYA as it killed almost all those in Eurasia but not those in Africa (C'DE'F split was at the time of Toba).
''Indians have (C)(DE)(HIJK(K1K2))''
ReplyDeleteIndians do not have DE.
Blogger Davidski said...
ReplyDelete" The argument regarding R1a1a is much less complex, since we're only going back about 6,000 years, and the phylogeny in question is very basic. On top of that we have ancient R1a from a variety of Eurasian remains that fit the modern phylogeny and the PIE steppe hypothesis."
Exactly the common denominator of a sample that is 45000 years old and modern R1a; both are found in the vicinity Neaderthal, both share Neaderthal genes. However the difference between Western Siberia K-M526 and modern Siberian R1a; the segments are longer. Authors dated the ancestoral contact occurred @ 52000-58000+/- with Neaderthal. Isn't 60000+/- years close to IJK/ F time frame ?
"Assuming I'm correct that K-M526 originated outside of SE Asia (which I am) and at least a couple thousand years passed between the origin of K-M526 and its introduction into SE Asia, all of the major branches (those leading to NO, P, K2b1, K2c, and K2d) would already have been in existence prior to their arrival in SE Asia".
ReplyDeleteThat is an absolutely impossible scenario. Do you mean to say that K2a, K2b, K2c and K2d sorted themselves out into the separate islands in the Sunda chain having arrived there full formed? That is a crazy suggestion. And Ebizur has already showed that 'a couple thousand years passed between the origin of K-M526 and its introduction into SE Asia' is incorrect. The diversification within the whole IJK haplogroup was rapid, implying its expansion was rapid.
"There should not have been any significant autosomal difference among the ancestral K2 population, the ancestral K2a population (or even the ancestral NO population), and the ancestral K2b population because these all emerged within an extremely short timespan".
Yes. The branches did not arrive in the regions they are now found already formed.
"the geographic location and age of the Ust-Ishim specimen do show that some early descendant of K2 ended up in Western Siberia".
The date of 45,000 years indicates the Y-DNA molecular clock is calibrated incorrectly. The date is quite consistent with K's presence in SE Asia. Certainly humans were in Timor by that time and in Australia some time before then.
"I do not know the internal structure of each of the Australasian-specific subclades of K2 in sufficient detail to comment on whether they show evidence of significant bottlenecks".
Maju had a list at his blog 'For what they were ...' but I can't find it at present. It is very revealing.
"Yes, a lot of branches are dead. The thing to explain here is why SEA/Oceania has the descendants of more brothers close to the root, and West Eurasia has the descendants of only one brother close to the root".
Yes. That is a huge problem for the above theory.
"56.3% of Papuans have P (3 subclades)"
And Australians have 6 (P5, P6, P7, P8, P3a and P4b), the last 2 shared with New Guinea (p3b and P4a). That leaves New Guinea with just 1 separate P P1''10), that shared with the Philippines. P is not a primarily New Guinea haplogroup.
"K2b is much more common in PNG than it is in Australia"
Yes. As I said K2b is not primarily an 'Australian' haplogroup.
I guess we’d need to decide what is more parasiminou: that undifferentiated K diffused throughout SE Asia then differentiated into different subclades in each locale, or that already differentiated subclades differentially colonized different parts of SEA. Iit is also possible that the colonizers carried both basal K, K2 as well as further downstream clades; but this wuld not explain why in each case it was younger daughter clades which became fixed and not older , more basal clades.
ReplyDeleteI haven't had my first coffee yet, so keep that in mind, but isn't this in line with statistical parsimony?
ReplyDeleteOlder, basal clades are less common than younger daughter clades, and since only a small fraction of clades survive, then these are more likely to be the younger clades.
Another point to keep in mind is that ancient populations in places like the Mammoth Steppe were more likely to suffer massive founder effects, because competition was probably more intense than in the tropics, due to more limited resources and harsher conditions.
"ancient populations in places like the Mammoth Steppe were more likely to suffer massive founder effects"
ReplyDeleteAnd, in the case of 45,000 year old Ust Ishim, subject to extinction as the coldest part of the ice ages was yet to hit.
"I guess we’d need to decide what is more parasiminou: that undifferentiated K diffused throughout SE Asia then differentiated into different subclades in each locale, or that already differentiated subclades differentially colonized different parts of SEA".
The first is far more likely to be the case. It is difficult to envisage a scenario where the second alternative holds.
"Older, basal clades are less common than younger daughter clades, and since only a small fraction of clades survive, then these are more likely to be the younger clades".
But the nodes still exist.
"some of the nodes in the ISOGG tree are more evolutionarily significant than others".
I suggest that each node implies at least some level of expansion which makes the K nodes just as interesting as the C and F nodes.
"Maju had a list at his blog 'For what they were ...' but I can't find it at present. It is very revealing".
Here it is:
http://forwhattheywereweare.blogspot.co.nz/2014/06/y-dna-macro-haplogroup-k-m526.html
Summing up:
K2*-M526 in Sumatra and Sulawesi, possibly actually separate haplogroups in the two islands.
K2a-M214 (NO if you prefer) China.
K2d-P402 Java
K2c-P261 Bali
K2b-P333 splits in two, as follows:
K2b2 (P if you prefer) South and Central Eurasia.
K2b1-P397 splits into four, as follows:
K2b1*-P397 Eastern Indonesia, Malays, Papua.
K2b1b-P336 West Timor, Borneo, Eastern Indonesia.
K2b1c Aeta of the Philippines.
M-P256 Malays, Melanesia, Eastern Indonesia.
K2b1a-P405 splits into four. Except for S each haplogroup is not assigned a separate nomenclature:
K2b1a^ Sumba (Next to Timor). S-M230 New Guinea, Philippines, Eastern Indonesia.
K2b1a-P60 New Guinea and Australia.
K2b1a-P79 Melanesia and Timor.
K2b1a-P315 both branches are 'Papuan'.
I hope you find that helpful Ebizur. But everyone can see that K2b2-P295 is very much the odd one out in that bunch. Surely it is extremely unlikely that every other K2 haplogroup entered SE Asia in an already differentiated form.
@ Davidski
ReplyDeleteWhy should older, basal clades be more likely to drop out of a pop than younger ones? Basal to what? P* is basal to R, but all the clades in R are also basal to the branch that led to P*. The point is not that SEA carry clades basal to R; the point is they carry many clades, including P*, period.
The thing that matters is representatives of how many trees survive. Just because an individual had a newfangled mutation in R doesn't mean that he is more likely to pass on his Y-DNA than another who had a newfangled one in another clade in K2. What the rest of the pop carries does not matter. Once a bottleneck event skews the distr so that descendants of a single brother R are overrepresented, yup this makes descendants of other clades less common, making them more likely to be lost to drift; but that was because of a powerful bottleneck, the existence of which is what we are trying to puzzle out the first place.
@John Smith
Good points, thanks.
According to Reich the Andamanese split from ASI 34,000-25,500 yrs ago (an estimate from the autosome), with no gene flow from Eurasia afterward, and the Andamanese carry D. I think this makes it incredibly likely that ASI also carried D, but this was lost in the pop turnover in India, with exp of H, L, T, R, J and whatnot.
I think the Toba eruption theory one of the more plausible objections to what I am suggesting. However, this would imply that India was less diverse than SEA/Oceania as a whole across the autosome, and from what I know at least this is prob not the case. SEA has many basal K2, but K1, which is hap LT, is in India only, and India is the only region with descendants of K1 and K2 both.
Going right back to the beginning:
ReplyDelete"Maybe this throws out the K2b developing in SE Asia, mess".
From the regional breakdown of the phylogeny Maju supplied it is difficult to come to any conclusion other than that K2b2 developed anywhere other than SE Asia. It's a simple matter to reconstruct the history of K2's expansion into and through SE Asia.
A population containing K2-M526 must have entered East Asia from South Asia and spread down the Malay Peninsula. K2a remained in the north while drift gave rise to separate haplogroups in each island to the south: K2* in Sumatra, K2d in Java, K2c in Bali and K2b Timor (across Wallace's Line).
From that last region a population containing K2b1-P397 moved up the western margin of Wallacea giving rise to K2b1b in Borneo and K2b1c in the Phillipines, while M and K2b1a remained for a time near Timor. The Eurocentrists amoung us believe that K2b1's 'relation', K2b2, was somehow miraculaously transported from western Wallacea to somewhere near Europe.
From near Timor (or Sumba) a population containing K2b1a-P405 entered New Guinea forming several haplogroups around and through that island. From there members of K2b1a-P60 entered Australia and K2b1a-P79 entered Melanesia. Much more recently members of the last haplogroup, as well as members of M, moved further out into the Pacific.
It's amazing what a little 'faith in interpretations of modern Y-haplogroup data' can reveal if you're prepared to actually look. Can any of those who wish to see K2b having originated anywhere near Europe offer any scenario at all for the history of the haplogroup without invoking extensive extinctions for which there is no evidence?
"K2b2, was somehow miraculaously transported from western Wallacea to somewhere near Europe".
ReplyDeleteSorry. Wrong way round. It is easy to see how a population that has just learned to cross open sea was able to expand rapidly from SE Asia to somewhere near Europe, 'miraculaously' or otherwise. What is far more difficult to see is how members of K2 both upstream and downstream of K2b2 could have miraculously travelled from somewhere near Europe to SE Asia yet maintained distinctive lineages that sorted themselves out regionally when they arrived.
@ All
ReplyDeleteHere is something Shaikorth spotted in Razib's blog which I find intensely interesting.
"It’s very likely Tianyuan was not really closer to East Asians than Ust-Ishim. U-I branches with East Eurasians with 100% bootstrap support, but branches off before Oceanians and Asians (supplements, page 57-58). This is similar to how Tianyuan branched, which was the main basis for that paper claiming it’s proto-East Eurasian. I think they did not realize during the Tianyuan paper’s writing that this could be simply because Tianyuan lacked the Middle Eastern/Basal/African admix of Europeans used in the tree (French and Sardinians).
Formal testing (f-statistics from Rasmussen et al 2014) which were not done in Tianyuan’s own paper show no difference between Tianyuan’s relation to East Asians and Europeans.
http://www.nature.com/nature/journal/v506/n7487/fig_tab/nature13025_SF5.html
(F & G)"
If you check the Tianyuan Axis in the final two diagrams, Europeans are as related to Tianyuan as Asians.
