October 18, 2010

On the high mobility of mtDNA macro-haplogroups in Eurasia

This review of Eurasian mtDNA is triggered by some rather interesting debate in this other thread. Not really unexpectedly, the structure of the scatter of mtDNA N is being questioned by an habitual reader, who seems to prefer (without any evidence I know of) a "northern route" for the migration of this haplogroup. 

Here I just visually depict the main patterns of earliest scatter of the three pan-Eurasian mtDNA macro-haplogroups: M and N, both directly derived from African L3, and R, derived from N.For that I use exclusively those basal sublineages of these three macro-haplogroups which are one control region (CR) mutation downstream of the main node. Why? For simplicity and because these sublineages are the ones that show, most likely, the earliest sub-expansions of each of the three macro-haplogroups. 

Also, if you look at the other sub-lineages, you do not see strong deviations from the general pattern described here and, when that happens, it can and must be attributed to peculiar histories at the stem of those sub-haplogroups. This is for instance case of mtDNA A, a basal sub-lineage of N, which expanded in NE Asia but that shows a rather long stem with 6 CR mutations, indicating a lengthy "invisible" history of its own between the N expansion and that of A.

However I must emphasize that, in spite of the similitude in their spread patterns it is rather likely that the three patterns of diffusion were not simultaneous. At least seen from the viewpoint of the shared ancestral L3 node (East Africa), M is only detached by 3 CR mutations, while N counts 5 and R 6 (the five of N plus one of its own). This is at least suggestive of successive rather than strictly simultaneous process but we cannot totally exclude that the three expansions were more or less simultaneous (or separated only by short periods) because the so-called "molecular clock" is not actually bound to tick with strict regularity but actually has some strong uncertainty.

For the source of the information used to build these maps (ultimately from the reference mtDNA site: PhyloTree), see Leherensuge: Reconstruction of mtDNA spread in Eurasia (again), which shows also a possible time structured sequence based on a hypothetical strict ticking of the "molecular clock" and also those haplogroups derived from M, N and R at longer CR mutation counts (presumably later in time).




What I find most interesting in these three maps is the similitude of the main pattern for all three macro-lineages from NW South Asia to New Guinea, via SE Asia. 

The differences are as follows:
  • M shows a clear distinct early trend northwards along the East Asian coast, which probably also carried the ancestors of mtDNA A. 
  • Instead R shows an early trend westwards, which is probably quite more recent than the M expansion, and does include, once you look at more derived sub-lineages (not shown here), three other N subclades (N1, X and W), two other R subclades (U and JT, the one shown is R0), and at least one M subclade (M1). The fact that all these Western lineages are rather derived, also suggests a later time-frame for the expansion of R, some time (10-20,000 years?) after that of M.
It is interesting also that N(xR) shows no scatter peculiarity of its own, in spite of having left a notable legacy.  It is also interesting to see that all three macro-haplogroups left an early mark in Melanesia, what suggests that they did indeed have some decent boating skills from early on, and is indeed supportive of the "rapid coastal migration" hypothesis, regardless of whether this model is not needed to explain some other aspects of the Eurasian spread.

44 comments:

  1. Interesting work Maju. However:

    "This is for instance case of mtDNA A, a basal sub-lineage of N, which expanded in NE Asia but that shows a rather long stem with 6 CR mutations, indicating a lengthy 'invisible' history of its own between the N expansion and that of A".

    The same holds true for other Ns of course. We have N13 in Australia with 13 mutations, yet S with just one. S makes it to your diagram, but N13 doesn't. Does that imply that N13 is relatively recent to Australia? Or is it more likely it remained for some considerable time in Australia before its own expansion within that continent? Similarly with N14 (10 mutations) and O (3 mutations). And we have N21 in Malaysia with 8 mutations and N22 (7 mutations). Again, does this mean they are relatively recent to Malaysia, arriving there after S had reached Australia? I happen to think that is quite possibly so, but it seems unlikely in your scenario.

    "It is interesting also that N(xR) shows no scatter peculiarity of its own, in spite of having left a notable legacy".

    Only if you ignore the other three N haplogroups: X (7 mutations), N2/W (5 mutations) and A (now with 8 mutations in the Phylotree). Is it not quite likely that these haplogroups coalesced in separate regions somewhere near where they are found today, and are part of an original N expansion?

    Just in passing I note that in the Phylotree you've linked to R is now given two mutations (12705 and 16223). Doesn't this mean the haplogroup should be eliminated from consideration?

    "who seems to prefer (without any evidence I know of) a 'northern route' for the migration of this haplogroup".

    I have long believed the distribution of mtDNA displays exactly such a route. The evidence of long-term human (whether 'modern' or not) through Central Eurasia offers evidence in support of the possibility of such a route. So not 'without any evidence'.

    ReplyDelete
  2. We can't know anything but the mutation list between the upstream node and a secondary (downstream) one. We can reasonably infer where a haplogroup (node) began scattering around, at least in most cases, but we cannot know what happened in between.

    The four minor Australian and SE Asian haplogroups only inform us of the overall basal diversity of N (and poorly so, because of the reasonable doubts that may exist about the actual structure of such ill-researched lineages). They do not tell us anything about what happened between the N multifurcation and the corresponding modern lineages, known only in one variant.

    As they are not found elsewhere, we can reasonably assume that they have been there for long but the genetic data is not informative about how long is that.

    "Is it not quite likely that these haplogroups [X, N2'W and A] coalesced in separate regions somewhere near where they are found today, and are part of an original N expansion?"

    N2'W is similar to N1'5, Western R and M1'51 in its spread. It looks much younger but otherwise it fits perfectly in the Western scenario described by these three other populations. Similarly X has a clear West Asian (Palestine?) origin and can be explained in full as part of the West Eurasian expansion, correlating at X2 level with the Central and North Asian expansion also visible in other lineages, male and female, even extending to America.

    The only 'odd man out' is A and just because it is found in the context of NE Asian colonization, where no other N subclade is found (but several M ones are). Again it poses no big problem when you look at the broader picture and see all those M lineages migrating northwards along the Western Pacific coast. But even if you wish to argue some other route for A, it would not change anything in regard to N as a whole because A after all is just one of 12 lineages.

    "Just in passing I note that in the Phylotree you've linked to R is now given two mutations (12705 and 16223)".

    No changes, I am using only the control region (coded in black at PhyloTree) and not the often misleading hyper-variable region (in blue, often beginning by 16-), which mutates too fast and I consider unreliable even often for mere haplogroup identification purposes. I decided to work that way some time ago and I still find it more solid and meaningful than mixing both kind of mutations.

    Consider HVS as a some sort of randomly placed "fractional mutations" if you wish to compare with the C.R. ones. Not enough to make a difference, IMO.

    ReplyDelete
  3. "I have long believed the distribution of mtDNA displays exactly such a route. The evidence of long-term human (whether 'modern' or not) through Central Eurasia offers evidence in support of the possibility of such a route. So not 'without any evidence'".

    You mix Neanderthals, H. erectus and H. sapiens too happily. That is not evidence of any H. sapiens colonization in the Altai (the best documented are and the oldest one probably) before 40+ Ka. We have older confirmed dates for New Guinea and Australia, and it seems most likely (per the archaeology) that the the presence of our species in Tropical Asia dates to at least 75+ Ka in South Asia (Jwalapuram, etc.) and 65+ Ka in SE Asia (Luijiang, etc.). It may be even older.

    ReplyDelete
  4. "I am using only the control region (coded in black at PhyloTree)"

    Thanks for that clarification.

