New blogs

Leherensuge was replaced in October 2010 by two new blogs: For what they were... we are and For what we are... they will be. Check them out.

Tuesday, January 27, 2009

On Eurasian mtDNA


A quite lengthy
discussion with (primarily) Terry Toothill at Dienekes' Anthropology Blog on Eurasian genetics and particularly mtDNA and the possible routes of migration and coalescence areas has caused me to review and re-consider Eurasian mtDNA genetics, often poorly understood because of higher complexity, less clear-cut nomenclature and somewhat less interest in general as of late.

After careful consideration of the genealogical tree (see Ian Logan's mtDNA site for reference), I came with the following simplified map:


Legend:
  • Red large dots represent top level subclades of M, small red dots mean significative presence of an M subclade likely to have originated in some other region.
  • Blue large dots represent top level subclades of N(xR), small blue dots mean significative presence of an N subclade likely to have originated in some other region (notably Australian N* related to West Eurasian W), blue squares represent N* (normally a single top-level N subclade but hard to confirm). An exception is East Asian N9-Y, actually related to R at its genesis but considered a separate haplogroup everywhere, which has been here considered like any other N(xR) subclade for simplicity.
  • Green large dots represent top level subclades of R, small green dots mean significative presence of an R subclade likely to have originated in some other region. R is a derivate subclade of N, linked to N9-Y at its origin but because of its widespread distribution it is always considered separately.
  • Grayed out areas are not really considered for various reasons.
  • Regions are marked by gray borders and 3-letter abbreviations in black: WEA (West Eurasia), SOA (South Asia), AND (Andaman Islands), SEA (SE Asia), AUS (Australian Aborigines), MEL (Melanesia), EAS (East Asia), CAS (Central Asia), KET (Ket people of the Yenisei basin), NAM (North American Natives), SAM (South American Natives).

Discussion:

It seems self-evident from the top-level diversity (and central geographical position) that South Asia was with all likehood the place of coalescence of haplogroups M and R. The case is less clear for haplogroup N and the genesis of R itself is directly associated with a major East Asian clade (N9-Y).

Pre-R (i.e. the N-derived ancestor of R and N9-Y) could have been South Asian or not (in which case we would have to look to East Asia). But considering the context, I'm more inclined to think that pre-R lived in South Asia and that R is its direct derivate in that region, with N9-Y representing maybe an early migration into East Asia.

The context is not just geography and common sense but also the problem of N as such. If we are to follow the clue of highest top level diversity, West Eurasia stands (by narrow margin admittedly) as the most likely candidate for the origing of N. It is not a too solid conclusion but it does make some good sense anyhow.

Specially if, as I do, you suspect, based on archaeology, that the presence of H. sapiens in Asia may be as old as c. 100,000 BP, long before the Toba supervolcano event and the expansion of H. neanderthalensis into West and Central Asia.

I do understand that the main expansion of H. sapiens in Eurasia happened after these two events: towards the east after Toba (i.e. after c. 74,000 BP) and towards the west after a phase of losing ground to Neanderthals (i.e. after c. 60,000 BP).

But some H. sapiens may well have survived in West Asia after the Neanderthal expansion and, in my opinion they may well have carried haplogroup N, either still undifferentiated or in the process of evlution towards the modern West Eurasian N subclades (all them quite rare): N1, X and W. These N carriers would have also migrated eastward after Toba through South Asia, mixing to an extent with the carriers of M, who were already in the subcontinent. Together (more or less, the exact process is not known) they must have migrated towards East Asia and Sahul, roughly at the same time when N was evolving into "pre-R" (R plus N9-Y) and into "pre-W" (W plus directly related Australian N*), as well as when pre-R was evolving into R (soon after).

This expansive epysode must have happened soon after Toba, I think, because only something of the dimensions of the Toba catastrophe would have left so much empty land for such a massive migration, evident in the huge ammount of new clades at this genealogical "point".