As Tianyuan is 40000 yrs bp, this might mean that the Ur-Crown-Eurasian-Population that existed in South/SE.Asia may have persisted for some time in sending multiple failed radiations into temperate Eurasia, of which Tianyuan is one and Ust-Ishim another.
All credit to Shaikorth.
I think one would have to resort to special pleading to argue that macrohaplogroup K arose somehwere in northern Eurasia.
ReplyDeleteI remember the same thing was said about R1a a couple of years ago.
ReplyDeleteThe supposedly high diversity of R1a in the Indus Valley doesn't look so crash hot nowadays though.
@ Davidski
ReplyDeleteHap diversity is only one thing going for this argument.
That was different . The high diversity based on STRs is very different reality to SNP diversity, although I am not proclaiming any definite conclusions
ReplyDeleteAnyway from the study authors "This individual would then represent an early modern human radiation into Europe and Central Asia that may have failed to leave descendants among present-day populations".
Explains the early occur of K so far north occurred during a warm phase of pioneer colonisation to the north; one whicj was tentative at best.
But the authors don't imply anywhere that Usti was a migrant to western Siberia from southeast Asia.
ReplyDeleteRather, they imply that his ancestors were sitting in the Middle East only a few thousand years earlier, because he had long chunks of Neanderthal ancestry.
@ Davidski
ReplyDeleteThe authors traced it to the OOA migration, which Indians had to go through.
Yes, I know, but his not too distant ancestors didn't mix with Neanderthals in India or southeast Asia, but in the Middle East.
ReplyDeleteSo unless they went running around Eurasia for a few thousand years afterwards, it seems that they basically moved straight into western Siberia from the Middle East. Some of their relatives probably migrated south and east from Central Asia or Siberia, mixing with Denisovans along the way. If so, that's how south and southeast Asians came to be.
@ Davidski
ReplyDeleteSo no Siberia?
The Middle East is more plausible I guess. Two things the authors need to do: they need to confirm that the Neanderthal admix happened in a single pulse, and not in a few that they averaged over.
Otherwise the Middle East is still too far West IMO. And this still does not change the fact that it was K* found in Siberia, not K2*; so the radiation of MNOPS is still up in the air.
From the present study's Supplementary Information 9:
ReplyDelete"The Ust’-Ishim Y chromosome sequence clusters with the K(xLT) haplogroup. The Ust’-Ishim sequence shares all the mutations common to the K macrohaplogroup and has one additional specific mutation rs2033003/M526 which defines the group K(xLT) (Figure S9.1, blue part). The Ust’-Ishim Y-chromosome carries no additional mutations belonging to any of the sub-haplogroups of K(xLT); however, there are 6 additional mutations that are not observed in the 23 present-day humans to which we compare. K(xLT) currently has a 54% frequency in Eurasia (Lippold et al. 2014, bioRxiv doi: 10.1101/001792)."
K(xLT)-M526 is synonymous with K2-M526 or MNOPS-M526.
@ Ebizur
ReplyDeleteThanks for pointing this out.
@ Davidski
Sorry my bad.
Not so. The recent paper reveals the phylogeny in considerable detail. The link to Maju's blog above will tell you everything you need to know on the subject.
ReplyDelete"I remember the same thing was said about R1a a couple of years ago".
The R1a phylogeny has been considerably refined since those days though. In fact it was that very refinement that showed R1a could not have originated in South Asia. It is difficult to imagine that the K2 phylogeny will be much altered from the latest research.
"Explains the early occur of K so far north occurred during a warm phase of pioneer colonisation to the north; one whicj was tentative at best".
Makes sense. I still can't see why anyone has a problem with the date. Humans, presumably including ones carrying Y-DNA K2, had certainly crossed Wallace's line by then.
"they imply that his ancestors were sitting in the Middle East only a few thousand years earlier, because he had long chunks of Neanderthal ancestry".
You seem to forget that the Y-chromosome is not the whole genome.
Yeah, I didn't get exactly from the paper that the authors explicitly think that the Neaderthal admixture occurred in the Middle East. Even if it had, this doesn't preclude a subsequent migration path from central - southern Asia to the north.
ReplyDeleteIndeed, to go from Middle East to north Eurasia, you'd either have to cross the Caucasus - a significant geographic barrier for early homonids, or the deserts of Khorezm, etc, which were very much a no mans land until recent prehistory.
Id guess a more eastern origin point for the treck north,although I'd have to research further into the palaegeological conditions 45 kya to not risk falsely extrapolating back from today's climactic conditions.
Ebizur. Thanks for the clarification. Means the Ust’-Ishim Y-chromosome is not ancestral to P but an extinct side branch formed after LT but before the diversification of K2. Presumably that places its origin somewhere in South Asia.
ReplyDeleteRyukendo,
ReplyDeleteThanks for bringing up Shaikorth's excellent observation from the Rasmussen paper that Tianyuan shares as much drift with modern-day Europeans as with modern-day Chinese. Ust Ishim and Tianyuan may not be dead ends as you suggest. Ust Ishim may well be an ancestor to WHG and to East Asians, perhaps even an ancestor of Tianyuan.
Just to add to my previous comment, Ust Ishim's Y-haplogroup may be extinct but he could still have left descendants that survive to the present.
ReplyDelete@ Balaji
ReplyDeleteNote also that they are far from bedouin and palestinian.
That he left descendants cannot be ruled out, but it is almost certainly untrue that UI was an ancestor to Tianyuan IMO.
Tianyuan, like Ust-Ishim, is a representative of a population that lay at the root of East and West Eurasian. To say that this pop was the ancestor of WHG-ANE and ENA is prob true, and to say both Tianyuan and Ust-Ishim are desc of this pop is also prob true. But to say that 1) Tianyuan and U-I are old ancestors of modern pops is prob untrue, and that 2) U-I is an ancestor of Tianyuan is prob also untrue. 1) would mean that we root U-I and Tianyuan much more securely in either ENA branch or the WHG-ANE branch than at present, when U-I appear to root outside altogether and TY may root outside when Basal is factored in, as Shaikorth mentioned. Also, the Euro-Asian divergence time is much shallower compared to these times, which prob means later waves were the ones that really mattered. 2) would mean that East Asians emerged from a northern pop, and we have covered that already.
One thing that troubles me is that Tianyuan roots basal to ENA as Shaikorth mentioned, which he shares with U-I. But he, unlike U-I, is symmetrically related to Present Day Europeans vs Asians in f-stats, when U-I is more related to Asians than present day Europeans due to Basal contribution in European, and is only symmetrical w.r.t La-Brana vs Dai.
ReplyDeleteReally not sure what is going on here.
terryt,
ReplyDeleteI think all that can be said with certainty at this point is that the Ust-Ishim specimen's Y-DNA belongs to K2-M526(xK2b). Whether or not it is located near the base of a branch of K2 that leads to K2a (and thence NO), K2c (now found on Bali), or K2d (now found on Java) is not entirely clear yet. Maybe also K2e-M147 (now found in South Asia), but Maju seems to think that K-M147 is the same as the pre-NO ("X-F650") haplogroup that has been identified by Magoon et al. (2013), which may just as well (or perhaps more appropriately) be called K2a*-F650(xNO-M214) for now. Does anyone know Maju's basis for equating K2e-M147 and K2a*-F650(xNO-M214)?
Thank you for your link to Maju's blog post, by the way. However, I really am looking for a dendrogram based on a larger number of Australasian K samples with the number of SNPs for each branch length (space between nodes) indicated so that I may perform TMRCA estimates.
Also, this is rather off-topic (but actually somewhat relevant to the subject of this thread if one considers the broader picture): have you seen the paper by Chaubey et al. (2011), "Population Genetic Structure in Indian Austroasiatic Speakers: The Role of Landscape Barriers and Sex-Specific Admixture"? They studied many Mundaic/Kolarian ethnic groups of India (in addition to the Khasis, the Tibeto-Burman Garos, and the Baigas, who have been associated by some authors with Kolarian peoples despite currently speaking Indic languages/dialects that are supposedly influenced by Dravidian Gondi), and found that O2a1-M95 reached maximal frequency among the most isolated of the Kolarian tribes, such as Bonda (40/42 = 95%), Gadaba/Gutob (24/27 = 89%), Juang (48/54 = 89%), and Birhor (55/62 = 89%). The Bonda/Bondo/Remo and Gadaba/Gutob speak similar languages of the Koraput Munda group. The Juang language is usually considered to be closest to the Austroasiatic Kharia language, but the Kharias clearly have been influenced more greatly by outsiders (a subgroup of them has spoken an Indic dialect for a considerable time already), and the Kharia sample notably has a much lower proportion of O2a1-M95 (only 18/37 = 49%). Birhor (literally "forest people") are the most isolated and primitive speakers of the North Munda/Kherwarian group, and correspondingly have the highest frequency of O2a1-M95 among peoples of that language group, who are otherwise probably the most civilized and well-integrated into Indian society of the Mundaic groups. (For example, the Santhal sample contains only 7/20 = 35% O2a1-M95.)
If you are not familiar with the Bondas, Juangs, Birhors, etc., I urge you to do some research about them before claiming that O2a1-M95 has spread from Northern China since the Neolithic. I am sure they are nothing like what you would expect. The more isolated and primitive tribes have been hunter-gatherers until very recently.
http://www.flickr.com/photos/alfreddaniels/6267168059/ (a Birhor mother and child)
http://upload.wikimedia.org/wikipedia/commons/6/66/Inde_bondo_8593a.jpg (a Bonda girl)
http://en.wikipedia.org/wiki/Juang_people#mediaviewer/File:Juang.jpg (an engraving of a Juang man and woman with modern hand coloring)
Yes, I agree with above two or so discussions. Finding out exactly where the K2 from Utsy lies is essential for any further discussion as to where (approximately) we should hypothesize that K arose. Whether Utsy's k is an entirely separate clade or one amongst the variety of extant K2s will suggest very different scenarios.