    "The only 'odd man out' is A"

    Not really. It's actually very difficult to make a case for a SE Asian origin for N9/Y. It has just one mutation before it breaks up into N9a, N9b and Y. N9b is Japanese. Y is especially common in the Nivkh, Ainu etc., as well as some in SE Asia where it is probably a relatively recent immigrant. N9a is East Asian, Central Asian and SE Asian, again probably a relatively recent immigrant at the southern geographic margin of its range.

    "Instead R shows an early trend westwards, which is probably quite more recent than the M expansion"

    And that's where I agree completely. Where was the source? Presumably not New Guinea, although P is obviously an early arrival there. Perhaps from somewhere very nearby?

    I've had a look at the Phylotree for R and I guess that you've used haplogroups R0, R2/JT, R6, R11/B, R30, R31 and P, but that makes just seven, unless you've split R11/B6 into two?

    That's interesting if you did. That haplogroup is definitely East Asian, and P (with one mutation) is Melanesian. Most of the Indian haplogroups, apart from R6, R30 and R31, are actually downstream mutations. And R6 is concentrated on the east coast of India, in Orissa. So that leaves just R30 and R31 as the only basal haplogroups definitely coalescing in India. And when we widen the view to take in the two mutation level we find just R12'21 (with R12 in Australia and R21 in Malaysia and in the Negritos), and R9/F2, another Eastern haplogroup. So I totally agree, 'R shows an early trend westwards, which is probably quite more recent than the M expansion'.

    ReplyDelete
  5. I believe we have discussed before the scatter of N9 and that the conclusion was that there was enough N9b diversity in SEA to suggest a coalescence of N9 there rather than in the North. But I am unsure as of why of this conclusion, so I can't discuss till I confirm either way.

    An N9 coalescence in NEA would make the geography of earliest N more like that of earliest M, replacing the emphasis in SEA for one in NEA. It can be read in a northern interpretation for N9 and A but only them: the several N clades in SEA and Near Oceania don't really support a Northern route for the whole pre-N lineage.

    "Where was the source? Presumably not New Guinea, although P is obviously an early arrival there. Perhaps from somewhere very nearby?"

    In South Asia, where the diversity at all levels is the highest by far. Please! For R and M there is really no room for doubts.

    "I've had a look at the Phylotree for R and I guess that you've used haplogroups R0, R2/JT, R6, R11/B, R30, R31 and P, but that makes just seven, unless you've split R11/B6 into two?"

    I considered R11'B7 and B4'5 as separate haplogroups, as they are only held together by a single HVS-1 mutation, which I am ignoring for practical purposes. Does not make much of a difference anyhow.

    "And R6 is concentrated on the east coast of India, in Orissa".

    R6 is widespread, from Kashmir to Tamil-Nadu (Metspalu 2004).

    ReplyDelete
  6. "R6 is widespread, from Kashmir to Tamil-Nadu"

    Sorry. I meant R8 concentrated in Orissa.

    "You mix Neanderthals, H. erectus and H. sapiens too happily".

    And you are too convinced of their complete separateness. Somewhere between the two positions is probably close to the actuallity.

    "In South Asia, where the diversity at all levels is the highest by far. Please!"

    You really are committed to an India origin. To me the evidence is overwhelming for an SE Asian origin for mtDNA R. If you could accept that you'd see that there is then no Indian basal N. India was occupied by M before R entered. From your link:

    "With the exception of the diverse set of largely Indian-specific R lineages, the most frequent mtDNA haplogroup in India that derives from the phylogenetic node N is haplogroup W [13]. The frequency peak of haplogroup W is 5% in the northwestern states – Gujarat, Punjab and Kashmir. Elsewhere in India its frequency is very low (from 0 to 0.9%)"

    Therefore quite likely an immigrant from the northwest.

    "Over 90% of the mtDNAs found in Iran belong to haplogroups HV, TJ, U, N1, N2 and X, commonly found in West Eurasia"

    So N1 and N2, basal N haplogroups, are common in Iran. Back to your comments:

    "Similarly X has a clear West Asian (Palestine?) origin"

    So there you have it: N's expansion route from Africa. X in Palestine, N1'5 in SW Asia, N2/W in Iran, A in Central Asia, N9/Y in NE Asia, R in SE Asia, O and S in Australia.

    ReplyDelete
  7. "To me the evidence is overwhelming for an SE Asian origin" [of mtDNA R]...

    What evidence?

    ReplyDelete
  8. "I meant R8 concentrated in Orissa".

    Surely spread from further south: Andrah Pradesh.

    But anyhow R8 is not one of the early expanding clades of R: it has a stem made up of 5 CR mutations.

    ReplyDelete
  9. Thanks for the link.

    "Surely spread from further south: Andrah Pradesh".

    Concernig R8 the authors actually say:

    "the highest frequency of this haplogroup occurred towards East India, especially within Orissa (12%) (Figure 4), whereas low frequencies occurred in the Gujarat (1.8%), Madhya Pradesh (0.53%), Uttar Pradesh (0.22%), Andhra Pradesh (0.9%), Chhattisgarh (6%), Jharkhand (1.04%) and Tamil Nadu (0.18%) populations"

    They're even kind enough to provide a map:

    http://www.plosone.org/article/slideshow.action?uri=info:doi/10.1371/journal.pone.0006545&imageURI=info:doi/10.1371/journal.pone.0006545.g004

    So, as I said, 'concentrated in Orissa'. Looks suspiciously as though it arrived via the coast, from further east.

    "What evidence?"

    The evidence of haplogroup distribution, for a start. In the link the authors say:

    "Existence of a comparatively high frequency of R8 in Orissa populations, especially among the AA-speaking Mundari tribes, strongly suggests that this haplogroup might have originated among the maternal ancestors of the contemporary AA speakers of the region".

    In spite of their belief that the AA languages are the earliest spoken in India I have my doubts. The possibiltiy of discerning a connection with SE Asian languages going back more than 40,000 years is extremely unlikely.

    "But anyhow R8 is not one of the early expanding clades of R: it has a stem made up of 5 CR mutations".

    True. Which makes its arrival from SE Asia even more likely. But not only R8's distribution. Other R haplogroups have equally interesting distributions. From Wiki:

    http://en.wikipedia.org/wiki/Haplogroup_R0_(mtDNA)

    "Haplogroup R0 occurs frequently in the Arabian Plate with its highest frequency in Socotri (Population 50,000 Yemen) 38% [4] and its also found in a high frequency in the Kalash (Population 6,000 in Pakistan) with 23%[5] smaller frequency in North Africa, the Horn of Africa, Anatolia, Iranian Plateau & Dalmatia. Its greater variety in the Arabian Plate suggests R0a originated in and spread from there".

    Socotra? Must have got there across the sea. Possibly arrived on the Yemen coast by a southern coastal migration, but from east to west.

    continued

    ReplyDelete
  10. "In South Asia, where the diversity at all levels is the highest by far. Please!"

    I don't know where you get that idea from. You provided eight haplogroups in your map, but it's interesting to include the two haplogroups with just two mutations before they expand. We have:

    1) P, which breaks rapidly into 7 new subdivisions. Surely that indicates there is nothing long distance about its arrival in New Guinea.

    2) R12'22, two mutations but it separates immediately into R12 in Australia and R21 in Malaysia and the Negritos. Admittedly both with long tails but that doesn't mean they were not present in each region for a very long time before they each expanded, or perhaps drift has eliminated relative haplogroups.

    3) B4'5, separates rapidly into the two haplogroups, but B4 rapidly separates still more. All haplogroups being Far Eastern. Again it looks unlikely it originated very far away.

    4) R11'B6, one mutation and it separates into R11 and B6, both with long tails, but definitely eastern.

    5) R9/F, two mutations again, but F especially has a rapid diversification, and the haplogroups are definitely SE and E Asian.