After R had already coalesced (necesarily in South Asia), there was surely a "back-migration" towards West Asia that brought R (already evolving into U and maybe also into R0, aka pre-HV, and pre-JT). This wave eventually colonized Europe, as it's well known, and we do have a clear date for that: c. 45-40,000 years ago - and at least U subclades participated in that migration. So the whole process of divergence of N and then of "pre-R" and of R itself happened probably between 74,000 and 40,000 years ago. This was surely also the moment of diversification of M (I'd say that rather to the beginning, soon after Toba).

And this timeframe (c. 60-40,000 BP) is when we do have the best available archaeological evidence for expansion of H. sapiens east of South Asia and west of West Asia. Though archaeological evidence in some cases may be of later date than the actual events, specially in those areas not throughtly researched.

West and South Asia though, as well as North Africa, do have some evidence pointing to older presence of H. sapiens. In the case of West Asia and North Africa this includes the oldest skulls north of the Sahara and are dated to c. 90,000 BP (North Africa, modern solid datation) and c. 130,000 BP (West Asia, old contoversial datation).

In the case of South Asia, it has the oldest evidence worldwide of stone blade technology (dated c. 103,000 BP) as well as strong indications of survival and continuity through the Toba epysode. There are no human remains though until much later, possibly because the tropical climate (wet and hot) makes preservation less likely even for bones.

Below there is a simplified map for a plausible scenario:


Notes:
  1. I have depicted migration into Sahul (Melanesia and Australia) as two separate events. This matter is not fully agreed upon and, admittedly, I depicted it that way on a mere whim.
  2. There is an interesting problem regarding SE Asia: it must have been necesarily a passage for the migrations into East Asia and Sahul, yet it appears to keep a relatively low index of high-level diversity (excepted the Andaman islands, where not just two unique subclades of M but also a distinct fossil clade of Y-DNA D appear to suggest that the regional diversity was much higher in the past). In comparison East Asia appears as much more diverse. I guess that Terry might want to make a case out of this matter but, sadly for his hypothesis, Central Asia does not appear to show any autonomous top-level diversity at all. Only the rare Ket N* might suggest that some N-derived clades (but not anything else) might have been in Northern Asia (rather than Central Asia) since old but how N arrived there and exactly when is a total mystery.
.

8 comments:

Manju Edangam said...

Thank you for this elaborate summary.
In the case of South Asia, it has the oldest evidence worldwide of stone blade technology (dated c. 103,000 BP) as well as strong indications of survival and continuity through the Toba epysode.

If I remember correctly there was a Genographic report that a hominid (likely Homo Hiedelbergensis) roamed around South Asia around 250,000 years ago. What about their continuity? There was also another study that speculated other hominids like Homo Neanderthalensis were capable of Blade technologies.

I just wonder if digging Himalayas may give some idea of colonization of South Asia.

Maju said...

Yes, technically there is no direct evidence of human presence in South Asia: the oldest skull of the region is dated to c. 36,000 BP (from memory), quite later than almost any other region in the Old World.

But lack of evidence is not evidence of lack, specially when there is indirect evidence in the form of:

1. Technological continuity before and after Toba, with assamblages that appear very similar to South African MSA
2. Documented presence of H. sapiens in neighbouring regions (i.e. West Asia)
3. The DNA pool with the highest diversity at such old stages

There was also another study that speculated other hominids like Homo Neanderthalensis were capable of Blade technologies.

H. Neanderthalensis is documented in Asia only AFTER H. sapiens: at least 40,000 years later. H. erectus does not appear to have evolved enough brains to do anything half advanced and its remains never appear in relation with Middle Paleolithic assamblages (Mousterian or the like). Also, there is no direct evidence of these hominids in South Asia (or anywhere else, AFAIK) after the date you mention.

I just wonder if digging Himalayas may give some idea of colonization of South Asia.

Why? Too high and too cold, right? But I would certainly ponder the Brahmaputra-Yellow River route (i.e. the Ledo Road). In that area the moutains are lower and much more transitable, even if they still act as some sort of barrier. If I'm not wrong the highest pass (Pangsau) is only some 1000 m. high, what is not really impassable at all.