ReplyDeleteIf Utsy's K is paraphyletic, and thus extinct, his lineage could then indeed represent a dispersal from further west radiating across central , southern, and SE Asia as well as Siberia. The K which went toward SE Asia survived, proliferated and diversified (in SE Asia's favourable tropical and subtropical niches), whilst that in central and northern- central Asia became extinct, drifted out, got replaced by fresh waves from the south or west (eg by newer Hg R' populations).
Whatever the case, this find of Utsy was almost certainly one of the earliest pushes of AMHs into Siberia, according the R C14 dating by Russian archaelogists.
(To piggy-back)
ReplyDeleteWe should also not forget K1 (ie L & T); found more or less mostly in south Asia; which could gives further clues as to trajectories.
Moreover, as above, it is relatively clear that north Eurasia was simply not colonized until 40 kya. Must have something to do with climate and ability to adapt to it. Whether Europe, Siberia, or NE Asia, northern Eurasia was colonized much later than south.
Had we a better fossil record to show tangible proof, it might be that AMHs existed in Southern Eurasia (sth Levant to SE Asia) for thousands of years before moving north.
Exactly where this north -south line was is an issue to more clearly elucidate.
Western Asian and Northern Caucasus are quite old and should not be discounted to the ancestral genetic contribution to Ust-Ishim. If Ust-Ishim sample does in fact have Neaderthal incorporated in his dna it could only have come from a couple of regions either through direct mating or actual splitting of line. In theory Malayasia would have been a genetic sink 60,000 B.P.from Western Asia or South-Western Asia, for Neaderthal. I have seen no studies implying Malayasia connected with the spread on Neaderthal genes. On the contrary, the region in North Caucasus with a wide range of ancient morphology;already experienced a wave of early homo-[Dmanisi] 1.8 million years prior. http://www.sciencemag.org/content/342/6156/326.abstract
ReplyDeleteThe region also has samples that are related to Neanderthal which have been connected to much of the human population today[implied study] . I would not be surprised if the Ust-Ishim sample was also connected to the Mesmaiskaya/Adygei
http://en.wikipedia.org/wiki/Mezmaiskaya_cave
http://www.nature.com/nature/journal/v505/n7481/fig_tab/nature12886_ST2.html
Adygei
http://anthropogenesis.kinshipstudies.org/2014/10/ancient-ust-ishim-dna-as-seen-from-the-americas/
"One thing that troubles me is that Tianyuan roots basal to ENA as Shaikorth mentioned, which he shares with U-I. But he, unlike U-I, is symmetrically related to Present Day Europeans vs Asians in f-stats, when U-I is more related to Asians than present day Europeans due to Basal contribution in European, and is only symmetrical w.r.t La-Brana vs Dai."
ReplyDeleteI can't really say whether that is actually an important difference without a better Tianyuan sequence. The D-test of Ust-Ishim and the f3 test of Tianyuan are a bit different, and there's a clear quality gap which could also cause discrepancies.
@mike Thomas
ReplyDeleteSince K1 group is older than "Ust" and seems to have origins stretching from eastern persia to India, then does it make sense that "Ust" waited these areas.
If Ust has m526 and its origins are in SEA as per Karafet does it not make sense that his line came via an area stretching from Burma to Malaysia?
There are 2 timesframes to consider, Ust and M526
Huh, maybe I was wrong. It looks like we have a date for the divergence of K2b from K2a, and for pre-P1 from P*, but not for the origin of other K2 subclades (unless in Karafet et al 2014 which I have not yet read).
ReplyDeleteIJK-M522 ~54 kya.
K-M9 ~53 kya
K2-M526 ~50 kya.
P-295 ~48 kya (!).
I hadn't realized the split of P was *that* early. This leaves the door open for P to have already existed during the initial expansion of K2, with one branch of it ending up in SE Asia with the most successful population of K2, and the other one happening to survive in Central/South Asia to expand later. That Indian pre-NO branch, if it checks out, would be another remnant of the original wave, since overrun by H, R, LT, C1, etc.
If the SE Asian sub-branches of K2 turn out to be as old as or older than P, then SE Asian origin remains most likely. However, if they are somewhat younger, then they could represent later expansions of the pre-existing basal diversity of K2.
So the boring theory of P1 might be correct after all. Bah. Probably we will find out before long.
@ Shaikorth
ReplyDeleteThanks for the info. Makes more sense now.
@ Ebizur @ Balaji @ Mike Thomas @ Shaikorth
The U-I paper comes up with a slow mutation rate. What does this do to the previous estimations of East-Asian/European divergence times?
@ Capra
My logic is very similar to yours. PQR(K2b2) splitting off in SEA/East India or splitting off in further west is equally likely, as K2b is bifurcated into PQR and K2b1, so each region contains one branch only.
But clades outside of K2b, incl. K2*, K2a, K2c, K2d, K2e, NO as well as the position of K1 all imply that K2 itself must have had an eastern center, and that other pops to the NW and NE are subsets of this K2 diversity.
In particular, the K2e-NO connection, the P*-(P1(QR)) connection, and the C((((1,6)5)3)C*(C2)) branching order all imply that the period of 'dense branching' in the C-K pop occurred in the (WHG/ANE)/(ENA) root pop, and that this should be south of the Himalayas or somewhere close by.
"Ust Ishim's Y-haplogroup may be extinct but he could still have left descendants that survive to the present".
ReplyDeleteExactly.
"Does anyone know Maju's basis for equating K2e-M147 and K2a*-F650(xNO-M214)?"
His own imagination, I suspect. The paper itself is very tentative in placing K-M147 anywhere in particular.
" I really am looking for a dendrogram based on a larger number of Australasian K samples with the number of SNPs for each branch length (space between nodes) indicated so that I may perform TMRCA estimates".
Do you have access to the original Karafet paper? I don't but perhaps it has the information you require. Thanks for the mention of the Munda paper. I had seen it but notice once more in the conclusion:
"The presence of a significant (approximately one-quarter) southeast Asian genetic component among Indian Munda speakers is consistent with this model, implying their recent dispersal from southeast Asia followed by extensive admixture with local Indian populations. The strongest signal of southeast Asian genetic ancestry among Indian Austroasiatic speakers is maintained in their Y chromosomes, with approximately two-thirds falling into haplogroup O2a. Geographic patterns of genetic diversity of this haplogroup are consistent with its origin in southeast Asia approximately 20 KYA, followed by more recent dispersal(s) to India".
Even though some Munda-speaking groups could hardly be called 'Neolithic' we know it is reasonably easy for farming groups to revert to hunter-gathering in appropriate conditions.
MT,
ReplyDelete"Correct me if I am wrong, but you have used the same arguement about diversity of R1a in EE as Rukendo (sorry, I know I keep misspelling it) has used about K more broadly in central-southern Asia . If we are going to be consistent, we either have to reject or accept such an arguement for all cases, and not use it selectively (?)"
Impressionistic phylogeography is fine as far as it goes. It just doesn't go as far as many internet commenters believe. This should not be hard to understand: Y chromosomes present in living humans coalesce into fewer and fewer lineages going back in time. Go back 70,000 years or so, and we have a sample size of one in Eurasia. We can infer nothing about population movements in Eurasia prior to 70,000 years ago from extant Y chromosome variation. Moving forward from 70,000 years ago, we do not instantly gain statistical power to make the sorts of inferences many here imagine they're able to make. The information just isn't there. Closer to the present, the tree contains more information, making phylogeographic inferences in the absence of formal statistics more often justifiable.
Ebizur,
"However, looking both upstream and downstream from the K2 node, I see a lot of "evolutionarily insignificant" nodes that are not associated with any clear bottleneck. K, IJK, HIJK, GHIJK on the one side, and at least K2b2/P, NO, O2, and O3 on the other; I do not know the internal structure of each of the Australasian-specific subclades of K2 in sufficient detail to comment on whether they show evidence of significant bottlenecks. Considering the phylogenetic position of K2 within this sort of dense branching pattern, how can you claim that "all but a few early branches of K-M526 died out everywhere"? Your claim seems to be inconsistent."
There are likely to have been hundreds or thousands of K-M526 descendants 1000 years after K-M526, and obviously we do not have hundreds or thousands of other haplogroups today in the same temporal plane as P-P295. This is just how coalescents work.
rk,
"This doesn't work because population genetics."
Population genetics, the branch of statistics, and population genetics, the rk/terryt pasttime that involves looking at pie charts and pronouncing confusedly on the internet, aren't really the same thing.
terryt,
"A population containing K2-M526 must have entered East Asia from South Asia and spread down the Malay Peninsula. K2a remained in the north while drift gave rise to separate haplogroups in each island to the south: K2* in Sumatra, K2d in Java, K2c in Bali and K2b Timor (across Wallace's Line). . . .
From near Timor (or Sumba) a population containing K2b1a-P405 entered New Guinea forming several haplogroups around and through that island. From there members of K2b1a-P60 entered Australia and K2b1a-P79 entered Melanesia. Much more recently members of the last haplogroup, as well as members of M, moved further out into the Pacific."
This is another area where you're confused. The tree displayed in the Karafet paper is not based on complete sequences, so you're assuming information about timing that's not actually there. For the Oceanian section of the tree, all Karafet had to go on was a couple of low coverage genomes and markers discovered incidentally while typing a handful of relevant known SNPs (the older ones of which had themselves been discovered through surveys of only small fractions of the Y chromosome): "A set of 13 mutations was extracted either from sets of known polymorphisms by comparisons that include low-coverage genome sequences 18 or discovered here by resequencing when genotyping previously known mutations (Supplementary Table S1)"
So it's likely many of the polytomies will be resolved with additional sequencing. For example, K2c, K2d, and Sumatran K2* could all turn out to be part of a single lineage. And there will likely also be additional consolidation under P397.
@ n/a
ReplyDeleteWhatever your motivations might be, I would have hoped that you entered into this debate without that tone. I am pretty sure that neither you nor me reach our conclusions by looking at 'pie charts'.
1) You have raised many good points, but I would actually really like to hear how you would reconcile all of this with the relatively clear picture of autosomal branching orders in Eurasians that we have now, and how this constrains our placement of various Y-radiations in a way they have not previously. This could be ground for some actually productive noise. After all the Y chromosome does not exist in a vacuum.
2) Ebizur said to you:
"I would also suggest that n/a might have one foot in the trap of accepting the ISOGG or Full Y-DNA tree as a Bible and reading too much significance into it because of having lost sight of the arbitrary nature of the nomenclature."