    6) R30, Indian, not particularly diverse, and subgroups have long tails.

    7) R31, also Indian but with a long tail before it diversifies.

    8) R6, Indian again but not particularly diverse in basal haplogroups.

    9) R2/JT, spread through west Asia including India, Baluchistan, Iran, Georgia. JT breaks up fairly rapidly.

    10) R0, separates rapidly into R0 and HV and is another rapid expander, but see above.

    That leaves R14 (New Guinea and Nicobar), R22 (Indonesia, Lesser Sunda Islands), U, R1 and R3 (West Asia), R8 (Orissa) R5 and R7 (ahh, at last, India).

    So there you have it: mtDNA R's expansion.

    ReplyDelete
  11. Re. R8, I just forgot to add a note mentioning figure 3, which shows that highest diversity, specially basal one, is from Andrah. Orissa lacks R8a(xR8a1), while Andrah includes all important subclades.

    In that graph the pattern of spread of R8 is quite obvious:

    (Origin: Andrah)
    R8b: Andrah-Orissa-Madhya-Gujarat
    R8a: Andrah
    >R8a1 Andrah-Orissa-Jarkhand/Madhya

    It migrated from Andrah northwards with the more dynamic branch (R8b) reaching as far as Gujarat, while the other remained in the East coast.

    There's no connection with Austroasiatic peoples other than the Munda. South Asian Austroasiatics are characterized by having high SE Asian Y-DNA but almost only South Asian mtDNA. Hence the association with Austroasiatics (mild) is one of an adopted lineage.

    Re. R0, the largest part of R0 is HV and the largest part of HV and R0 is H. That quote only speaks of R0(xHV), which is essentially R0a. R0(a) is the former pre-HV, which got renamed. R0 branched into R0a (West Asia, rather derived) and HV (West Asia and Europe, more basal). R0a should, per the my rustic approach to the "molecular clock", contemporary of H in its expansion. And H to me looks European or almost: a signature of the colonization of Europe by H. sapiens, maybe 40 Ka ago.

    So, not sure what you're seeing in R0a but it expanded at the time of H more or less (but in a more gradual fashion). It's likely that the coastal expansion you are talking about (or whatever it is) happened rather at the R0a2 stage or something like that.

    ...

    ReplyDelete
  12. "You provided eight haplogroups in your map"...

    We have gone through this before - see here. The diversity found in South Asia has no comparison.

    I agree with the content of your points re. each of the sublineages of R. While it's obvious that R expanded from South Asia (highest basal diversity by far), it didn't have too much room for its sublineages to have secondary expansions. Instead in East and West Eurasia it did. (Why? I don't care: there was less effective demic pressure, whatever the exact causes).

    I recapitulate here:

    West+South Eurasia: 9 R sublineages (8 present in SA and 5 exclusive of SA).

    East Asia+Sahul: 3 R sublineages (2 in East Asia, 3 if you divide R11'B, and one in Sahul).

    There can be no reasonable doubt for R coalescence area being SA.

    WEA+SA weight 3, EA+OCE weight 1 only. Within WEA+SA, WEA weights 1 and SA weights 3.

    You should recognize this before we can continue.

    ReplyDelete
  13. Erratum: within WEA+SA, WEA weights 7 and SA wights 15, almost 1:2 (not 1:3).

    This places the centroid of Western R in NW SA. The centroid of Eastern R is clearly in SEA but weights only 1/4, It places the overall centroid of R at the middle Ganges (Son valley?)

    ReplyDelete
  14. "Hence the association with Austroasiatics (mild) is one of an adopted lineage".

    Agree 100%.

    "West+South Eurasia: 9 R sublineages (8 present in SA and 5 exclusive of SA). East Asia+Sahul: 3 R sublineages (2 in East Asia, 3 if you divide R11'B, and one in Sahul)"

    Such numbers are hardly relevant. Surely the sort of basal diversity you're claiming here is a function of geographic extent, rather than being evidence for origin. It looks as though 'original' R became widely spread through all the regions, then diversified with daughter haplogroups replacing the parent line. The numbers of haplogroups tell us nothing about where it originated. Besides which there is actually no evidence for any centre of expansion for the haplogroup from anywhere in India.

    "I agree with the content of your points re. each of the sublineages of R. While it's obvious that R expanded from South Asia (highest basal diversity by far), it didn't have too much room for its sublineages to have secondary expansions. Instead in East and West Eurasia it did. (Why? I don't care: there was less effective demic pressure, whatever the exact causes)".

    The cause should be obvious to you. It originated there, but you refuse to accept it. It conflicts with your Out of India theory.

    The basal separation between Australian R12 and Malay Negrito R2 is surely significant. We see no such two-way split in India. But we do see a region of great basal haplogroup diversity in the region between where the two separate R12/R2 haplogroups finished up.

    The greatest diversity in any of the basal haplogroups is actually in SE Asia. In the Phylotree the two haplogroups, R11'B6 and B4'5, are connected by a single hypervariable mutation. But basically B4'5 splits immediately into B4, B5 and R24. B4 then immediately splits further, so we have considerable basal haplogroup diversity in the one region. B4 is commonest in Taiwan, Philippines, Tengarra and Java. It is less common through much of the rest of East and Southeast Asia. B5 is commonest in Sumatra and Java, also less common through the rest of E and SE Asia. R24 is found in the Philippines, although it has a long tail. This tail is quite possibly a product of its long isolation on the island environment.

    Even R9/F diversifies within two mutations of the 'original' R. F is concentrated in Taiwan and Sumatra, F1 in the Philippines, Moluccas, Tengarra, Borneo, Java, Nicobar (lesser proportions in India, Sumatra, Malaysia, Vietnam and Thailand), and R9 in South China and especially in Aboriginal Malays.

    So the SE Asian R haplogroups show the greatest basal haplogroup diversity, specifically those closest to Wallacea. And we have other R haplogroups there besides the three or four mentioned. R22, R23, R14 and P. So eight haplogroups around Wallacea (P, R2/12, R14, R22, R23, R11'B6, B4'5 and R9/F), seven haplogroups in India (R6, R30, R31, R5, R7 and adding R8 and R2/JT, which are both quite possibly more recent immigrants) and five haplogroups in SW Asia (R0, U, R1 and R3, with R2/JT appearing twice, in India and SW Asia). Nineteen R mtDNA haplogroups which, even by your own reasoning, have their greatest diversity round Wallacea.

    ReplyDelete
  15. "The numbers of haplogroups tell us nothing about where it originated".

    Of course they do.

    If you drop a bunch of insects and measure where are they 1 min. after, most will remain near the initial place (call it laziness or economy) and only a few would have spread all around.

    I am sure you already know this.

    That's why diversity and not mere frequency is what matters after all. Because diversity (in a phylogenetic context) tells us about the past, while frequency only tell us about the present.

    "The greatest diversity in any of the basal haplogroups is actually in SE Asia".

    That would be pretty much irrelevant even if true (as it details a second or later episode, not the R node). And it is not true (at least your explanation makes no sense that I can understand).

    ReplyDelete
  16. Also, please re-read this Leherensuge article, where Oriental N* and A appear to originate in SE Asia. And where not just N9a but also Y2 appear as common in SE Asia.

    ReplyDelete
  17. "your explanation makes no sense that I can understand"

    That seems to me to be because you don't wish to understand it. I've considered all 18 R haplogroups (or 19 if you split R11/B into two), not just the ones that suit a particular theory.

    "If you drop a bunch of insects and measure where are they 1 min. after, most will remain near the initial place (call it laziness or economy) and only a few would have spread all around".

    And that's why we find so many R-derived haplogroups around island SE Asia. Eight, in fact. Indian R haplogroups give the impression of having evolved separately from and expanding R* population.

