Wednesday, December 17, 2014

Musing On The Y-DNA Of The First Farmers

The predominant Y-DNA haplogroup of the first farmers of Europe was Y-DNA haplogroup G2, together with some Y-DNA haplogroup I, the predominant Y-DNA haplogroup of European hunter-gathers (probably integrated into the farmer community from hunter-gatherers), and I believe, a bit of Y-DNA halogroup T.  This is consistent across the LBK farmers who were the first farmers of Central and Eastern Europe, the Cardial Pottery farmers who were the first farmers of Southern Europe, and the early Neolithic megalithic farmers of Western Europe.

Ancient DNA evidence strongly suggests that these first farmers replaced or demographically overwhelmed and assimilated in small percentages, the pre-existing hunter-gather men of Europe.

Relict populations, rich in Y-DNA haplogroup G, all of which likely date to this first wave of farmers, are found in several populations in the modern Caucasus mountains and Sardinia.  Y-DNA haplogroup G is also found at elevated levels in Corsica, Gascony (in Southern France), and Tuscany (which was home to the Etuscans of Italy before they suffered ethnocide within the Roman Empire).

Y-DNA haplogroup G has its greatest diversity in the Levant, which was part of the Fertile Crescent where farming was invented, suggesting that it originated there.  But, it is now most common in the Caucasus and the vicinity to the north of these mountains, where it is far less diverse.

Ancient DNA informs us that Y-DNA haplogroups R1a and R1b were almost entirely absent from the first farmers of Europe, appearing only in the Copper Age when they became the predominant Y-DNA haplogroups of Europe, diluting Y-DNA haplogroup G (and the small percentages of Y-DNA haplogroups I and T found in association with it) to low percentage frequencies in most of Europe.

Y-DNA R1b-V88, found mostly in Chadic people in Africa, arrived around 5700 BCE in the vicinity of Lake Chad, show signs of a linguistic connection to and wife taking from Cushitic people, probably in Southern Sudan.  But, Y-DNA R1b-V88 was not shared by any other African population to any great extent.

Y-DNA haplogroups J1 and J2 were also associated with early Neolithic peoples, although just how this played out is a bit unclear.  Today, J1 is confined largely to Semitic peoples who were traditionally herders in SW Asia.  Meanwhile J2 is associated with the highlands of the Fertile Crescent - Anatolia, the Caucasus, Armenia, and the Zargos Mountains, none of whom have ever been linguistically Semitic in attested history. There is a cline of relative J2 to J1 proportions from these highlands where J2 is more common, to the Bedouins of the Arabian desert, where J1 is predominant, but in most of SW Asia, there is a mix of both types in any given population.  Y-DNA haplogroup J is rare in North Africa and East Africa (and almost absent elsewhere in Africa).  Where it is found, Y-DNA J1 predominates.  Y-DNA J1 in East Africa is probably attributable mostly to Ethio-Semitic conquest ca. 1000 BCE, and in North Africa is probably mostly attributable to Islamic expansion and intercourse after 630 CE.  The genetic diversity of Y-DNA J is similar throughout Europe, West Asia and SW Asia, leaving a precise point of origin unclear.

The nearly complete absence of Y-DNA haplogroup J in the ancient DNA of the first farmers of Europe suggest that like Y-DNA R1a and R1b, that its presence outside SW Asia is due to a later migration wave.  It is also the case that outside SW Europe, Y-DNA J is predominantly Y-DNA J2, Y-DNA J1 in Europe is largely confined to Jews who migrated there in the historic era, and to areas where there was Muslim rule in the historic era.  The expansion of Y-DNA J2 into Europe may have been as a minor component of the initially Y-DNA R1a dominated expansion of the Indo-Europeans into Europe in second or subsequent waves of farmer and herder migration into Europe.

Europe was not the only receiver of the Fertile Crescent Neolithic package of farming, herding and other technologies and cultural traits.  The Fertile Crescent Neolithic package also spread to the Indus River Valley via diffusion at roughly the same rates seen in Europe across Iran, and to Egypt and beyond in North Africa.