Precisely; in the tree every branch at every level has equal taxonomic significance at that level. Older clades might be distinguished by one mutation; younger clades may also be. So the same statistical arguments used to support OOA from Y-DNA apply as strongly to K. The first branches from Ychrom adam are theoretically speaking simply a much older replica of the branching in K. While there are not thousands of descendants of K, neither are there billions of descendants of DE or CF, but we make very solid judgements based on A, B , DE and CF already. They are in no way different in interpretation just because the earliest are in some way 'deeper'. The 'stochastic death' scenario does not scrub information in the way you described from the tree then, and it does not now.
In any case, 'population genetics' could hardly involve espousing a belief in dynamics that can be taken to support an into-africa scenario.
These are the two most relevant points raised in the above arguments, and you have not addressed either.
Let's just wait and see. My conclusions are not far from those reached by Karafet herself.
rk,
ReplyDelete1) You have raised many good points, but I would actually really like to hear how you would reconcile all of this with the relatively clear picture of autosomal branching orders in Eurasians that we have now, and how this constrains our placement of various Y-radiations in a way they have not previously.
There's nothing that needs to be reconciled. If I understand what you're trying to argue, you're claiming that non-"Basal Eurasian" Eurasians must have originated in East Eurasia and "Basal Eurasians" in West Eurasia. This is not the case. "Basal Eurasian" is likely of North African origin, and Eurasian is Eurasian.
"2) Ebizur said to you:
"I would also suggest that n/a might have one foot in the trap of accepting the ISOGG or Full Y-DNA tree as a Bible and reading too much significance into it because of having lost sight of the arbitrary nature of the nomenclature."
Precisely; in the tree every branch at every level has equal taxonomic significance at that level. Older clades might be distinguished by one mutation; younger clades may also be."
I'm not completely clear on what you're trying to argue here, but the bottom line is that everyone in NO descends from a single man who lived not long after the man in whom K-M526 occurred. No matter how many branches of O exist, or how early they coalesce (and they in fact coalesce many thousands of years after K-M526), for the period immediately after K-M526 they represent only a single data point.
So the same statistical arguments used to support OOA from Y-DNA apply as strongly to K.
No. Again, Alan Templeton's rule of thumb is that "the evolutionary tree of genetic variation for a particular DNA region is informative only for the second half of the tree's time depth". The full time depth of the Y chromosome tree is something like 200,000 years, while the time depth of the Eurasian portion is something like 70,000. Which places the split relevant to "out of Africa" in the informative half of the tree. This is not the case for K-M526 (which occurred around 50 kya) with respect to Eurasian Y chromosomal variation.
That said, many of the arguments for out of Africa based on uniparental markers were also confused (e.g., many assumed lack of archaic Y or mtDNA indicated lack of archaic admixture), and differences in effective population size certainly play a role in the African vs. non-African levels of genetic diversity. Into Africa events in the past also are not ruled out by the Y or mtDNA phylogeny, and I think it's as likely as not (based on Denisovan and Neanderthal mtDNA) that the human Y and mtDNA lineages trace back to Eurasia at some point prior to their more recent coalescences in Africa.
My conclusions are not far from those reached by Karafet herself.
ReplyDeleteWhat Karafet et al. actually say:
In sum, our results support the hypothesis of a Southeast Asian/
Oceanian center for the diversification of Oceanian K-haplogroup
lineages and underscore the potential importance of Southeast Asia as
a source of genetic variation for Eurasian populations. We propose
that the patterns of Y-chromosome variation in the K haplogroup
reflect a process of population fragmentation, likely associated with
the early expansion of modern human populations into island
Southeast Asia, and possibly also with rapidly changing sea levels, 30
followed by a subsequent dispersal from the same area. While limited
in their inferential power, our results warrant the exploration of a
demographic model that includes a population expansion from island
Southeast Asia into mainland Asia.
A demographic model involves math and formalized assumptions (as opposed to blind assertions about how strong the evidence is). Karefet et al. are well aware they have not proved anything. But it never hurts to exaggerate the importance of your results when trying to get a paper published. The fact that they didn't test any formal model themselves suggests they are probably not all that confident in the scenario they lay out.
For comparison, here is the very sound conclusion of a Karafet and Hammer paper from 1998 (pdf):
ReplyDeleteFuture Directions
One of the clear advantages of a nested cladistic
approach using haplotype trees is that spatial and tem-
poral patterns of genetic variation can be examined con-
currently. Notice, however, that natural selection is not
directly addressed in our present application of the nest-
ed cladistic design. Therefore, studies of variation at
other nonrecombining segments of the genome are needed
to confirm our inferences as well as to clarify the
biocultural and demographic contexts responsible for the
differing evolutionary trajectories exhibited by
maternal- and paternal-specific data. In the short term we need
to examine as many single-gene trees covering as many
parts of the genome as is logistically feasible.
Concordant gene trees may then lead to the construction of
robust population trees.
In detecting recurrent and historical events through
nested cladistic analysis, recall that the timescale is de-
termined by the coalescence time of the DNA region
under examination. It is important to keep this consid-
eration in mind when comparing data from independently
segregating segments of the genome. [. . .]
In addition, it is imperative to look for concordance
among genetic, linguistic, ethnohistoric, fossil, and/or
archeological data sets to lend support to particular hy-
potheses about human population history, because genetic
data alone cannot inform us completely about that
history. Ultimately, we need to develop model-based
tests applicable to different kinds of data that have the
power to discriminate among alternative hypotheses
and, specifically, to distinguish among the various com-
peting models of human population origins. [. . .]
In addition to the aforementioned indirect, infer-
ential procedures for reconstructing human evolutionary
history, the portentous replicated recovery of Neander-
thal mtDNA detailed in Krings et al. (1997) has ushered
in a much more direct kind of evidence with the poten-
tial for actually distinguishing among competing evo-
lutionary scenarios (Ward and Stringer 1997). This is,
indeed, an exciting time for students of human evolu-
tion!
As a small addition, over on John Hawkes' blog, he argues that ties individual left no descendants. Ie it was a failed colonization attempt into northern Eurasia.
ReplyDeleteFinally, palaeoanthropologists, whether believers in a 60 K OoA AMH dispersal, or an earlier, 100K one, argue that SE Asia was colonized later, with a 'lag' between HUman arrival in India and their subsequent movement in SE Asia, which had a very different ecology.
@ n/a
ReplyDeleteI am very glad that you have responded to my points directly in a civil manner. I am as curious as you are; I have no problem changing my mind when I see my evidence being challenged by something robust.
" "Basal Eurasian" is likely of North African origin, and Eurasian is Eurasian."
I agree that North-Africa vs Eurasia instead of ME vs India is a possibility; I have myself argued this elsewhere. My only qualm is that 1) TianYuan is also at the root, but very far east very early, 2) that the tree in ENA is (Papuan(ASI, Dai)) and not (ASI(Papuan, Dai)), rooting the whole ENA half of the (WHG-ANE-ENA) clade east of India, and so moving the root of the parent clade east as well, since we have only 1 other branch, and that 3) autosomal diversity seems to peak in Eurasia in India.
"A demographic model involves math and formalized assumptions"
It seems that autosomal DNA is far more amenable to this type of analysis than Y-Haps or mtHaps. I have a feeling that, with a few more ancient genomes and some more ancient pops made apparent, we would be matching Y- and mt-haps to autosomal trees, instead of what we used to do--matching autosomal ancestry to Y-trees, which was always going to be a rather futile exercise.
In any case, I think you would agree with me that ME, India, or SEA are the most possible candidates for K, while Siberia is more unlikely than any of the three above.
I will now add Alan Templeton to my reading list.
@ MT
I am aware of that; I am also aware of the fact that SEA and India had very different archaeologies for a very long time, and that this was always kinda puzzling to archaeos, which led to the 'perishable bamboo theory' of tool use.
However it also seems that the ENA-(WHG-ANE) divergence is much later than the above timeframes, and that early radiations like U-I and Tianyuan left no/few descendants, like you have highlighted. I suspect that we are all the descendants of only the most recent radiation of the ur-crown eurasian pop.
However it seems like the slow rate of mutation that evidence is building up for might just hammer this on the head. Question to anyone who can illuminate this.
People used to think that Europeans diverged from East Asians only as recently as 20,000 years ago. But if the mutation rate is half what was previously assumed, then the divergence must have started 40,000 years ago, or in Tianyuan's time.
ReplyDeleteOn another topic, if BEA was the result of a second OOA migration, or if it was some people from the original OOA who were somehow sequestered in North Africa, then people with BEA (Europeans, Middle Easterners and people of the Caucasus) would have markedly less Neanderthal admixture than other non-Africans.
The Ust-Ishim genome might be a good way to probe for BEA in different populations. Table S11.1 in the Supplementary Information of the Quomei Fu paper gives a D value of 0.0258 comparing French and Onge to Ust Ishim. But for Sardinians and Onge compared to Ust Ishim this value is higher at 0.0294. This is because Sardinians have higher BEA than French. This method can be used to estimate BEA values for populations that Lazaridis has not published using his values for calibration.
I hope Davidski will find some time to get Ust Ishim's genome and calculate some D statistics!
@balaji
ReplyDeleteEsp about neanderthal, populations like bedouin do not have 80% less neanderthal even though they are approx 80 basal at least. The split point must be in ME, since there is no evidence of nean in North Africa. This is a very good point.
Thank you for clarifying the split time. This must mean that the ancestors of the crown Eurasian branch tips today reached their geographical position later than that time, which prob means that u-i and ty did not contribute significantly to modem pops.
As a final aside, I just found out that Mike Hammer, co-head of the same lab as Tatiana Karafet in the U of A, gave a presentation at FTDNA 2013 on R1b, coming out in favour of a post-neol intro into Europe. During that presentation, they traced the movement of the ancestors of Hap R from Africa to SEA, and the SEA back into West Eurasia. I suspect that they in that lab have been ruminating over this model for quite some time.
ReplyDelete"they traced the movement of the ancestors of Hap R from Africa to SEA, and the SEA back into West Eurasia".
ReplyDeleteThe only possible scenario.