    "That would be pretty much irrelevant even if true (as it details a second or later episode, not the R node)".

    Maju, none of the R haplogroups on their own can tell us anything about the placement of the R node.

    "Oriental N* and A appear to originate in SE Asia".

    From the first diagram:

    "Red for the Tonkin Bay area, Green for South China, Orange for North China, and Blue for Southeast Asia"

    The blue SE Asian is the very rare haplogroup N22, which no-one diputes is Malay. And that diagram tells us nothing about N having come via India.

    "And where not just N9a but also Y2 appear as common in SE Asia".

    Check again. The haplogroups in the diagram are:

    "Blue stands for Y2, and red for N9a6".

    Both are derived haplogroups, known to have made it to SE Asia. The map tells us nothing about the origin of N9 or Y.

    ReplyDelete
  18. I think you may be referring to info I am not looking at such as the latest version of Wkipedia's page on R, right?

    That data is new to me. And whoever made it did not take in account PhyloTree (for instance he makes R6 and R7 belong to a single clade when they are not). Even if R1 is moved to WEA and R2'JT and U are considered WEA clades, you still get only 4 haplogroups in the West, 7 haplogroups in South Asia (exclusive) and 9 haplogroups in SEA.

    I can only make sense of your argument following that Wikipedia page but most of the data is new to me, so I have not double checked. Still it'd make an unclear case, yet it would indeed make a stronger case for a SEA origin, more or less simultaneous to N and flow within N rather than autonomously.

    We would then have to conclude that the explosion of R and N happened closer in time (and geography) than would be expected by molecular clock averages (1-2 Ka instead of 5-6?). We would then have an eastward flow lead by M followed by a Westward flow led by N/R. That would be interesting: it would simplify all and would make mtDNA N and R migrate westwards with Y-DNA P, pushing definitively the origins of Y-DNA R and Q to the early colonization of West Eurasia.

    "Red for the Tonkin Bay area, Green for South China, Orange for North China, and Blue for Southeast Asia"

    In my language: red, green and blue for SEA, orange for NEA.

    You still may be right for N9 but not for A.

    ReplyDelete
  19. "I can only make sense of your argument following that Wikipedia page but most of the data is new to me"

    For the number of haplogroups I used Phylotree (as you've done), and Wiki supplied the distribution data (but not only Wiki). Phylotree has eighteen R haplogroups but I'm prepared to split R11'B6 and B4'5 as you've done. And considered R6 and R7 to be separate. Makes nineteen.

    "it would indeed make a stronger case for a SEA origin, more or less simultaneous to N and flow within N rather than autonomously".

    And would remove R as being Indian in origin, leaving a distinct lack of N haplogroups along any route through that subcontinent. Other N haplogroups in India (N1'5 and N2/W) look more like later expansions into India from the west, rather than being indigenous. I agree that R's expansion is basically simultaneous with that of N but, rather than being 'within N', it appears to be a continuation of that expansion, into India from SE Asia.

    "We would then have to conclude that the explosion of R and N happened closer in time (and geography) than would be expected by molecular clock averages (1-2 Ka instead of 5-6?)".

    Not necessarily so. As I said R can really be considered just a continuation of N's expansion.

    "We would then have an eastward flow lead by M followed by a Westward flow led by N/R. That would be interesting: it would simplify all and would make mtDNA N and R migrate westwards with Y-DNA P, pushing definitively the origins of Y-DNA R and Q to the early colonization of West Eurasia".

    Finally!!! That's what I've been trying to tell you all along. It also explains the two different haplogroup combinations in Australia and New Guinea. As you say, 'it would simplify all'. I've always considered it to be so obvious that it amazes me why every-one is so opposed to the idea.

    ReplyDelete
  20. If both N and R come from SE Asia, then I would think it's better to consider both expansions near simultaneous. Also all N lineages in West Asia (except X and R0 - but this one is part of R) are shared with South Asia. None is shared with Eastern Eurasia (except the X erratic into North America).

    While we can see the signature of expansions into Central and North Asia in other haplogroups (G, H, U, CZ, D, A, etc.), we do not see a single N basal sub-haplogroup that looks like expanding from Central/North Asia into West or South Asia. Neither we do in Y-DNA.

    For that and other reasons I still consider that at that stage most or all the action happened only in tropical and subtropical Asia.

    ReplyDelete
  21. "That's what I've been trying to tell you all along".

    Not really. What you have been saying all along is:

    1. A northern migration into Eastern Eurasia (false as far as I can tell).

    2. The development of basic boating only in Island SE Asia and not before (most probably false and anyhow unproven).

    3. Neanderthals magically become modern Homo sapiens by miraculous "introgression" (plainly impossible).

    ReplyDelete
  22. Anyhow, I think it's the way you present your arguments, assuming on one side that the reader knows everything you do or is thinking in whatever you are thinking, and also the stubborn defense of ideas against all evidence.

    In this particular case, now but not before, it seems that the evidence has slided into favoring your idea but this was not the case before and you still made your position once and again and never conceded. So the logical reaction by the other is, as with 'the boy who cried wolf', disregard your arguments as probably wrong - for excess of vehemence and for lack of evidence (or clearly presented evidence).

    ReplyDelete
  23. "I think it's the way you present your arguments, assuming on one side that the reader knows everything you do or is thinking in whatever you are thinking"

    Sorry about that. But I have at times become very frustrated at your continued refusal to consider the possibility of an outside-India origin for haplogroup N, not to mention R, or Y-hap C.

    "it seems that the evidence has slided into favoring your idea but this was not the case before"

    In fact it was the 'case before', but there was perhaps less easily seen because of less supporting evidence than we have now. I saw the connection when I first examined the 2005 McDonald maps. I'd been researching Polynesian origins and just continued back, unravelling the haplogroups. If you will remember you objected strongly to the concept when I made the suggestion in an essay at Tim's blog. I suggested Y-hap K at the time but the phylogeny has been refined and I'd now plump for MNOPS.

    "all N lineages in West Asia (except X and R0 - but this one is part of R) are shared with South Asia. None is shared with Eastern Eurasia (except the X erratic into North America)".

    Exactly. And that argues against the original arrival of N in East Asia as having been via India.

    "we do not see a single N basal sub-haplogroup that looks like expanding from Central/North Asia into West or South Asia. Neither we do in Y-DNA".

    Of course not. None have moved in that direction. What we do see is apparent evidence for an expansion of mtDNA N and Y-hap C through Central/North Asia into East Asia. Then to South Asia, and so to West Asia. Around in a big circle.

    "Not really. What you have been saying all along is: 1. A northern migration into Eastern Eurasia (false as far as I can tell)".

    Only for mtDNA haplogroup N and Y-hap C.

    "2. The development of basic boating only in Island SE Asia and not before (most probably false and anyhow unproven)".

    Something has to have given those East Asian haplgroups some advantage, probably technical. The only reasonably likely one I've been able to come up with is some huge improvement in boating ability.

    "3. Neanderthals magically become modern Homo sapiens by miraculous 'introgression' (plainly impossible)".

    Well, perhaps not impossible. What about your post on the Chinese find? And I've never claimed that Neanderthals were the only source population for modern humans. Obviously a large proportion of our ancestry is African, although how much is still open to question.

    ReplyDelete
  24. I or anyone will not accept any weird-sounding theory unless it is well exposed and has strong evidence behind it.

    Your ideas had not evidence behind (though you seem to have got lucky about this aspect and now some evidence is accumulating) and you have never bothered proposing them in an orderly manner in a space of your own (blog, web page, etc.). You cannot expose such a complex and controversial hypothesis as yours in short snippets such as these comments. You cannot even add graphs... so go figure!