Beyond these areas, the Fertile Crescent Neolithic package hit ecological and geographic barriers.  Fertile Crescent crops did not thrive in the climate of the African Sahel or Sub-Saharan Africa, and domesticated Fertile Crescent animals could not survive the tropical diseases of tropical Africa.  Fertile Crescent crops likewise were not suitable for the climate of tropical, monsoon driven Southern India.  The European and Central Asian steppe could support the herding of Fertile Crescent domesticated animals, but was too dry for the farm crops in the Fertile Crescent Neolithic package of the first farmers to thrive without inventing arid region optimized irrigation technologies (which eventually were invented).

The secondary centers of the Neolithic revolution that ultimately produced the Harappan civilization of the Indus River Valley, and the secondary center in Egyptian and North Africa, however, follow a very different demographic pattern than the demographic pattern observed in Europe.

In the Indus River Valley, prior to the arrival of Indo-Aryan invaders around 2000 BCE, Y-DNA haplogroups L and R2, as well as autochronous South Asian Y-DNA haplogroups like Y-DNA haplogroup H were common, while Y-DNA haplogroups R1, J2 and G that are predominant in early European and Near Eastern farmers were largely absent until the Indo-Aryans arrived in the wake of the collapse of Harappan society in the face of a severe climate event.  Y-DNA T is found in India, but largely in Dravidian areas of Southeastern India that did not adopt agriculture until around 2500 BCE using mostly African Sahel crops.  Y-DNA halpogroup T probably arrived in India at that time, and is otherwise largely absent from South Asia.

We don't know if Y-DNA L and R2 arrived in the Indus River Valley at the same time, or if they represent separate waves of migration to the region.  We do know that Y-DNA L is the sister clade to Y-DNA T and that the two clades may have geographically close origins to each other.  And, we do know that there are strong indications that the split of Y-DNA R into subclades R1 and R2 and possibly also into R1a and R1b, took place in Iran, which has maximal Y-DNA diversity today.  All Y-DNA R appears to have remote origins in the early Upper Paleolithic era in Southeast Asia and a very basal form of Y-DNA R was found in the ancient DNA of Ma'alta man from ca. 24,000 years ago in the Altai Mountain region at the far Southeast of the Eurasian steppe.

It is reasonable to hypothesize that the first farmers were Europe were from the Levant at the west end of the Fertile Crescent, while the first farmers of the Indus River Valley and Iran were from ethnically and genetically distinct populations probably originating from Mesopotamia and the Zargos Mountains at the east end of the Fertile Crescent (an area that in the Copper Age had thriving maritime trade with the Indus River Valley).

We also know that the Fertile Crescent Neolithic package involved several independent domestication events at different centers of population within the Fertile Crescent, that were then assembled into a comprehensive package, probably through trade and exchanges of small numbers of farming experts.  It was not invented by a single anthropological culture and the Fertile Crescent was rarely under the control of a single state or ruler in historic times.  So, it makes sense that different populations that developed particular components of the Fertile Crescent Neolithic package might have different population genetics.

In Egypt, which received the Neolithic revolution from the Fertile Crescent at about the same time as the Indus River Valley and the Balkans in Europe, there are traces of Y-DNA haplogroups G and T, but there is far more Y-DNA haplogroup T than there is Y-DNA haplogroup G.  But, there is very little Y-DNA haplogroup R1 in Egypt that is not traceable to the historic era. And the Y-DNA haplogroup E clades, typical of linguistically Afro-Asiatic areas that probably pre-date the Neolithic revolution are very common.  Thus, in Egypt, rather than the population replacement we see in Europe, we see a much milder demographic impact.  (King Tut was probably Y-DNA R1b, but that was likely attributable to the historic area Semitic Hyskos invasion and establishment of their Egyptian dynasties, or to Greek contacts with Egypt, not to the deep population genetic history of Egypt.)