"it's likely many of the polytomies will be resolved with additional sequencing. For example, K2c, K2d, and Sumatran K2* could all turn out to be part of a single lineage. And there will likely also be additional consolidation under P397".
That would significantly alter nothing. Those haplogroups are effectively a single lineage anyway, and K2b2 is still downstream from it.
"they traced the movement of the ancestors of Hap R from Africa to SEA, and the SEA back into West Eurasia".
ReplyDeleteI don't follow. Surely IJK arose in Western Asia , not Africa
Also , when discussing haplogroup K, we have to consider the other East Asian Hg •C".
Needless to say, there is a vast no mans land between current Hg C and Hg DE/ E.
@TerryT
ReplyDeleteAre you essentially arguing macrohaplogroup K arose in SEA?
If so, I'd have to disagree if you follow the brother Clades. Rather , K arose somewhere in the Tigris-Euohrates-Indus alluvials Andover eastward, became a prolific and successful lineage which survives to this day. It's more northern expansion was less successful (!)
@n/a "Basal Eurasian" is likely of North African origin"
Possibly , but this is doubtable at this stage . Nirthern Africa is yet to be confirmed to harbour any ancient AMH fossils, and it's genetic profile is arguable the result of back migration from Western Asia . So, then, this back migration could actually have been a (near-)Virgin colonisation.
"Surely IJK arose in Western Asia , not Africa"
ReplyDeleteYes, but presumably Mike Hammer was considering the very deep origin.
"K arose somewhere in the Tigris-Euohrates-Indus alluvials Andover eastward, became a prolific and successful lineage which survives to this day".
IJK-M523 certainly arose in SW Asia, then K-M9 must have entered South Asia where K1-P326 (LT) arose. But K2-M526 undoubtedly arose in SE Asia where the clades within then K2 arose. That includes K2b-P331 within which K2b2-P295 lies.
@ Mike Thomas
ReplyDeleteThey were following since the OOA migration, so IJK prob did arise in West Asia/India.
IJK as a whole must have arisen in West Asia, but probably in a crown Eurasian population, not a Basal Eurasian one and therefore not too far West. K might have arisen outside SEA, but the most western area for K2 possible is Bengal IMO.
The tree for C now is ((((C1,C6)C5)C3)C*(C2)) (is this correct Ebizur?), where 1 is in Japan, 6 in Europe, 5 in India, 3 and 2 in East and SE Asia, C* in India. Once again we have a pattern of basal clades in India/SEA, and derived subsets diversifying in all of Asia+Mesolithic Europe, but not ME.
Ebizur said: "Of course, this case would require the presence of C-M130, D-M174, or E-M96 in either the Ust-Ishim population or some other very closely related population from which both ENA and {ANE+WHG} are descended". I have a similar idea, that Crown Eurasian pops carried IJK + C + D, but not E + G.
Of course, we must bear in mind n/a's warnings about time depth and loss of information, esp in Northern Eurasia. We can probably exclude scanarios due to lack of autosomal gene flow, e.g. "C6 came to Europe from East Siberia", since this is easy to falsify as Japanese have no ANE and mesolithic Europeans have no ENA. However we cannot exclude scenarios that propose autosomal gene flow, or that postulate loss of Y-Haps from a pop, e.g. "C once existed in the ME".
Balaji,
ReplyDelete"On another topic, if BEA was the result of a second OOA migration, or if it was some people from the original OOA who were somehow sequestered in North Africa, then people with BEA (Europeans, Middle Easterners and people of the Caucasus) would have markedly less Neanderthal admixture than other non-Africans."
Europeans do in fact have less Neanderthal admixture than East Asians.
But the North African source population of "Basal Eurasians" would have been Neanderthal-admixed, to one degree or another. There's evidence of early back migration to Africa, in the form of Y haplogroup E and Neanderthal admixture in sub-Saharans, and ancestry derived from this back migration was probably the primary component in the ancestry of "Basal Eurasians".
rk,
"I will now add Alan Templeton to my reading list."
Here are links to some other material that might be of interest. Joe Felsenstein also has pages with lecture notes and audio for all his courses, including one on phylogenetic inference (which has some discussion of coalescents): http://evolution.gs.washington.edu/gs570/2014/
"As a final aside, I just found out that Mike Hammer, co-head of the same lab as Tatiana Karafet in the U of A, gave a presentation at FTDNA 2013 on R1b, coming out in favour of a post-neol intro into Europe. During that presentation, they traced the movement of the ancestors of Hap R from Africa to SEA, and the SEA back into West Eurasia."
A year before that Hammer had R arriving with Neolithic farmers. Going back another year or so, it was Paleolithic Western European R for him. Now Hammer will get to watch as ancient DNA yet again undermines his chosen narrative.
And where are you coming up with Bedouins being "80 basal at least"?
MT,
"Possibly , but this is doubtable at this stage . Nirthern Africa is yet to be confirmed to harbour any ancient AMH fossils, and it's genetic profile is arguable the result of back migration from Western Asia . So, then, this back migration could actually have been a (near-)Virgin colonisation."
Again, Y haplogroup E and the presence of Neanderthal admixture in sub-Saharans leave little doubt of an early back-migration to Africa, and archaeology, physical anthropology, and Y haplogroup E leave little doubt of a significant North African influence on Epipaleolithic Middle Easterners. It's not plausible that two separate populations existed side by side in the Middle East for 50,000 years (or even 30,000 years, under the rk/terryt scenario) while maintaining the level of distinctiveness found between "basal West Eurasians" and "basal Eurasians". So it's clear there was a geographic barrier. The African component in Natufians is the only plausible source of "Basal Eurasian". No other possibility comes close.
@Terry
ReplyDelete"But K2-M526 undoubtedly arose in SE Asia where the clades within then K2 arose. That includes K2b-P331 within which K2b2-P295 lies."
GIven Utsy's finding, I'd emplore that it is more parasiminous that K2 arose also in Sth Asia, then radiated to east and north. It failed in the north..
@Ruki.
What do you mean by "Crown Eurasians'? ?
n/a
"The African component in Natufians is the only plausible source of "Basal Eurasian". No other possibility comes close."
What is your evidence for this ? (as opposed to basal Eurasian being in ME/ west Asia )
MT,
ReplyDeleteI thought I just got done explaining why "basal Eurasian being in ME/ west Asia" doesn't work. Gene flow within the Middle East would homogenize populations on time scales much shorter than 50,000 or 30,000 years. There must have been a strong geographic barrier, and there's nowhere in West Asia "Basal Eurasians" could have remained genetically isolated from other Eurasians for that length of time.
@ MT
ReplyDeleteAgree that either scenario is most parsimonious, and that it is difficult to tell between them, and both are more parsimonious than any other scenario, but do not preclude the rest, for the reasons n/a/ brought.
Crown refers to the radiation at the tip of a lineage, so 'crown eurasian' refers to WHG+ANE+ENA, the same way that 'crown insects' include beetles and moths but exclude silverfish.
@n/a
There is indeed no way of sussing out formally the amount of Basal vs UHG in ME; however we can get nonformal confirmation. ADMIXTURE generally model Stuttgart as 50% European and 50% Bedouin, where the Bedouin themselves score 100% bedouin and mesolithic European samples score 100% European, and approx 50% Basal is the figure produced by formal tests in Stuttgart, which means that this Bedouin ADMIXTURE cluster is very Basal-rich. Laz, in their estimation for the amount of Basal in Stuttgart, decided to use Bedouin and were so sure that they went so far as to entertain only African admix into bedouin in the calcs.
There is no doubt that Basal Eurasian existed in North Africa. The question we were talking about is how far east they existed in West Asia, which gives us a western limit on the extent of the WHG-ANE-ENA root pop, because we know for sure that even the Westermost European Mesolithics did not encounter Basal in their first migration to Europe.
It is unlikely that Basal has much less Neand than Crown Eurasian, because Ust-Ishim and mesolithic Europeans, who have no Basal, have indistinguishable amount of Neand as later Europeans, who do have Basal.
If Basal Eurasian is old in ME and N.Af, which I think is likely due to Neand Admix in Basal and evidence for a single admix pulse in Basal+Crown Eurasian, then it must mean that the pop of Europe+Siberia occurred northward movement through Central Asia, which adds credibility to the ENA+WHG+ANE clade being rooted in India, and explains why the first bifurcation on the crown Eurasian tree occurred on either side of the Himalayas.
@ n/a
ReplyDelete"I thought I just got done explaining why "basal Eurasian being in ME/ west Asia" doesn't work. Gene flow within the Middle East would homogenize populations on time scales much shorter than 50,000 or 30,000 years. There must have been a strong geographic barrier, and there's nowhere in West Asia "Basal Eurasians" could have remained genetically isolated from other Eurasians for that length of time."
There are two possible scenarios
1)Basal after admix with Neand, retreated into NAf and were isolated from Crown Eurasian; however this is basically the same scenario in terms of implications on the root of Crown Eurasian, because i) the split occurred in the ME, and ii) the same issue with permeable isolation over long periods occurs across the Sinai, though this is arguable.
2) Basal actually admixed with Crown in a continuum in the ME, but the WHG-ANE-ENA expansion occurred far in the east away from locations where the continuum existed. In fact WHG and ANE trekked to Europe + Siberia without exchanging genetic material with continuum pops. This also does not change the situation.
The solution you are suggesting, if I'm understanding you right, seems to be that Basal and Crown split across the Sinai, with Basal have less Nean admix. However this is not supported by the level of Neand admix in pops like Bedouin, and in modern vs mesolithic European genomes.
N/a & RK
ReplyDeletebased on what you chaps are proposing (give or take your own peculiarities), then how would you modify the diagram from lazaridis (http://dienekes.blogspot.com.au/2013/12/europeans-neolithic-farmers-mesolithic.html)
n/a
ReplyDelete" physical anthropology, and Y haplogroup E leave little doubt of a significant North African influence on Epipaleolithic Middle Easterners"
I dont think you're correct here. Firstly, there is the arguement that E diffused from WA to NA.
Secondly, the Neolithic in NA and middle/ NEar East were their own, independent yet linked developments. But if anything, the Mid East influenced NA, and not vice-versa.