    ReplyDelete
  25. "And that argues against the original arrival of N in East Asia as having been via India".

    I don't see how. See how you are being stubborn about something (northern route hypothesis) that you have zero evidence to support. You just make things sound like they somehow support your conjecture but you are lying when you do that.

    And really I'm not comfortable with lies. That's probably why I am so bad at the Capitalist power game, in which deceit is central to success.

    "What we do see is apparent evidence for an expansion of mtDNA N and Y-hap C through Central/North Asia into East Asia".

    No. We do not see that. All we see in mtDNA N and Y-DNA C is expansion from SE Asia. There is nothing, absolutely nothing telling of any presence of these lineages prior to expansion in Central Asia or any other non-tropical spot.

    Plus, crossing Siberian rivers without boats would be suicidal, not sure if you have ever considered this problematic fact to your conjecture. It is still "possible" to cross Bab el Mandeb or the Ganges Delta by mere swimming but no way you can cross the Obi or the Lena or Lake Baikal that way - not even in the hottest day of summer.

    "Something has to have given those East Asian haplgroups some advantage, probably technical".

    In truth we do not know if those groups had any advantage at all. Anyhow, I'd bet for atlatls or needles before boats if anything. Or even just a more individualist sociology or a more violent society... it does not need to be technical, it can be sociological. Similarly today we do not evaluate the world on mere technological parameters but we have to consider primarily the economic and social parameters instead. Technology can always be imported from other groups or created ad-hoc or circumvented by other means. Or it can happen, as with the steam engine in Antiquity, that a technological advance finds no room to become important now and then it does in a different socio-economical context.

    Inventions are not that important alone, people and their attitudes towards life (culture and even individual personality) is.

    "The only reasonably likely one I've been able to come up with is some huge improvement in boating ability".

    It looks extremely short-sighted to me, not to mention that boats are for us like any other tool: something we use for convenience. A tool and not a condition. We use tools, tools do not use us (unless the socio-economic frame re-shapes things that way).

    "Well, perhaps not impossible".

    It's impossible. While genes can go through the species "barrier", you cannot get one species from another: you get hybrids. Just like you cannot get a Gascon cow from a Frisian one, no matter you can have hybrids.

    "What about your post on the Chinese find?"

    It's not even clear if its mixed: the few "archaic" traits may just reflect, as happens in other cases, archaism in our species in that time (and much later, even today probably in some individuals) and nothing else.

    But, even if it's a mixed specimen, it's obvious we are not, so the apportion of mixed ancestors on us is minimal in any case. So we are not essentially descendant from any species other than H. sapiens. And that means culturally too.

    "Obviously a large proportion of our ancestry is African, although how much is still open to question".

    So far some 97.5% in the case of Eurasian peoples. 100% it seems in African peoples.

    For phylogenetic purposes we are 100% H. sapiens and that is what shows up in haploid lineages (logically).

    This also applies to culture.

    ReplyDelete
  26. "Your ideas had not evidence behind (though you seem to have got lucky about this aspect and now some evidence is accumulating)"

    Luck has nothing whatsoever to do with it Maju. It came from studying the evidence with an open mind, not being committed to 'the great southern migration route' theory.

    "See how you are being stubborn about something (northern route hypothesis) that you have zero evidence to support".

    Who is being stubborn here? You've consistently used R's presence in South Asia to justify your belief in N's migration eastward through that region. With R's coalescence now moved to SE Asia your evidence dissappears. All we're left with is members of N1'5 and N2/W being present in South Asia. However their presence there can easily be explained as movement from SW Asia.

    "You still may be right for N9 but not for A".

    A is still a huge difficulty for you. It remains very difficult to make a case for an SE Asian origin for it. Try for a moment to imagine a small pre-Toba population hanging on through the depth of the ice age in a south-facing valley somewhere between the Altai and the upper Amur basin. This would easily explain A's very long tail. And the Y-chromosome that eventually became C3 is the most likely candidate for its male companion. It is quite possible, of course, that its original male companion's haplotype has been replaced by a later arriving one. So the route N took to SE Asia is, as I said before: 'X in Palestine, N1'5 in SW Asia, N2/W in Iran, A in Central Asia, N9/Y in NE Asia, R in SE Asia, O and S in Australia'.

    "While genes can go through the species 'barrier', you cannot get one species from another: you get hybrids".

    Aren't you making the assumption here that each group of 'Archaic humans' was a separtae 'species'?

    "Just like you cannot get a Gascon cow from a Frisian one, no matter you can have hybrids".

    Bad news Maju. You can. It's called 'breeding up'. Start with a Gascon cow and cross with a Frisian bull. Cross the offspring again with Frisian bull, and so on. By the time the offspring are 15/16 Frisian (four generations) they're considered Frisain. Well, that goes for most breeds anyhow. Some purists do not allow breeding up, but those breeds are mostly suffering inbreeding depression. Fertility problems especially.

    ReplyDelete
  27. "Luck has nothing whatsoever to do with it Maju. It came from studying the evidence with an open mind, not being committed to 'the great southern migration route' theory".

    That's insultingly untrue. You are committed to your own pet hypothesis in a way that is clearly excessive, irrational.

    "You've consistently used R's presence in South Asia to justify your belief in N's migration eastward through that region".

    False. There's no particular genetic reason to say anything between L3 and M/N. And certainly R is pointless in this story.

    I think that the southern route is correct because mtDNA M says so and archaeological evidence supports it too. The role of R is less relevant once established that N coalesced in SEA.

    What I have said often is that R "unmistakably coalesced in South Asia", what now seems incorrect, but that has nothing to do with the OoA but with the backflow.

    That you thoughtlessly mistake everything related to South Asia as being the same thing is not my problem but yours.

    "A is still a huge difficulty for you. It remains very difficult to make a case for an SE Asian origin for it".

    It is no problem for me, it's your fetish. For me A is like every other N: ultimately original from SEA.

    Don't make accusations: you insist on A and I either stop listening to what you say (what may be something healthy to do after all) or discuss it. If your fetish would be M8a or JT*, I'd be in the same situation.

    "This would easily explain A's very long tail".

    There's nothing to explain about A's long stem: it's not the only haplogroup in such circumstances: its foremother N is in the same situation for instance. And the best explanation is minority lineage that randomly survived until it made a founder effect, which is why we see it now.

    "And the Y-chromosome that eventually became C3 is the most likely candidate for its male companion".

    There's a whole list of mtDNA lineages in the C3 area: D1 and CZ specially. I don't know why you chose A of all them.

    "So the route N took to SE Asia is, as I said before: 'X in Palestine, N1'5 in SW Asia, N2/W in Iran, A in Central Asia, N9/Y in NE Asia, R in SE Asia, O and S in Australia'".

    That's not how we do things here. I don't define a "route" by the resulting haplogroups, what I define is a 'centroid', the likely origin of all the shared haplogroup (N in this case). You can't find "a route" from the phylogeny unless we are talking

    The phylogeny and spread tells that N coalesced, probably, in SEA and that its foremother L3 did so in East Africa. All we can do in regard to that is to draw a dotted line between Asmara and Hanoi.

    ReplyDelete
  28. "Aren't you making the assumption here that each group of 'Archaic humans' was a separtae 'species'?"

    Not sure of what you may be talking about. I know that H. neanderthalensis and H. erectus were different species from H. sapiens: too different to be just the same thing. Just like donkeys and horses are different species too.

    "Bad news Maju. You can. It's called 'breeding up'. Start with a Gascon cow and cross with a Frisian bull. Cross the offspring again with Frisian bull, and so on. By the time the offspring are 15/16 Frisian (four generations) they're considered Frisain".