This isn't to say that the Neolithic revolution didn't bring about immense demographic change in Egypt and North Africa.  Population densities increased by as much as a hundred-fold in a matter of a few centuries.  Any population that did not participate in the demographic population expansion that North Africa experienced as a result of the Neolithic revolution would be almost invisible in modern North African population genetics.

But, unlike Europe, the Neolithic Revolution in Egypt and North Africa was not largely a story of pure male population replacement by outside populations.  There is good evidence from the distribution and phylogeny of Y-DNA haplogroup E's subclades that it originated in the vicinity of Ethiopia, not in SW Asia as a back migrating clade.

Perhaps the fishing economy, and abundance of the hunting and gathering in the Nile Valley gave Egypt staying power in the face of Levantine Neolithic migrants in a way that was not true in Europe.  And, Egypt was the gateway through which all Neolithic expansion in Africa was filtered.

It is also possible, for example, that Y-DNA G and Y-DNA T may have represented two distinct populations in the Levant.  Perhaps Y-DNA G was made up of farmers and Y-DNA T was made up of herders.  In Europe, which was well suited to farming, Y-DNA G dominated, while Y-DNA T left a thin shadow of its range with much less demographic impact.

In contrast, in Egypt and beyond in North Africa and East Africa, herding was a much more significant component of the Neolithic revolution as there wasn't nearly as much arable land available to farm upon due to the narrowness of the Nile River Valley.  So, the impact of the Y-DNA T herder migrants from the Fertile Crescent may have been comparatively great and the impact of the Y-DNA G migrants from the Fertile Crescent may have been comparatively small.

Furthermore, historical evidence seems to suggest that it is much easier for people to transition from a terrestrial hunter-gather mode of subsistence to a herder mode of subsistence, than it is to transition from a terrestrial hunter-gather mode of subsistence to farming.  So, cultural transmission of the Neolithic revolution may have been easier in North Africa where that mostly involved teaching people how to be herders, than it was in Europe where that involved teaching people how to be farmers.  Hence, there may have been less economic pressure for the people who were the source of the Neolithic revolution to replace the existing population of North Africa than there was to replace the existing population in Europe (and probably also in Iran and Indus River Valley).

On the other hand, that specific scenario doesn't square very well with the fact that the Y-DNA T clades found in Southern Arabia are much younger than those in the Levant, or the fact that in SW Asia, low land herding is largely conducted by men who are Y-DNA J1 rather than Y-DNA T.

In general, the timing and circumstances by which Y-DNA J entered into its large demographic role in SW Asia, especially Y-DNA J1, is not clear.

UPDATED December 18, 2014 to correct numerous spelling, punctuation, grammar, and usage issues, as well as a few inadvertent errors, and to better spell out some incomplete thoughts.  Every once and a while when I'm in a really pinch for time, I come up with a real stinker in terms of formal writing errors like these.  I've also looked at Maju's comment and done a bit a research of Y-DNA J2 since this post, but I am not ready to see if I can add some substance to the discussion of Y-DNA J in this post.


Maju said...

J2 is also very important in Italy and the Eastern Balcans, while in South Asia it seems particularly associated to Neolithic flows from the NW. So to me J2 seems very much Neolithic even if it has not been yet sampled in ancient remains.

J1 on the other hand seems to me strongly associated to Afroasiatic peoples but only secondarily so. I.e. the first expansion of J1 (particularly to NE Africa, where it is very diverse) was probably Paleolithic and only part of it can be related to Afroasiatic expansion as such, along with some E1b.

Ebizur said...

Andrew wrote,

"We don't know if Y-DNA L and R2 arrived in the Indus River Valley at the same time, or if they represent separate waves of migration to the region. We do know that Y-DNA L is the sister clade to Y-DNA T and that the two clades may have geographically close origins to each other. And, we do know that there are strong indications that the split of Y-DNA R into subclades R1 and R2 and possibly also into R1a and R1b, took place in Iran, which has maximal Y-DNA diversity today. All Y-DNA R appears to have remote origins in the early Upper Paleolithic era in Southeast Asia and a very basal form of Y-DNA R was found in the ancient DNA of Ma'alta man from ca. 24,000 years ago in the Altai Mountain region at the far Southeast of the Eurasian steppe."