Actually the claimed as oldest AMH skull in East Asia (a quite modern Mongoloid type) was found in that area (southern China near the Burmese border) but sadly the interest on its antiquity did not develope until much later, so the only datations are extremely contoversial (albeit very old: pre-Toba).

Maju said...

PS- Correcting myself. Actually the Liujang skull has been dated (indirectly) to c. 67,000 BP (not pre-Toba, as I said before) and was found at Guangxi-Zhuang (i.e. much closer to Vietnam than to Burma).

If the datation is correct, H. sapiens would have entered East Asia (by whichever route) soon after the Toba event, which would be in agreement with a pre-Toba presence in South Asia but would be hard to explain if H. sapiens was still limited to Southern Arabia, as the post-Toba model seems to suggest.

Manju Edangam said...

the oldest skull of the region is dated to c. 36,000 BP (from memory), quite later than almost any other region in the Old World.

And that too in Sri Lanka. I wonder if it's from north-west of the subcontinent or from south-east Asia to South Asia.

The DNA pool with the highest diversity at such old stages

Many studies have consistently shown mtDNA coalescence age is higher in East Asia and Oceania than in South Asia.

H. Neanderthalensis is documented in Asia only AFTER H. sapiens: at least 40,000 years later

If we agree that H. Heidelbergensis is the ancestor of H. Neanderthalensis then another descendant of H. Heidelbergensis in South Asia would have equal mental capacity.

Maju said...

"the oldest skull of the region is dated to c. 36,000 BP (from memory), quite later than almost any other region in the Old World".

And that too in Sri Lanka. I wonder if it's from north-west of the subcontinent or from south-east Asia to South Asia.


This doesn't need to mean anything. The lack of skulls may derive from several reasons:

1.- Tropical (hot and humid) conditions are awful for the preservation of any sort of organic materials. Not sure if this may apply to South Asia but acid soils are also extremely destructive of bone remains (it happens in Portugal, for instance, for much later dates).

2.- Lack of sufficient research.

3.- Luck.

Many studies have consistently shown mtDNA coalescence age is higher in East Asia and Oceania than in South Asia.

Not sure which are those "many studies" nor on what are they based. In any case, to my eyes the highest top-level diversity (of unique clades, totally or mostly: it's not diversity caused by immigration from various origins) is clearly in favor of a South Asian urheimat for R and M. In the case of R there's just nothing comparable and in the case of M only East Asia may approach it somewhat but still South Asia keeps the prominence.

If we agree that H. Heidelbergensis is the ancestor of H. Neanderthalensis then another descendant of H. Heidelbergensis in South Asia would have equal mental capacity.

No, the ancestor of Neaderthals is H. antecessor (a local European evolution from undifferentiated H. erectus), ask Arsuaga. H. heidelbergensis (East African variant of the genus) is ancestor (probably) of H. sapiens. There are one or two skulls in East Asia that show some intermediate features between H. sapiens and the H. erectus, indicating that some of the same direction of evolution was taking place there as well, but far from the capacities of Neanderthals and Sapiens in any case.

There's no particular evidence supporting that South Asian middle UP cultures were work of any specific Homo species, except one: the extreme similitude with African MSA, which is generally accepted as work of H. sapiens. For me there is little doubt that Jawalpuram culture (and much of South Asian MP) was work of H. sapiens, especially for that reason (and also because H. sapiens is found in West Asia long before that -and long before Neanderthals made their own "out-of-Europe" brief epic- and because the short chronology of a post-Toba OOA is way too short to explain things properly).

Manju Edangam said...

Regarding M:
There are at least couple of studies that have found mtDNA M age is greater in East Asia and Oceania than South Asia. I tried to find them at Quetzalcoatl but no luck. I remember commenting on those studies.