    Exactly my point. It's 95% Frisian and not Gascon anymore. In fact it stopped being Gascon at the first cross (pure hybrid) and IMO by the second cross it was already clearly in the Frisian side.

    What you are saying is that a Gascon cow became absorbed by a Frisian population. Exactly as a H. sapiens individual might be absorbed by a Neanderthal tribe or vice versa. This does not change the nature of the larger group: the community remains either Neanderthal or Sapiens or hybrid. You will never get a Neanderthal from crossing Sapiens with each other, the same you don't get a Frisian cow by crossing Gascons with each other.

    ReplyDelete
  29. "That's insultingly untrue. You are committed to your own pet hypothesis in a way that is clearly excessive, irrational".

    I'm committed to my own pet hypothesis because I have looked at the evidence, with no pre-conceived ideas.

    "False. There's no particular genetic reason to say anything between L3 and M/N. And certainly R is pointless in this story".

    You have consistently used R's presence in India to support a southern route for N.

    "What I have said often is that R 'unmistakably coalesced in South Asia', what now seems incorrect"

    So what haplogroup did N leave behind as it moved through South Asia?

    "I think that the southern route is correct because mtDNA M says so and archaeological evidence supports it too".

    Just because M moved via a southern route says nothing about any other mtDNA haplogroup.

    "There's a whole list of mtDNA lineages in the C3 area: D1 and CZ specially. I don't know why you chose A of all them".

    Because D and C definitely originated in India. They fit your southern migration theory.

    "For me A is like every other N: ultimately original from SEA".

    By the way. You've missed another expansion. Have a look at A in Phylotree, ignoring the blue mutations. After five clicks A expands instantly into A3, A4a, A4b, A4c, A4d, A2, A6, A9, A11, A5, A8 and A10. Some expansion. Twelve, I think. Nearly all in northeast Asia or America. Does that suggest a SE Asian origin?

    "it's not the only haplogroup in such circumstances: its foremother N is in the same situation for instance".

    True. And almost certainly developed from an N that remained in the region where A eventually expanded from. The same with the other N haplogroups.

    "I don't define a 'route' by the resulting haplogroups, what I define is a 'centroid', the likely origin of all the shared haplogroup (N in this case)".

    But a centroid is more likely to be a product of a large proportion of the population containg the haplogroup, rather than showing the region of origin. In fact you have criticised me in the past for suggesting centroids indicate origin. For example Y-hap D's centroid is in the hill country between Szechwan and Tibet. Y-hap C has two centroids: West Australia and in the northern bend of the Yellow River. Regions of origin for those haplogroups? I think 'perhaps', you think 'definitely not'.

    ReplyDelete
  30. "I know that H. neanderthalensis and H. erectus were different species from H. sapiens"

    The word 'species' is a human construct, sometimes with evidence, sometimes with emotion.

    "What you are saying is that a Gascon cow became absorbed by a Frisian population. Exactly as a H. sapiens individual might be absorbed by a Neanderthal tribe or vice versa".

    Exactly. In fact with just one more cross we get 31/32, just a little less than the mixture of H. sapiens and something else in Eurasian populations. And doesn't that rather suggest that modern humans and Neanderthals were not separate species?

    "You will never get a Neanderthal from crossing Sapiens with each other, the same you don't get a Frisian cow by crossing Gascons with each other".

    So? What's your point?

    ReplyDelete
  31. "I'm committed to my own pet hypothesis because I have looked at the evidence, with no pre-conceived ideas".

    You are not able to produce any evidence that supports your pet hypothesis, so it is not something cold, rational and unbiased: it's strongly emotional.

    "You have consistently used R's presence in India to support a southern route for N".

    No, I have not.

    "So what haplogroup did N leave behind as it moved through South Asia?"

    None. When pre-N migrated left no traces between the L3 core area and the N one.

    I do not see migrations the way you do: I only see population expansions when they happen. 'Silent' migrations of minor lineages leave no trace often between origin and 'destination'.

    "Because D and C definitely originated in India. They fit your southern migration theory".

    MtDNA D and C did not originate in India, their ancestor M did. D and C coalesced in NE Asia (or maybe SE Asia in the case of D?)

    "By the way. You've missed another expansion. Have a look at A in Phylotree, ignoring the blue mutations. After five clicks A expands instantly into A3, A4a, A4b, A4c, A4d, A2, A6, A9, A11, A5, A8 and A10".

    Ok. I take note of the larger size of this star. Just consider that A4c and A4d are dubious subclades by the CR count, as they are only defined by HVS (blue) markers (one each). Still it's 10 vs. the 6 or 8 I counted last year.

    "Does that suggest a SE Asian origin?"

    For pre-A it's N which defines the origin, not A. How do you know that all those clades are from NEA or America?

    "And almost certainly [A] developed from an N that remained in the region"...

    That's not a logical reasoning. It's as good as saying that "N developed from an L3 that remained in the region"...

    Try to apply the same logic to other lineages, please. You'd allow anything but would become unable to pinpoint any single origin, as every single haplogroup could have originated anywhere where it has a descendant. It's a sabotage against reason: M would have coalesced simultaneously in all Asia, R could be claimed to be European, L3 could have even coalesced in Asia or who knows... Wallacea?, Andaman islands? Everything becomes "possible" with your leftover lineages pseudo-reasoning.

    One needs to look at the whole picture: in the case of N to all derived basal lineages and not just one.

    ...

    ReplyDelete
  32. ...

    "But a centroid is more likely to be a product of a large proportion of the population containg the haplogroup"...

    That depends on how you estimate the centroid. I've seen what you say but I understand that if you reconstruct in an upstream direction: from lower level lineages upwards (weighting all lineages the same if they are at the same phylogenetic level, regardless of population), you get valid centroids that suggest origin.

    I'd even go a step further and I correct these "raw" centroids towards the upstream one in a second phase of the process (because of the implicit vector of the direction of the expansion) but this is just adjustments and the parameters of this correction (perfecting of the method) are by the moment uncertain for me.

    "In fact you have criticised me in the past for suggesting centroids indicate origin".

    I criticized the method used to estimate those centroids, based on mere population weight and not phylogeny. I am using the same word and may even have borrowed the essence of the concept (unsure) but it is a different kind of product, of "centroid" which I obtain. Mine ignores population figures and only consider clades by phylogenetic levels (each lineage weights one at its level).

    For example, for L3, M and N weight one each, exactly as L3a, L3b'f, etc. In this particular case I'd downplay M and N but I get a centroid in any case not far from the Red Sea. In the case of N, the several Australian lineages (which weight collectively 1/3) force it to be in the SE. It's the Australian (and SEA-specific) lineages, rather than A or N9 or R which do that mostly. If in the future it is discovered that they are not four but one, then things would change.

    But the overall population weights nothing in my "centroid" method.

    "Y-hap C has two centroids: West Australia and in the northern bend of the Yellow River".

    The midpoint of those two is the overall centroid (roughly) and falls near Guangzhou if I'm correct.

    "The word 'species' is a human construct, sometimes with evidence, sometimes with emotion".

    Re-phrasing that: I think that Neanderthal-Sapiens divergence is 0.9-1.3 million years old (what justifies the use of the term species). I follow Atapuerca's team on that and most literature on the matter I know of.

    "And doesn't that rather suggest that modern humans and Neanderthals were not separate species?"

    Bos taurus and Bos indicus can mix (not sure of how well) and yet they are often considered separate species. Same for wolf and coyote, etc. Your concept of species is old and rigid.

    "So? What's your point?"

    That there's no way that the Neanderthal population of, say, Altai evolved into the Sapiens population that replaced it. Same for West Asia, Europe...

    ReplyDelete
  33. "Bos taurus and Bos indicus can mix (not sure of how well) and yet they are often considered separate species".