(NA= north Africa)
ReplyDeleteMy suggestion to explain how Middle Easterners remained separate from all other non-Africans for tens of thousands of years is the following. The first successful settlement of the Out of Africa migration was not in the Middle East but in Pakistan via coastal migration from East Africa. Admixture with Neanderthal-like archaic humans must have happened there. Later some migrated to the Middle East and settled there and in North Africa. However they could not settle Europe because of water barriers and because of the presence of Neanderthals there. Europe was settled by a later migration of WHG again from the Indian Subcontinent. Geographical barriers then kept the Middle Easterners separate in the Middle East and North Africa until the Neolithic. With this scenario, the true BEA is to be found in the Indian Subcontinent and the Middle Eastern component can better be called AME (Ancient Middle Eastern).
ReplyDelete@ Balaji
ReplyDeleteI am rather confused by what you are saying. If Basal European settled in ME and NAf, and WHG arose from India, the how can BEA be in India? Isn't India the root of Crown Eurasian then?
@ MT
Nothing is modified from the tree. The structure is just
(African (Basal (ENA(WHG, ANE)))))
which is the same as the paper.
The only thing we are trying to do is 1) trying to match the branching to geographical correlates based on modern and ancient autosomal DNA and 2) attempting to match Yhaps to the topology of the branching to infer which ancestral YHap diversified in which ancestral pop, and thus where it diversified. 2) is done with caveats mentioned previously; Y-DNA is subject to high drift and high loss of information; autosomal DNA is subject to as much drift as is needed to create the length of the branches, and thus to diverge the two pops. Autosomal can falsify Y-flows, but Y-phylo cannot disprove autosomal flows. So while we can conclude for various reasons that Siberian autosomal flow to SEA is unlikely from Y-DNA and the reverse movement is somewhat more likely, we cannot prove it, while we can prove that C6 in Europe could not have come westwards from Japan during the HG peopling of Europe as the nearest autosomal connection between both is via their parent population, the (ANE-WHG-ENA) root.
E* is found in some Lebanese and Syrian samples if I rmb correctly, while E* in Africa has turned out to be spurious. E prob has an Asian origin IMO. Also, in pca many Africans, even Yoruba, are pulled strongly to Eurasians; but Middle Easterners are pulled weakly towards Africans, which is exactly the behaviour that ANE exerts on Europeans and Asians: Siberians+Amerindians pulled strongly to Europeans, and N.Euros pulled weakly to Asians, by a pop that roots close to Europeans contributing to both.
In Afrasia Basal can play the same role, rooting close to Eurasians and contributing to both African and Eurasian pops, pulling Afr strongly and ME weakly. I have always suspected hidden Basal ancestry in many Africans for that reason.
The genome is now out, and rumor has it that Ust-Ishim is K-M2335.
ReplyDeletehttp://yfull.com/tree/K-M2335/
That's ancestral to N1c1 y'all.
@Davidski
ReplyDeleteWhere did the rumor come from?
Here...
ReplyDeletehttp://www.anthrogenica.com/showthread.php?3344-DNA-yields-secrets-of-human-pioneer&p=56850#post56850
"On YFull's haplotree, then, this man from 45,000 years ago has some but not all of the SNPs associated with K-M2335, which is essentially pre-NO."
ReplyDeleteEbizur, what is the estimated age of NO, and does this fit?
Judging from the SNPs, could we come up with how long prior to NO he lived, and how long his branch is from K? Was he on the branch that led to X?
Last of all, this implies that there was no Denisova in a branch leading to NO, which strengthens your interpretation of events. N/A's caveats regarding to y - haps vs autosomal still applies though.
Back To what you fellas were saying about the barrier between Basal Eurasian and basal West Eurasian would have been the Nile valley itself. Basal Eurasians , if you're corrct about them being North African , would have this harboured in the wet Sahara. There is plenty of palaeo climactic data about this . Then, when humid phases came to be, they would have again moved east into Nile delts and Arabia
ReplyDelete*west sahara*
ReplyDeleteAnd if Utsy's genome was indeed 'pre-NO', then macrohaplogroup K might indeed have spread throughout eastern and northern Asia as already differentiated sub-groups, and not diversified solely within SEA.
ReplyDeleteThe ratio of positives and negatives for SNPs under "pre NO" suggests that the Ust-Ishim man's line diverged from the rest of K2a quite early.
ReplyDeleteEither way it says very little about its cousin, K2b line, and the available evidences still overwhelmingly point to the proposition that K2b is SE Asian.
There are only 2 SNPs linking him to K2a but still this is the worst possible scenario for Euro-chauvinists and I am amazed by your ability to turn every argument and fact in favor of your "theory".
Does ANYTHING make you abandon your idea? "your theories" are unfalsifiable and therefore garbage.
Too much has been made of the affinity between "Ancient Northern Eurasians" and West Eurasians.
ReplyDeleteUst-Ishim man seem to lend support for my long held view that ANE were not born as West Eurasians but "became" one because their descendents mainly contributed to the gene pool of the West Eurasians.
In terms of the initial branching order I think ANE cluster with East Eurasians or even East Asians. The fact they do not cluster with East Asians more than with SE Asians/Oceanians may be due to the fact they are heavily influenced by the latter in the male line.
Hector,
ReplyDeleteYou can see the branching order here...
https://drive.google.com/file/d/0B9o3EYTdM8lQbXpsNGg2a05oSHM/edit?usp=sharing
East Asians are the eastern non-Africans who contributed to Amerindians.
@Hector
ReplyDeleteYour first point is well taken. If u-i were indeed so basal, then this really doesn't change the situation very much; we're back where we started in some sense.
However there is simply no point in trying to cast aspersions on the autosomal branching order. Unlike y haps, the autosome cannot lie.
The autosomal branching order is generally based on SNP frequencies and it is dependent on the presupposed population model.
ReplyDeleteMost of these theorists are not population geneticists with rigorous scientific training but generally are from other disciplines ostensibly related but actually unrelated like biochemistry.
Until I see extensive studies utilizing things like massive arrays of linkage disequilibrium blocks(which tell a little bit more about the history of each block beyond mere frequency distributions which can be confounded by things like admixture) I will remain unconvinced by most of these "theories".
That is because I actually know some of these researcher personally and have seen how stupid and ignorant they can be.
This sounds like a bluff because you don't know who I am and let's leave it at that.
The part that says ANE contributed to both Native Americans and Europeans seems uncontroversial.
ReplyDeleteBecause of the general absence of ANE component in the older European layer(WHG) theoretically it may be possible to guess the phylogenetic relationship between ANE and WHG by studying ancient WHG samples and comparing them with specimen like Malta boy.
Some may claim to have done that but I laugh at such ideas. I have not seen any satisfactory work in that direction.
Hector, How is hypothesizing a possibly south -central Asian rather than SEA origin for K2 being "Euro-chauvanist". Check your geography and also please check the apparent chip on your shoulder . If you have something insightful to say, then please do.
ReplyDeleteThe presence of a basal branch of K-M526 in Siberia that left no descendants neither proves nor disproves the Southeast Asian theory of K2b. Simply put, the importance of the discovery itself is overrated. The controversy remains.
ReplyDeleteUst-Ishim doesnt show much other than that K2 is older in East asia than C or D is, and that K y-dna is generally associated with R mtdna. its hard to say but if we are going to know the origins of K we need more super ancient y-dna samples including some from Australia.
ReplyDelete"The tree for C now is ((((C1,C6)C5)C3)C*(C2)) (is this correct Ebizur?)"
ReplyDeleteISOGG now has C1a (C1a1 is the old C1 and C1a2 is the old C6), C1b (C1b1 is the old C5 in Gujarat, C1b2 is new and in Bangladesh), C1c (the old C2) and C1d (the old C4). C2 is the old C3.
"Ust-Ishim doesnt show much other than that K2 is older in East asia than C or D is"
Does it really show that? Don't forget, an individual can only belong to one Y-DNA clade. We have no idea what haplogroup other members of his 'tribe' had. With C1a-11043 (old C1 and C6) being found only at either end of Eurasia it is difficult to conclude its presence is other than ancient there.
"A year before that Hammer had R arriving with Neolithic farmers".
Which it probably did as far as most of Europe is concerned.
n/a wrote:
"The tree displayed in the Karafet paper is not based on complete sequences, so you're assuming information about timing that's not actually there. For the Oceanian section of the tree, all Karafet had to go on was a couple of low coverage genomes and markers discovered incidentally while typing a handful of relevant known SNPs"
'Timing' is actually irrelevant here. Karafet et al are simply considering the phylogeny which tells us:
"In sum, our results support the hypothesis of a Southeast Asian/
Oceanian center for the diversification of Oceanian K-haplogroup lineages and underscore the potential importance of Southeast Asia as
a source of genetic variation for Eurasian populations".
Although based on a limited sample size it does not substantially alter the 2005 phylogeny used as the basis of the McDonald maps of 2005:
http://www.scs.illinois.edu/~mcdonald/WorldHaplogroupsMaps.pdf
Even at that stage, before either T or MNOPS had been discovered, it was apparent that K was at the base of both P and NO. And K was primarily a Southeast Asian haplogroup.
@ John SMith Have to agree with Terry. Your comments are non sequitur
ReplyDelete@ Terry T.
I wholly understand what you are saying, but archaeologically/ demographically I am not convinced.
At present, we don't have any evidence for colonization/ population of SEA much before 45-50 kya, and that's going on Australian evidence , as fossil evidence from SEA is lacking or dubious at best for anything before 20 kya.
With this dating in Siberia, then, the colonization of Siberia and SEA could be seen to have been part of a series of chronologically close colonization attempts from a focus in central-southern Asia, radiating Hg K.
@Lathrindor "Simply put, the importance of the discovery itself is overrated".
Sure, we cannot hang our hats contently on just one piece of evidence, but lets recognize the antiquity and success of this work, and the fact that it nevertheless is highly informative.
r.k. wrote,
ReplyDelete"The tree for C now is ((((C1,C6)C5)C3)C*(C2)) (is this correct Ebizur?)"
terryt wrote,
"ISOGG now has C1a (C1a1 is the old C1 and C1a2 is the old C6), C1b (C1b1 is the old C5 in Gujarat, C1b2 is new and in Bangladesh), C1c (the old C2) and C1d (the old C4). C2 is the old C3."
Sorry I'm late to the party today.