Y-DNA haplogroup L is no more closely related to T than R is related to O. In fact, Hallast et al. have estimated the TMRCA of L + T to be the same as the TMRCA of K; the phylogeny shows that the TMRCA of L and T must be somewhat less than the TMRCA of K, but it is apparently only slightly so. Members of L and T might have remained in close contact with each other for a long time after their MRCA, or they might have parted ways almost immediately and never seen each other again for tens of thousands of years.

As for haplogroup R, the TMRCA of R1 + R2 should be approximately 26,600 years. The Mal'ta specimen, who lived in the vicinity of Lake Baikal approximately 24,000 YBP, belonged to R(xR1'2); i.e. his lineage is more closely related to R1 and R2 than any of R1, R2, or Mal'ta is related to Q, but R1 and R2 are more closely related to each other than either is related to the lineage of the Mal'ta specimen. So, some other derivatives of R did exist at least in what is now Irkutsk Oblast (just north of the middle of Mongolia) approximately 24,000 years ago, but it appears that only the R1 and R2 derivatives of R1'2 have left any readily detectable number of direct patrilineal descendants in present-day human populations. Note that the genealogical split between R1 and R2 appears to predate the time of deposition of the Mal'ta specimen (but not by much).

On the other hand, if we do not adjust the TMRCA estimates to accord with the published estimate of the time of deposition of the Ust'-Ishim specimen and accept the mutation rate of Xue et al. (2009), then the TMRCA of R1 and R2 should be only about 19,300 years. However, in that case, the times of deposition of both the Ust'-Ishim specimen and the Mal'ta specimen should be adjusted downward to accord with the genetic evidence, so the TMRCA of R1 and R2 should predate the time of deposition of the Mal'ta specimen either way.

Contemporaries of the MRCA of R1 and R2 should have included early forms of T1a-M70, O1a, O2a, O2b, O3-F742 (≈ O3*-M122(xO3a-M324)), and O3a-M324. The MRCA of O3a1-L127 and O3a2-CTS8236 should have lived around the same time as the MRCA of R1 and R2. All of extant L and all of extant N may descend from ancestors who postdate the MRCA of R1 and R2; however, the MRCA of extant N may have been approximately a contemporary of the MRCA of R1 and R2 if one considers the recent evidence of a basal derivative of N (found in Iron Age Hungary and in a few modern Serbs and Croats). The deep age of diversity within haplogroup O-M175 is really salient and begs some explanation.

Maju said...

Quite in agreement with what Ebizur says except in using STR-based molecular clock guesstimates, which are totally unreliable. If we follow Underhill 2012 (and even more if we recalibrate archaeologically, see: & ), as well as the few other "full chromosome" age estimates, R1 must be of c. 48 Ka BP, i.e. it branched out in the early Upper Paleolithic, coincident with the colonization of most of West Eurasia by H. sapiens (from Tropical Asia). That means that P and K must be older, and my hunch is that K spread soon after the Toba catastrophe, largely carrying with them mtDNA R (or vice-versa: no need to assume patrilineal dominance in this case).

Surely NO and O expanded in the East in similar chronologies as P and R did in the South-West. The root of IJ, I and J also seems of similar "early UP" age.

andrew said...

I have deliberately refrained from trying to date when splits happened in this post, focusing instead on our best available information about where splits happened.

It is fair to say with some confidence, however, that Y-DNA R1a, R1b and R2 all were present in or adjacent to Iran at the dawn of the Holocene ca. 10,000 years ago.