    Very well. In fact it's quite common here to talk about a 'taurindicus' breed. And the humped cattle of Africa are a hybrid between indicus and a native African species.

    "Your concept of species is old and rigid".

    What is 'old and rigid' about it? Surely it is you who have the very ancient Jewish belief in the absolute separateness of individual species.

    "Try to apply the same logic to other lineages, please".

    I have always done so.

    "When pre-N migrated left no traces between the L3 core area and the N one".

    What? Not X? not N1'5? Not N2/W?

    "How do you know that all those clades are from NEA or America?"

    I think it was Wikipedia.

    "It's as good as saying that 'N developed from an L3 that remained in the region'..."

    Isn't that basically what happened? N derives from L3. Which had a huge expansion.

    "You'd allow anything but would become unable to pinpoint any single origin, as every single haplogroup could have originated anywhere where it has a descendant".

    Rubbish. It's almost always possible to eliminate many of the possibilities as being relatively recent regions for its arrival. And other methods are available to narrow the field even further.

    "M would have coalesced simultaneously in all Asia, R could be claimed to be European, L3 could have even coalesced in Asia or who knows... Wallacea?, Andaman islands?"

    No. See above. That's how we are fairly sure that L coalesced in Africa, M coalesced in India, and how I was able to persuade you that R originated in SE Asia (I don't think I've actually convinced you of Wallacea yet, but it will come).

    "One needs to look at the whole picture: in the case of N to all derived basal lineages and not just one".

    Which I do, which obviously confuses you.

    ReplyDelete
  34. "What? Not X? not N1'5? Not N2/W?"

    I understand so. Those haplogroups must have migrated from SE Asia (in stem stage).

    I understand that there was only one N expansion in only one site. Even if there were two or three short-lived migrant populations (instead of 12) between N and the derived haplogroups, this would be the same: the center of expansion was probably in SEA and one group would have migrated westwards, leaving a shallow trail (still detectable though in R, N5, N2...)

    "I think it was Wikipedia".

    It says almost nothing on this haplogroup. Only A2 is found in America per the only relevant paper cited. A4 (ancestor of A2) and A5 are related to Japan. There are other three clades at least.

    "Isn't that basically what happened? N derives from L3. Which had a huge expansion".

    But you do not claim that L3 coalesced in SE Asia because N appears to have been born there to posterity. You average considering all seven basal clades, not just one. Similarly A only tells 1/12 of the info re. N overall.

    "Rubbish".

    No so rubbish: your "method" is no method: it's pseudoscience of the kind you bend to whatever direction your emotions direct you to. It provides no answers but all possible answers to pick at convenience.

    Why should I consider mtDNA A above N1'5 or N14 or X or R or O? There's no reason at all but you pimp up A because it seems to favor your emotional choices, your preconceptions.

    ReplyDelete
  35. "Why should I consider mtDNA A above N1'5 or N14 or X or R or O?"

    Obviously you shouldn't consider A above any of the others. But neither should you ignore it simply because it doesn't fit a particular theory.

    "your 'method' is no method: it's pseudoscience of the kind you bend to whatever direction your emotions direct you to".

    And that statement is rubbish too. My method is to look at ALL the evidence, and draw conclusions that fit ALL the evidence. Anthropology, zoology, botany, genetics, geology, ecology, prehistory, archeology and paleontology. I don't just consider the evidence that supports a particular theory. Obviously each single piece of evidence on its own is open to a wide range of interpretation. And it's possible to come to almost any conclusion by selectively ignoring evidence. The conclusion I've come to is the result of considering a huge amount of evidence, but each single piece is open to several different interpretations. It's when all the evidence points towardfs a common conclusion that the conclusion is likely to be correct. The volume of evidence is much too great to compress to a comment on a blog.

    "I understand that there was only one N expansion in only one site"

    Certainly.

    "Those haplogroups must have migrated from SE Asia (in stem stage)"

    With R? Leaving no other N haplogroups in India? I don't think so. Why do you believe it is impossible that they moved in the other direction?

    "the center of expansion was probably in SEA and one group would have migrated westwards, leaving a shallow trail (still detectable though in R, N5, N2...)"

    It's rather difficult to make a case for N5 and N2 having originated in Se or S Asia, but I'm relieved you now accept R is SE Asian in origin. But don't you think it is rather surprising that N is the ONLY mtDNA haplogroup to have a disjointed distribution, unless you consider Central Asia as part of its original distribution.

    ReplyDelete
  36. "It says almost nothing on this haplogroup. Only A2 is found in America per the only relevant paper cited. A4 (ancestor of A2) and A5 are related to Japan. There are other three clades at least".

    It obviously wasn't Wiki then. All I could find there today on distribution is this:

    "Its subgroup A1 is found in northern and central Asia, while its subgroup A2 is found in Siberia and is also one of five mtDNA haplogroups found in the indigenous peoples of the Americas, the others being B, C, D, and X.[2]

    "Haplogroup A is the most common haplogroup among the Chukchis, Eskimos, Na-Denes, and many Amerind ethnic groups of North and Central America. 7.5% of the Japanese belong to haplogroup A (mostly A4 and A5).[3] 2% of Turkish people belong to haplogroup A. [4]"

    I got the distribution from a site which included the tree. I'll have another look.

    ReplyDelete
  37. I do not ignore A but it is just one dot out of 12 and is not even Central Asian nor clearly exclusive of NE Asia/America.

    "My method is to look at ALL the evidence, and draw conclusions that fit ALL the evidence (...) I don't just consider the evidence that supports a particular theory".

    I think you do: you are very stubborn about models that have no or extremely weak basis. So you choose first the model and then try to bend the evidence to it.

    You are doing that with A, which is like the most brilliant star by far in your "constellation N". But it's just another star: that's the objective reality.

    "Obviously each single piece of evidence on its own is open to a wide range of interpretation".

    That's why you must look at all the evidence and see where it converges and where does not.

    "With R?"

    Maybe. It's possible indeed. Of all Western N, only N1'5 and R are one CR mutation away from N-root. All the rest seem more recent (in their respective expansions-consolidations): 4-5 mutations away, almost as "young" as A.

    If R was "first daughter" (this is conjectural but plausible), at least within this Western group, it would explain somewhat why we see so little N(xR) in South Asia.

    This model would assume a rapid migration from SEA to WEA, in few millennia (specially to allow the expansion of N1'5 happen in WEA - if that is the case after all).

    Whatever the case, N2, A and X experienced relatively late expansions with long latency periods. The same can be said surely of N13, N14, N21, N22, whose "latency" lasts to present time AFAIK. Instead N1'5, N9, O, S and R show short latencies (1-2 mutations). The difference can mean 20 thousand years maybe.

    "Why do you believe it is impossible that they [mtDNA N] moved in the other direction?"

    Because the number of basal lineages (diversity) is highly concentrated around SEA (3/4). So SEA is the only center of expansion allowed by Occam's fatal razor. :/

    My hunch was around Thailand but, after moving N1'5 westwards (???) it may be more like Burma.

    "It's rather difficult to make a case for N5 and N2 having originated in Se or S Asia"...

    South Asia if anything (not SEA): at least they are found there. I cannot get to details I do not know well and it would not make much of a difference if they coalesced in Iran, for instance.

    "But don't you think it is rather surprising that N is the ONLY mtDNA haplogroup to have a disjointed distribution".

    Yes, of course. But this can be largely explained by the huge early success of M in South Asia. Still I am not so persuaded yet that N5 and N2a are not of South Asian origin. The evidence is unclear.

    But, as I said, this is not so important. If anything, it would allow an equally possible (and equally unproven) northern route in an East-West direction, which is counter-producing for your hypothesis of boaters' backmigration (the boaters of the steppes, hehe).