Actually, there is a discrepancy between the current ISOGG tree and the current YFull tree in regard to the branching order of the subclades of haplogroup C.
According to ISOGG, the branching order is as follows:
{C2-M217+{{C1a1-M8+C1a2-V20}+{C1b1-M356 + C1b2-M38} + C1c-M347}}
ISOGG has grouped C1b2-M38 (the Wallacean-Melanesian-Polynesian subclade) with C1b1-M356 (the South Asian/Central Asian/Southwest Asian subclade) on the basis of their sharing the F1370 SNP.
On the other hand, according to the current version of the YFull tree, the phylogeny of haplogroup C is as follows:
{C2-M217+{{C1a1-M8 + C1a2-V20} + C1b-M356/F1370 + C1c-M38 + C1d-M347}}
In addition, YFull indicates that the entire {C2-M217 + C1-F3393/Z1426} group that I have just described forms a monophyletic clade, C-Y1786, relative to C*-M130(xC-Y1786).
YFull has F1370 on the same level as M356, marking the South Asian/Central Asian/Southwest Asian subclade of C. On the other hand, ISOGG has F1370 one level higher, marking a clade that subsumes M356 and M38 (the latter of which marks the typical Wallacean/Melanesian/Polynesian subclade of C, although some other type or types of C is/are also common in Wallacea) as C1b -F1370 in contrast to Japanese/European C1a-CTS11043 and Australian C1c-M347.
I presume that ISOGG is the more up-to-date of the two trees in this case, and that the people in charge of the YFull tree simply do not have any data regarding the status of the F1370 SNP in a sample of C-M38 Y-DNA. You would have to ask them to be sure, however.
"The presence of a basal branch of K-M526 in Siberia that left no descendants..."
ReplyDeleteMost of ancient samples left no descendents.
People keep talking about Ust-Ishim being "basal" but given the time frame he lived in how else did you expect him to be?
"...please check the apparent chip on your shoulder..."
I must have seen this cliche at least 10 times in my internet career. The usage of this cliche usually meant that the user was deeply hurt, probably too young to realize how bad it looked and desperately trying to come up with a one liner. Utter fail in terms of the sarcasm-scale.
If I have to ever resort to using a cliche I would choose something shocking and deeply offensive, not something this boring, banal and bland... and most of all just pathetic.
Examining YFull tree should show you that the company has no sample that has M38 or M347. The tree is based on 1100 or so samples. That sounds a lot but it naturally misses many branches from under-represented regions in terms of the customer base.
ReplyDeleteThey probably used the older ISOGG tree as a supplement when they drew the tree.
@ All
ReplyDeleteI'm gonna summarise what we've taken away thus far IMO.
One consistent theme is that Y-haps are relatively unreliable; there are no descendants of U-I's Y in that part of Siberia today. So we are kinda in an epistemological vacuum here, with regard to either 1) there were many more U-Is, and all died out in Siberia, or 2) there were many early radiations from somewhere else into Siberia. Because, while we have many basal K in SEA, we also have one in Siberia, and it is impossible to say if 1) or 2) was the main producer of the picture we have currently. Whatever conclusion we can draw from modern distr is very weak.
At the same time, I think its safe to say that the very clean autosomal splits, along with pop hist, localize the WHG-ENA-ANE clade to somewhere East and South, and that this is really the only piece of evidence we have for localizing K to India or SEA.
This is however counterpoised against evidence for later gene flow into SEA and E.Asia due to dilution of Denisovan, and also the lack of Denisovan admix in U-I. The second flow in SEA and E.Asia produced both ASI/Onge and East Asians, overprinting Oceanoid pops.
The above is again counterpoised against the branch order, suggesting an extremely deep bifurcation betwween WHG-ANE and ENA in Eurasia, and thus suggesting that whatever pop flows into E+SEA arose from a pop that was far closer to Papuan than it is to an undifferentiated remnant in ANE-WHG-ENA, and thus that the Crown Eurasian root pop must have been in a geographical circumstance where it bifurcated early with two separate temperate colonisation events.
I think the balance of evidence is only slightly in favour of the SEA/India origin at this point.
Feel free to add to this picture.
"At present, we don't have any evidence for colonization/ population of SEA much before 45-50 kya, and that's going on Australian evidence , as fossil evidence from SEA is lacking or dubious at best for anything before 20 kya."
ReplyDeleteCheck this out. You may not have seen it:
http://dienekes.blogspot.co.nz/2011/11/42000-year-old-fishermen-from-east.html
From that we can tell humans were fishing in the sea in that region at least 42,000 years ago. And don't forget: Timor is east of Wallace's Line and so humans have to have reached there by boat.
"ISOGG has grouped C1b2-M38 (the Wallacean-Melanesian-Polynesian subclade) with C1b1-M356 (the South Asian/Central Asian/Southwest Asian subclade) on the basis of their sharing the F1370 SNP".
Thanks. That's an alteration since I last checked. That gives just three 'southern' C haplogroups although C1a can hardly be called so.
Thanks terry . As I said , we can only hope renewed vigor into this part f the world brings forth new fossil evidence.
ReplyDeleteAs of now, I am of the opinion, following what I've read by palaeoanthropologists, that an early (100 kya) southern dispersal might well have occurred , but this might have stopped at India for quite some time ...
As R.K. Agreed above, Much of India had a savannah-like environment which resembled Africa , which contrasts with the tropical ranforests which dominate SEA. This might have created a pause in colonization with prerequisite adaptation, and then only supporting relatively small-scale population groups, albeit rapidly disseminating throug SEA and Australia from c. 50kya.
I might be wrong, of course.
@ Mike Thomas
ReplyDeleteQuestion: How do you reconcile this with early dates for the colonization of Australia?
Sundaland at the time was agglomerated into a single landmass, and the savanna/monsoon climate found in inner Burma or East Java was much more widespread in the Sunda Shelf at the time, as the Shelf is basically a depression ringed by mts from each of the islands today.
Could you post your sources for what you read from Paleoanthropologists, esp after such old fossils are found in S.China as Liujiang and etc?
If a pause occurred in India than the Brahmaputra river would be a possible dividing point for the population. Across the river and down the coast you are in tropical SEA. Another group that lived closer to the base of the Himalaya could have headed North through the valleys that lead to the Tibetan plateau.
ReplyDeleter.k. wrote,
ReplyDelete"E* is found in some Lebanese and Syrian samples if I rmb correctly, while E* in Africa has turned out to be spurious. E prob has an Asian origin IMO. Also, in pca many Africans, even Yoruba, are pulled strongly to Eurasians; but Middle Easterners are pulled weakly towards Africans, which is exactly the behaviour that ANE exerts on Europeans and Asians: Siberians+Amerindians pulled strongly to Europeans, and N.Euros pulled weakly to Asians, by a pop that roots close to Europeans contributing to both."
Y-DNA haplogroup E is not the only common element in the haploid gene pools of modern Africans and modern Middle Easterners.
Y-DNA haplogroups B and Y*(xBT-M91) and especially mtDNA L(xM, N) are found regularly as far east as Sindh, and even further east in India in a more segregated fashion (among Siddi communities). In Asia, these haplogroups seem to have a coastal tendency, although I recall that Haber et al. (2012) have reported finding Y-DNA haplogroup B as far into the interior as the Hazaras of Bamyan and Ghor, Afghanistan.
Autosomal analyses of these populations that I have seen have attributed much of their ancestry to some "Southwest Asian" or "Red Sea" component, which I think may be masking significant amounts of recent admixture from Africa. Even accepting the hypothesis of an Asian origin of haplogroup E, I think we must admit that admixture has been a two-way street in more recent ages -- unless you want to argue that (at least some) of these Y-DNA B and Y*(xBT-M91) and mtDNA L(xM, N) lineages are native to southwestern Asia, which would verge on posing a challenge to the very foundation of the Recent OoA theory.
r.k. wrote,
"Ebizur, what is the estimated age of NO, and does this fit?"
Yan et al. (2014) have estimated the TMRCA of N-M231 and O-M175 to be 30,000 [95% CI 27,900 to 32,000] YBP and the TMRCA of NO-M214 and P-M45 (i.e. the TMRCA of K2a and K2b) to be 33,000 [95% CI 30,900 to 35,200] YBP. In response to the finding of K2-M526 Y-DNA in the Ust-Ishim specimen, which has been dated to approximately 45,000 YBP, parasar on Anthrogenica has suggested applying a corrective factor of 1.51 to Yan's TMRCA estimates. This would result in an estimate of 45,300 [42,129 to 48,320] YBP for the TMRCA of N-M231 and O-M175, which would encompass the Ust-Ishim specimen's estimated age.
More securely, it may be said that the TMRCA of N-M231 and O-M175 should be approximately 91% (or anywhere from 79% to ~100%) of the TMRCA of K2a and K2b. Furthermore, the TMRCA of N and O should be approximately 84% (or anywhere from 73% to 96%) of the TMRCA of G and HIJK.
r.k. wrote,
"Was he on the branch that led to X?"
No. The haplogroup X of Magoon et al. (2013) was based on the Y-DNA of HG03742, an individual of Telugu (Dravidian from roughly Andhra Pradesh/Telangana) origin sampled in the UK. This individual shares with all NO-M214 individuals at least three SNPs (M2313, M2339/Z4952, F650/M2346) that the Ust-Ishim specimen lacks. Thus, "haplogroup X" and NO-M214 are closer to each other than either is close to Usty's Y-DNA.
r.k. wrote,
"Judging from the SNPs, could we come up with how long prior to NO he lived, and how long his branch is from K?"
It seems that Usty shares only two SNPs with macro-NO (including "X") that he does not share with the rest of K2. He also has "6 additional mutations that are not observed in the 23 present-day humans to which we compare" according to the authors of the present study. So I would estimate that Usty is about eight SNPs away from the root of K2-M526, with one quarter of those being shared with macro-NO.
BTW, the MRCA of all NO-M214 is about 55 SNPs downstream of K2-M526, with about seven of those 55 SNPs being shared with the "X" group and about two (or possibly up to two more that have not been recovered or have not been genotyped, for a maximum of four) of those 55 SNPs being shared among the "X" group, NO, and Usty.