It is also reasonable, looking at the entire universe of ancient DNA estimates, to believe that on average, particular human communities were more homogeneous in uniparental lineages and had less uniparental diversity in the more distant past, than in the more recent past, particularly in the case of Y-DNA. We start from a "well sorted" situation where most communities had just one or two clusters of closely related Y-DNA haplogroups in the early Holocene and end up in one by the Iron Age where there are many clusters of closely related Y-DNA haplogroups in each community - a result that is partially due to founder effects with the intense demographic growth associated with the Neolithic revolution giving the initially small and genetically homogeneous communities that were boosted by it an outsized impact on the gene pool, while marginalizing the impact on the gene pool of other diverse communities that did not participate in this rapid period of demographic expansion. (In other word, the general trend may break down once you got back to the Upper Paleolithic in areas that didn't experience an LGM bottleneck.)

Maju said...

"It is fair to say with some confidence, however, that Y-DNA R1a, R1b and R2 all were present in or adjacent to Iran at the dawn of the Holocene ca. 10,000 years ago".

I must disagree: it does seem the case of R1a, but R1b must have spread much earlier (so each of its branches should be treated separately re. recent secondary expansions or re-expansions) and R2 was surely already in South Asia. With due caution in this last case because we lack data for R2 but with some confidence because R and R1 seem to stem from NW India or Pakistan.

Re. homogeneity of ancient populations: I think it's hard to judge because we have almost no ancient Y-DNA, especially not outside Europe/Siberia/China. But what we see in these areas, for example in China, is that there was a great deal of diversity in the wider region, even if some populations do seem to fall locally to your idealization of certain homogeneity. It seems to be the case that one population was Q and another just a few hundred kilometers away was N or a mix of several lineages. In Europe we see more dominance of I2 (Epipaleolithic) but there are also instances of other lineages like C1 or Q1, and this variability is surely not exhaustive.

Ebizur said...

I have not made any reference to TMRCA estimates based on Y-STRs:

"TMRCA and its standard deviation was estimated for clades within the PHYLIP outfile using the rho statistic (Forster et al. 1996; Saillard et al. 2000) implemented in a Perl script. A scaled rate of one mutation per 268.5 years was used, based on 1.0 x 10^-9 mutations/nucleotide/year (Xue et al. 2009) and the number of nucleotides in our regions of interest (3,724,156). We assumed a generation time of 30 years. In addition, to capture the uncertainty in the published mutation rate we calculated TMRCA based on the bounds of its
95% confidence interval: 3.0 x 10^-10 – 2.5 x 10^-9 mutations/nucleotide/year (Xue et al. 2009)." (Hallast et al.)

"Fig. S1. Phylogenetic tree of human Y chromosome. The tree is constructed for 78 samples sequenced in this study, together with three published East-Asian genomes (YH, SJK, GMIAK1) and a chimpanzee genome (Pan), which are labeled with ‘*’. Except for YCH145 (Spanish), SJK and GMIAK1 (both Korean), all the human samples are Chinese. The branch lengths (horizontal lines) are proportional to the number of SNPs on the branch, and the SNP numbers are labeled under the branches). The SNPs labeled on the horizontal lines are only representative. The SNPs labeled in green represent newly recognized clades in this study. The estimated coalescence time (in years) for the nodes are calculated only from good-quality (> 6× coverage) human sequences (in bold italic) by BEAST with relaxed clock (see SI Methods), and the numbers in brackets are for 95% confidence intervals (ignoring uncertainty in mutation
rate)." (Yan et al. 2014)

"To test the assumption of molecular clock for the tree, we used PAML package v4.4 with the GTR model. The null hypothesis of a molecular clock cannot be rejected (P.0.05) by comparison between the models. We used the GTR model of nucleotide substitution determined with MrModeltest 2.3 with a strict clock. The single nucleotide substitution rate was set as 1 x 10^9/nucleotide/year. The effective sample size of the coalescent prior was above 900. A relaxed clock was also employed for comparison and the results were similar." (Yan et al. 2014)

All my TMRCA estimates are based on Y-SNPs.