Regarding N:
Most of the daughter clades of N,
A, S, X and Y are observed in East Asia and Oceania. R is widespread. Only I and W are found exclusively in Near East and Europe(I'm not sure of the explanation here). I don't know why I should bother about R at all. It appears except R all other haplogroups experienced bottlenecks. I think understandable considering their distribution in Siberia and Oceania. However, you should equal importance to all those clades. In my opinion, that shows initial non-South Asian distribution of macroplogroups M and N.

Maju said...

There are at least couple of studies that have found mtDNA M age is greater in East Asia and Oceania than South Asia.

Never heard of them. Anyhow age estimation is a most inexact discipline of the gentic science.

I'm saying that there are more top-level M subclades (i.e. those derived directly from the M node) in South Asia than anywhere else. True that East Asia is second but all Melanesian M are derived from the same node and must be considered a single branch of M. Australian M is different but largely related to West Eurasian W anyhow and not particularly diverse otherwise.

Check Logan's graphs for details.

Regarding N:
Most of the daughter clades of N,
A, S, X and Y are observed in East Asia and Oceania. R is widespread. Only I and W are found exclusively in Near East and Europe(I'm not sure of the explanation here).


I mean top-level branches: those diectly derived from the N node (and not those named in the 1990s before the whole sequence was known, I for instance is absolutely trivial, as it's just a subclae of a subclade of a subclade of N1). These are, following Logan, exactly six:

1. R-N9-Y ("pre-R" hereafter): these clades share a common acnestor, hence they are a sigle N subclade. It later splits into N9-Y and R.
2. A
3. N1-X ("pre-X" hereafter): later split into N1, X and certain Indian N* sequences (N1 splits in several branches and one of the subclades of one of these branches, N1a to be precise, is I)
4. Ket N* (also found in Europe)
5. "pre-W": which includes West Eurasian W (also found in South Asia) and Australian N*
6. S (Australia)

Of these branches, Ket N* and A show many SNPs at the root, suggesting a long non-expansive history before consolidation. The other four clades are defined by a single SNP instead and therefore should be considered that they expanded rapidly after N split.

I notice that I have to refine the map to describe this N structure more properly. I also notice that, as N1 and X (and related Indian N*) are one single top-level clade, we do not actually have highest diversity in West Eurasia but that this peculairity is shared with South Asia and Australia.

While I review the map, this is the N top-level subclade distribution:

- West Eurasia: pre-X (N1 and X), pre-W (W), pre-R (some R derivates). Candidate for homeland of the two first clades listed.
- South Asia: pre-R (lots of R-derived clades), pre-W (W), pre-X (Indian N*). Most commonly acknowledged urheimat of R (but not necesarily of pre-R)
- East Asia: pre-R (N9-Y and a a couple of R-derived clades), A.
- Australia: pre-W (Australian N*), S.
- Northern Eurasia: Ket N*

Genesis of these oldest N subclades?

- pre-R must have developed between South and East Asia. R evolved in South Asia, while N9-Y is clearly East Asian. I'd think of pre-R as South Asian but could also be East Asian - impossible to determine.
- pre-X and pre-W must have evolved either in South or West Asia. There is a slim chance that pre-W evolved in Australia but I see no logic behind it. Considering its geogpraphy, South Asia must have played a major role in its spread anyhow.
- S must have evolved in situ in Australia or soon before arrival
- A and Ket N* may have evolved nearly anywhere because what we see now is a very derived version of the ancestral clades. It is possible that they represent early "adventurers" into middle or northern Asia and that they had limited chances for expansion due to climatic constrains.

This is what I see and will soon post again with a reviewed map to make it more clear.

Maju said...

Erratum:

I wrote: but all Melanesian M are derived from the same node and must be considered a single branch of M. Australian M is different but largely related to West Eurasian W anyhow and not particularly diverse otherwise.

This is outmost incorrect. Sorry.

W is N-derived (of course).

Not all Melanesian M is included in the same haplogroup: M27,28,29 is a single haplogroup but Q is distinct. Australian M is composed of Q and an Australian-specific clade, M42. Still the whole Sahulian diversity of M is of just 3 subclades, not larger than any other region but West Eurasia.