ReplyDeleteUsty is about 8 SNPs downstream of K2-M526, whereas the MRCA of {X + NO} is about 7 SNPs downstream of K2-M526, so Usty should have lived roughly around (or even slightly after) the time of the split between X and NO. Therefore, I would place the split between X and NO at roughly 45,000 YBP.
I estimate that the TMRCA of {X + NO} should be about 98.8% that of K2-M526 and the TMRCA of {Usty + {X + NO}} should be about 99.3% to 99.7% that of K2-M526. Therefore, the MRCA of K2-M526 may have lived as little as half a millennium prior to Usty.
What is the probability that they may have unearthed and tested the remains of an individual who is perhaps only 18 generations removed from the direct patrilineal ancestor of nearly half of extant humans? My head is spinning...
@ Ebizur
ReplyDeleteThanks for all the great info!
So it does seem that Usty was extremely basal in K, and extremely basal in WHG-ANE-ENA. This solidifies my conviction that the "direct patrilineal ancestor of nearly half of extant humans", aka Y-hap K, diversified in the Crown Eurasian root pop in that timeframe you suggested.
So now we have a very basal sample in ancient Siberia, and a series of very basal samples in modern SEA and India. I guess the thing now is to wait for ancient DNA to see what comes out, as far as the Y-DNA evidence is concerned.
@ Ebizur @ Strandloper
I still have a feeling that the perspective autosomal genetics presents, on the other hand, about the YHap K-Crown Eurasian connection is somewhat more conducive to a southern origin. However the fact that a basal K2a existed in a pop with no Denisovan means that we have to situate it a bit further to the West, with K2a entering ENA pops in S.India and SEA where a trace of Denisovan is already present, or taking part in the overprinting of Oceanoid pops with the new (Dai-Onge) branch.
Its all highly confusing, because Haplogroup M also has its highest basal diversity in Burma, and N and R in South Asia, and thus this Brahmaputra idea has been tossed about multiple times (@ n/a I suppose M, N, R is just old enough for the ~45 Kya timeframe to be informative??). But Denisova in both sides of the Bengal is going to be hard to explain by any scenario that has the bifurcation of Crown Eurasian occurring as far east as the Brahmaputra.
@ Ebizur
There most definitely is recent African admix in the Middle East; however the only pops in Eurasia affected by this far enough to show a strong pull to Africa on the same level as Yoruba show to Eurasia vis-a-vis Khoisan are North African populations.
I think the Eurasian contribution in Africans, on the other hand, is probably greater in quantity, more pervasive, and older such that it is difficult to infer at a glance.
By this method of calibration, TMRCA {N-M231 + O-M175} should be approx. 41,400 YBP and TMRCA {Q-M242 + R-M207} should be approx. 33,250 YBP.
ReplyDelete@Ruki -
ReplyDeleteW.r.t the 100 kya fossil frrom China - the evidence is debatable. Firstly it comes from just a mandible, and the basis on which some scientists argued that it is AMH is that it has a proment mental symphysis (ie a chine) something which archaics apparently lacked. However, this is not entirely true, and some latter arhcaic hominids eg neanderthals did start developing 'chins'. Thus if you look up any good scholar (eg have a look at John Hawkes blog), they;d point this out and argue that it is not yet sufficient to call this a homo sapiens.
W.r.t the data on palaeecology, it pains me, but I cannot remember where exactly,. Must have been the stuff by Petraglia - have a loook at Academie.edu - his papers are all posted there. It might have been this one specifically http://www.palaeodeserts.com/wp-content/uploads/2013/04/Human-dispersal-PDF.pdf
In the same authors GIS paper, he argues that "rapid dispersals along coastline and river valleys would have occurred upon initial expansion out of Africa, but slowed as population expanded demographically into South Asia and Sunda Shelf".
:)
Ah got it; from
ReplyDeletehttp://www.palaeodeserts.com/wp-content/uploads/2013/04/Human-dispersal-PDF.pdf
Pg 11:
"India is dominated by savannah and Sahel type vegetation,
as well as vegetation zones whose differences from these are not
radical, and similar food resources are attested to by disjunct taxa.
The region just east of India, however, features significantly different
environments that may have posed new challenges to dispersing
humans (Fig. 4). The northeast Indian region of Assam and
adjacent areas are characterized by high hills with dense tropical
forests, and seasonally flooded and swampy river valleys. Tropical
rainforests generally support low human population densities (cf.
Bailey and Headland, 1991) and necessitate specific rainforest adaptations
(e.g., techniques for hunting hard to catch arboreal game
such as monkey, see Piper et al., 2008; Perera et al., 2011). Swamps
may have supported lower populations due to higher disease loads
(cf. Sattenspiel, 2000), and if not true barriers would nonetheless
have reduced local populations. The impression of a major barrier
between the regions of India and Indochina is reinforced by biogeographical
and genetic studies suggesting that a wide range of
animals are differentiated or speciated across this region."
@ MT
ReplyDeleteThanks for the info. I have been a regular reader of John hawks, but must have missed that one.
It pains me to say this, but I have always been somewhat dubious of Chinese paleoanth, esp when comes to 'ancestors' of any kind.
Many ppl I have met in China still do not accept OOA. Oh well :)
@ MT @ Ebizur
I suppose this is adds to the center of gravity being west of SEA and towards India.
I can think of an alternate scenario of the divergence occurring exactly at the forest-savanna border, with ENA having to pass through the forest in SEA and South India (S.Indian forest never died, was a refugium in Ice Age); but this is just pure speculation.
Oh, @ Hector
ReplyDeleteWhat is 'banal and boring' is your 'champagne liberal' unwarranted, self-righteous huffing and puffing ad hominens on well-meaning contributors who clearly have absolutely no racial agenda, as anyone reading here the comments can see.
Put that on your tofu burger....
@ MT
ReplyDeleteA 'champagne liberal' would end up giving him too much credit.
Nothing against Koreans, but Hector is a Korean ultranationalist, which is much worse. No need for noble motives here.
As we are all clearly keeping an open mind , humans might have been in SEA for , say 80K, not a mere 40K. But the evidence is lacking .
ReplyDeleteI hope more , clear and unbiased evidence comes forth .
I should add that the phylogenetic position of the following types of K-M9 Y-DNA vis-à-vis the Ust-Ishim specimen's Y-DNA (and probably also vis-à-vis X-F650/pre-NO, though I do not have access to the full paper by Karafet et al. 2014) is unclear at present:
ReplyDeleteK2c-P261
15/641 = 2.3% Bali
K2d-P402/P403
2/61 = 3.3% Java
K2*-M526(xNO-M214, K2b-P331, K2c-P261, K2d-P402/403)
5/38 = 13.2% Batak Toba (Sumatra)
1/177 = 0.6% Sulawesi
K-M147
2/88 = 2.3% Pakistan + India (Underhill et al. 2000)
K-M147 has been described by Underhill et al. (2000) as being K-M9+/M147+/M175-/M70-/M11-/M20-/M22-/M61-/M46-/M128-/M4-/M5-/M106-/M45-/M74-.
In present terms, it would be K-M9(xL-M11/M20/M61, T1a-M70, M1-M4/M5/M106, N1c1-M46, N1c2a-M128, O-M175, P1-M45/M74). Karafet et al. (2014) have stated, "The haplogroup M147 has not been assigned to the M526 haplogroup" (in the sense that its status in regard to the M526 SNP is unknown).
I was wrong about it proving that the Earliest East Asians have K as it was really rather far from the East Asia population center. If you look at this map http://upload.wikimedia.org/wikipedia/commons/0/07/Irtyshrivermap.jpg it is clear that Ust-Ishim comes from a population associated with the western side of the Irtysh river. If Ust-Ishim was only 500 years about K2 was born, he probably descended from some early K+R mtdna population associated with the Volga and Ural rivers which included a man with K2 who is the direct ancestor of most modern human males currently (although this may change soon as Africa has the highest current growth rate and very little K2).
ReplyDeletehttp://genetics.med.harvard.edu/reich/Reich_Lab/Welcome_files/2014_Fu_Nature_UstIshim.pdf
ReplyDeleteis the fulltext of the article
"Thanks for all the great info!"
ReplyDeleteAs usual Ebizur has come up with the goods.
"an early (100 kya) southern dispersal might well have occurred , but this might have stopped at India for quite some time ... As R.K. Agreed above, Much of India had a savannah-like environment which resembled Africa , which contrasts with the tropical ranforests which dominate SEA".
I'm sure you are correct there.
"If a pause occurred in India than the Brahmaputra river would be a possible dividing point for the population. Across the river and down the coast you are in tropical SEA. Another group that lived closer to the base of the Himalaya could have headed North through the valleys that lead to the Tibetan plateau".
Once more I think that statement is correct.
@Terryt
ReplyDeleteI used to think that too. But the biggest problem for this theory is this: how do you reconcile denisova being in ENA on either side of the Brahmaputra, while ANE-WHG West of North India has no Denisova, while the earliest pops deriving from this root that we have unearthed also have no Denisova?
A last piece off evidence, from kunstamera.ru posted by parasar:
ReplyDelete"...Certain cranial features, including very narrow braincase, low and narrow face, marked prognathism (anterior protrusion of the midface), and very wide nose, are typical of tropical populations. The trait combination links the cranium with those of Papuans and Melanesians.
Certain other Upper Paleolithic crania from Europe, too, display “tropical†features. Bodily proportions of Early Upper Paleolithic people reveal southern characteristics as well. This also concerns the arm proportions of the Markina Gora individual, whose forearm was relatively long compared to the shoulder.
The meaning of those facts is yet unclear. Modern geographic human groups (so called “racesâ€) had not completely formed by the Upper Paleolithic, and some of their characteristics may have incidentally appeared in various parts of the world."
"M.M. Gerasimov, like G.F. Debetz, believed that people of such appearance or their direct ancestors had actually migrated to Europe in the early Upper Paleolithic from areas lying far south. Therefore he endowed the reconstructed individual with tropical characteristics including curly hair. The future will hopefully show whether or not this “artistic liberty†of the scientist and sculptor was warranted."
Referring to the Kostenki individual from the Don, 32000 BP.
"Much of India had a savannah-like environment which resembled Africa"
ReplyDeleteMan o Man. Please people look at the definition of Savannah. Don't make theories to support your ethnic agendas.