The expansion of NO is not in the same time frame as the expansion of P and the expansion of O is not in the same time frame as the expansion of (extant) R. If the expansion of R1 (to be precise, the genealogical separation of the R1a lineage from the R1b lineage) has occurred approximately 48,000 YBP as you suggest, then the expansion of O (i.e. the genealogical separation of the O1'2 lineage from the O3 lineage) has occurred approximately 78,000 YBP, and the expansion of NO (i.e. the genealogical separation of the N lineage from the O lineage) has occurred approximately 91,000 YBP. There is very little evidence (and ambiguous at that) of AMHs in eastern Eurasia around 90,000 years ago. Needless to say, this would require at least Y*(xBT), i.e. former "haplogroup A," to represent direct patrilineal descendants of various sorts of pre-AMH (unless you want to argue that each of the Y*(xBT) lineages has survived in a polymorphic population alongside one or another subclade of BT ever since the dawn of AMHs).

Maju said...

Alright. Still, which is the calibration for the age Pan-Homo split. Typically the literature claims 7-5 Ma, when in fact it is 10-13 Ma, almost exactly double: what implies that, even if everything else is correct, the results are almost half the realistic ages.

Also, just from the oven: (mutation rates can vary wildly among clades).

"The expansion of NO is not in the same time frame as the expansion of P and the expansion of O is not in the same time frame as the expansion of (extant) R".

Should be about the same. Another thing is that some O3 subclades may have expanded more recently but overall O should be of near the same age as R, Q, I, J, etc.

"There is very little evidence (and ambiguous at that) of AMHs in eastern Eurasia around 90,000 years ago."

Actually the evidence is overwhelming by no:

The evidence for India in the same date range is also unmistakable:

Furthermore: the evidence for presence of H. sapiens in Arabia and Palestine in the Abbassia Pluvial is very strong and logic almost demands that the "pump effect" of this climatic episode would have pushed the people outwards (some of them in eastward direction) when dry conditions were restored, i.e. c. 100-90 Ka BP.

"Needless to say, this would require at least Y*(xBT), i.e. former "haplogroup A," to represent direct patrilineal descendants of various sorts of pre-AMH"

I do think that at least the two oldest branches of the modern human Y-DNA tree are not H. sapiens proper, but hard to decide because there's not enough paleoanthropological data from Central-West Africa, where those lineages seem to have originated. In any case those would be close relatives of H. sapiens, quite closer than Neanderthals, surely H. rhodesiensis.

Similarly there's an X-chromosome lineage that almost certainly stems from Neanderthals. There was never a strictly pure H. sapiens "race": they punctualy mixed with close relatives, even if some of those are considered different species (with the reproductive complications implied). This would hardly be the case between earliest H. sapiens and its immediate "ancestor" H. rhodesiensis: the speciation process was then still in its infancy.

Maju said...

Also there's zero archaeological evidence suggesting later migration of H. sapiens from Africa to Asia (other than minor admixture in the border areas, mostly in the Holocene).

andrew said...

The evidence of H. Sapiens being the species in Asia ca. 100kya is not very solid at all.

Maju said...

It is VERY SOLID. Otherwise how do you explain Aterian in India or chins and other modern traits in China?

I assume that you accept H. sapiens presence in West Asia c. 125 Ka BP, right? That your objection is only East of Iran, right?

Maju said...

What is clear is that the so-called "molecular clock" guesstimates are not C14 nor anything of even remotely similar scientific certainty (and even C14's calibration is fine tuned now and then, causing dates to change significantly, although not too radically, typically towards further into the real chronology past).

There's no one MC modeling method but many, there's too many assumptions in the model that crash with reality: from mitochondrial branches with very unequal lengths to totally wrong calibrations with paleontological or archaeological facts, which are carried on only because of scholastic inertia. Beliefs such as that H. sapiens only crossed into Eurasia (emphasis in "Eur-") c. 50 Ka (or stretched back to maybe 60-70 Ka) are soooo 20th century! Same for beliefs such as the Pan-Homo split being of c. 5 Ma BP (sometimes stretched back to 6-7 Ma).

With all those problems we just cannot take, as so many do, the MC guesstimates as "rocket science" - in fact they are rather like pseudoscience, and the faith in MC guesstimates is not more scientific than the belief in god or the Loch Ness monster.

With all those fundamental flaws in the MCH, we can only say that, at least until a new refined molecular clock model is consolidated, the data from such models cannot be taken seriously.

Even when the STR method is replaced by full sequence or long chanin SNP method (no doubt much better), there are still other serious issues, notably calibration. Calibration is key: C14 had to be demonstrated to work once and again on remains of known age such as Egyptian mummies before it could be accepted. Instead MCH only seems to need enough theoretical argumentation behind, what makes me suspect that we are falling down from an age of lights to an age of darkness in scientific thought. Hardcore evidence does not anymore seem to matter but the parroted word of "scientists" is taken as that of priests was in the past, even when they have zero evidence.

The problem is the same now and in the Dark Ages: scholasticism!


andrew said...

My objection is East of South Asia before 75 kya. The Chinese finds are not credibly dated or identified as H. Sapiens.

Maju said...

If they arrived to South Asia c. 95 Ka BP, and they did, what is the problem with immediate arrival to the southern areas of East Asia? From the viewpoint of geography there's no barrier at all (just a hilly buffer). From the viewpoint of population genetics it is a must, as the Eurasian lineages do share a lot of diversity between the two regions from the very roots (Y-DNA CF, mtDNA M), even if India seems to be somewhat more central. Myanmar and Yunnan also seem very central to the spread of H. sapiens in East Asia, retaining very high diversity, as does NE India (technically SEA). Some lineages expanded directly from SE Asia (mtDNA N, Y-DNA D and K, probably also C): all that implies that Southern and SE Asian colonization was of about the same age, although the core was in South Asia initially (logically) and that there was a re-expansion from SE Asia with Y-DNA K and mtDNA R (in essence).

From the archaeological viewpoint it must be added the c. 81 Ka BP dates for the colonization of Australia, which also seems very solid, although with a large CI of 21 Ka up/down. The c. 60 Ka BP are in any case unquestionable. See:

Personally I think that the early colonization of Australia is of around the Toba episode (v. and generally related to the expansion of mtDNA N from SE Asia in that period.

But, regardless, the East Asian findings are very solidly H. sapiens. You can stand in denial but, to begin with, no other species ever evolved a chin. Chin = H. sapiens - there is no doubt (and it's not me who says that, it's repeated research by experts with very solid conclusions).

The dates are also very solid.

capra internetensis said...

@ Andrew

A little off topic - Y hgs roam pretty freely don't they? But in the next couple of years I expect this will all be far more tightly constrained by whole chromosome sequencing and more extensive Y-DNA surveying.


I don't think the Dark Ages have arrived quite yet. The molecular clock hypothesis has been criticized from its inception, even if sometimes researchers who ought to know better put too much faith in it (60 000 year old Austroasiatic tribes....) There have been many attempts to calibrate the Y-DNA clock - several in the past couple of years - but it isn't easy.

The best possible evidence for calibration is aDNA with absolute dates. We have only a few data points there, but (unless the results are completely wrong) they do rule out R1 dating to the initial Upper Paleolithic. It is rather C, D, E, and F and their most basal clades (K, D1a-c, C1a-c, pre-C2, etc) which date to that period. We have nothing older than Ust'-Ishim man (with Y-DNA K2a* at ~45 kya), but projecting the Ust'-Ishim-calibrated rate backward the separation of CDEF into C, D, E, and F should date to about 75-70 kya (possibly associated with Toba) and of BCDEF into B and CDEF to 110-95 kya (perhaps coinciding with AMHs in Asia being separated from those in Africa). Even these wide ranges may be wrong, of course, but wildly older or younger dates are not very likely.

There is certainly considerable rate heterogeneity within and between lineages, with various possible causes which have not been pinned down. (So you can't necessarily tell for sure whether nodes in different lineages are contemporary or not by SNP counting without aDNA calibration.)

The Y-DNA of Middle Paleolithic humans in East Asia may simply have gone extinct or become vanishingly rare, just as that of Neanderthals and Denisovans has. (Though very little C/D/DE/F* has been subject to whole chromosome sequencing, so the possibility of rare very old Y-hgs remains open, even within what has already been surveyed, which is only a tiny fraction of the male population.)

I guess that damn real life thing has interfered with your blogging, Maju? I always find your perspective very informative, even if I don't agree with you.

Maju said...

I agree with the first sentence: "molecular clock" is (or rather should be) a developing field plagued with question marks everywhere.

I disagree with the rest: lack of evidence is not evidence of lack, unless (if at all) the evidence is overwhelming - and even then there's still room for doubt. There's no direct evidence yet for R1 in the dates you say but there's also HUGE BLANKS in our direct knowledge of Pleistocene and even Holocene Y-DNA. What we do know is that R* did exist in Siberia before the LGM so in general terms R must be older.

Per the full sequence estimates (see: R1b is almost exactly half the age of P1, with R1 being like 70% of the age of P1 (R1a is quite younger than R1b by all accounts). The calibration points to be used here would be then:

→ P1 is older than UP (so at least 60 Ka BP) because we do see R1 associated with post-Neanderthal "Western" UP.
→ R1 is older than the 24 Ka date of Mal'ta. I'd say it's from the very roots of "Western" UP (i.e. 50 Ka BP or older) because it is centered in South Asia and not West or Central/North Asia and there's no archaeological evidence or even indications of secondary migrations from South Asia to Western Eurasia after the one that gave rise to the UP and triggered Neanderthal extinction.

Basically then P1 and R1 must have evolved in the 74 Ka BP and 50 Ka BP bracket (i.e. between Toba and the initial UP).

My own estimates using the work of T.D. Robb on the 1000GP data, but recalibrating for age(CF)=100 Ka BP (supported by the updated archaeological evidence), are that: age(P)=68 Ka BP, age(R1)=48 Ka BP and age(R1b)=34 Ka BP (just in time for the Gravettian second wave of West and North Eurasian colonization, which judging on the Mal'ta data (which belongs to Gravettian) would also include other R1 branches now rare or extinct).

Hence the first Western/Siberian colonization wave (Aurignacoid) was essentially I in Europe (J in parts of West Asia and NE Africa but probably Q1 in Siberia - else this lineage could not have spread UP in NE Asia c. 30 Ka and reached trans-Beringian America c. 17 Ka BP). R1 was also there, as was also probably the case of G, but the real expansion of R1 (R1b for West Europe and some other areas) took place only with the Gravettian expansion, having a re-expansion somewhere (M412, where?) with Magdalenian (see the update here:

That's well calibrated full sequence molecular clock.

The available sequences are just first evidence of the implicated haplogroups, just as Sahelanthropus is just the first evidence of Hominids. This does not mean that they are the oldest ones in their respective kinds, not at all, just the oldest ones we happen to know. The oldest one is ALWAYS older than the references we can find. Mal'ta's R1* is already heavily derived R1, so it can't be considered near the root of R1, as you seem to try to do. It just means that R1 is older than the LGM.


Maju said...


"The Y-DNA of Middle Paleolithic humans in East Asia may simply have gone extinct"...

It doesn't matter if some lines went extinct or became very rare. I'm not saying at any time that there were not demic replacements in Neolithic, just that they mostly seem unrelated to most haplogroup expansions. They are re-expansions showing only weaker founder effects, if any at all.

In East Asia we know that O3 was there early in the Neolithic but not in North China, as some wanted to believe, but rather towards the center-south. It replaced N and Q in the North, as well as (partly) O1 in the East. See:

What is clear is that the O and NO expansions were Paleolithic and even most of that of O3 is. Then there were replacements but with much weaker founder effects (settlers originated from many founder lines within certain genetic pool(s), there was no time for near-fixation as the population was growing fast).