    "... I'm relieved you now accept R is SE Asian in origin".

    It only weights even more for an Eastern origin of N. It's counter-producing for your hypothesis also.

    "... unless you consider Central Asia as part of its original distribution".

    I see no N diversity nor patterns claiming "Central Asia" anywhere. Most N basal subclades are very clearly southernly.

    ReplyDelete
  38. "So you choose first the model and then try to bend the evidence to it".

    You have a short memeory. Who was it had so much difficulty convincing you that MNOPS* did not originate in South Asia, but somewhere in SE Asia? Who is it had so much difficulty convincing you that mtDNA R did not originate in South Asia, but somewhere in SE Asia? I also remember you conceding that I had persuaded you of something concerning Madagascan prehistory (I forget now what exactly it was). Are you really still convinced that I have been correct in those matters in spite of seeing no evidence at all, just luck? There has always been enough evidence for those prepared to look. It's just that many were led astray by 'the great southern coastal migration theory'.

    "Yes, of course. But this can be largely explained by the huge early success of M in South Asia".

    So how come either N or R managed to pass through South Asia? There must be some explanation for it. Of course we can easily see R's trail through South Asia (in a westerly direction), but how could N have moved firstly east through South Asia, then back west without leaving any trail whatsoever?

    "South Asia if anything (not SEA)"

    I agree in part. Certainly not SE Asia. But that's exactly what you're in effect proposing. Haplogroups N and R moving together from that region.

    "My hunch was around Thailand but, after moving N1'5 westwards (???) it may be more like Burma".

    N1'5's origin has to be moved way further west than just as far as Burma.

    "Still I am not so persuaded yet that N5 and N2a are not of South Asian origin. The evidence is unclear".

    The evidence is fairly clear to me. Certainly N5 is Indian, but it has a five mutation latency period before it expands there. Its nearest relation N1 is widespread in West Asia and diversifies after just two mutations. Just a few downstream mutations are found in South Asia. So it is extremely unlikely that N1'5 originated in South Asia. Much more likely that it came in from the west. And N2a is a downstream mutation from N2. And W too is widespread in West Asia. N2 has a 4 mutation latency period, and W and N2a have a 7 mutation latency period though.

    "it would not make much of a difference if they coalesced in Iran, for instance"

    I agree with the importance of Iran, but I certainly don't agree that 'it would not make much of a difference'. It makes a huge amount of difference.

    "It only weights even more for an Eastern origin of N. It's counter-producing for your hypothesis also".

    Not counterproductive for my hypothesis, but certainly so for yours.

    ReplyDelete
  39. "You have a short memeory. Who was it had so much difficulty convincing you that MNOPS* did not originate in South Asia, but somewhere in SE Asia?"

    It must be that I have really short memory because I don't remember any moment since the discovery of MNOPS, when I did not think it was a SE Asian clade. I was even into that idea when it was still only known as NOP and I had even considered years ago (before inaugurating Leherensuge) if K as a whole might have orignated over there.

    The discussions with you about MNOPS were often about terminology (MNOPS or KMNOPS, or whether MNOPS included some K*) and maybe about Wallacea-wonderland.

    In fact I am one of the few people that has been arguing for a role in the Eurasian colonization, even if secondary or diffuse, for SEA in the last many years.

    I must have short memory because what you imply I said makes no sense to me.

    "Who is it had so much difficulty convincing you that mtDNA R did not originate in South Asia, but somewhere in SE Asia?"

    If you had posted the data (when it came forth, which is very recent), It would have been easier to understand. Data speaks volumes, discourses do not.

    It's not a matter of being correct or not. I could adventure that there is a planet X out there and maybe in some years I'll be vindicated. But right now I have no evidence in favor of its existence, so I remain cautious. Instead you adventure things much more extravagant without any evidence, just because you feel like. With time you may happen to be right in 5% or so of your proposals but that's not different than chance.

    And if you cared about making sure you have the data to back your hypothesis or, otherwise, dropping them (or just filing them in the corner of the highly speculative stuff), then we'd have less pointless discussions here.

    Because after all it's data what matters. Evidence is what makes the difference between something real or a fairy tale.

    "There has always been enough evidence for those prepared to look".

    FALSE. There was no evidence earlier. If I have been defending that R is a clear case of SA coalescence it is because the evidence pointed that way and no other reason at all.

    The evidence accounted for many more R basal sublineages in SA than in all Eastern Eurasia together.

    Now we know better? Fine. But it's something new, something that you could not know a few years or even months ago and something that, if you knew, you should have documented.

    But for you it's easier it seems to build a theory on the highly conjectural existence of boats now but not then, here but no there. Something for what you lack the slightest evidence whatsoever.

    And all our debates are around that hypothesis you drew from your magic hat out of nothing. It would have merit if it would be a fluffy rabbit but hypothesis are just words and concepts, unless supported by evidence, unless built on evidence.

    ...

    ReplyDelete
  40. ...

    "It's just that many were led astray by 'the great southern coastal migration theory'".

    It has nothing to do, the coastal model is very much self-evident. When you find that R effectively appears to have expanded from SEA, what you are in fact finding is how N and R expanded in SEA, within the M expansion, within the overall Eurasian expansion.

    The problem is precisely why haplogroups like N (and by extension R) coalesced in SEA without leaving earlier traces farther west (in India or Altai, I don't care). And the only half-decent answer experts have been able to muster is that the migration between Africa and SEA was quite fast, what could well been have aided by the existence of a population which was more specialized in coastal foraging and therefore primitive navigation.

    This also explains why people seem to have crossed the Wallacean straits so fast, because we see very old basal sublineages popping out in Melanesia (and nowhere else). If people would not have invented boats until arriving to SEA, then they would probably have taken a lot more time to cross to Australasia. They did very fast, so they had the know-how before, what takes no merit from the feat anyhow, just explains why.

    It's the same as with Columbus or Vasco de Gama or Zheng He or even Red Eirik: they did what they did because oceanic navigation had taken shape before them, they did not invent the ships they sailed on, not at all, they used technology that was older than them. Otherwise you would not be living in New Zealand, really.

    ReplyDelete
  41. "So how come either N or R managed to pass through South Asia?"

    LOL: coastal migration! Not N nor R but pre-N (L3n, so to say).

    It seems to me that you have not understood yet that it is SE Asian lineages precisely the ones which claim for the rapid coastal migration model. Only such an alternative, much thinner but also quite faster, route seems to explain the Eastern macro-lineages.

    "But that's exactly what you're in effect proposing. Haplogroups N and R moving together from that region".

    Yes, if R is oriental too, then N and R probably back-migrated together westwards.

    "N1'5's origin has to be moved way further west than just as far as Burma".

    Please try to understand what I said:

    "My hunch was around Thailand but, after moving N1'5 westwards (???) it may be more like Burma".

    This sentence refers about the origins of N as a whole, not N1'5. If you move the origin of N1'5 westward, then the origin of N must also do proportionally (1/12).

    "Certainly N5 is Indian, but it has a five mutation latency period before it expands there".

    Whatever. It means that 50% of basal lineages under N1'5 are South Asian. If there's no evidence of N5 in WEA, then the overall origin of N1'5 is between WEA and SA (a bit uncertain because we only have two branches, not 12). BUT, as the origin of N overall is in SEA, that may serve to calibrate the overall origin of N1'5 in SA.

    That N5 is small is exactly what we would expect if N sublineages had very limited room for expansion in the SA passage. And that is precisely what we see, right?

    "And N2a is a downstream mutation from N2. And W too is widespread in West Asia".

    But if N2a is exclusive of SA it's the same case as with N1'5 or any other such bifurcating haplogroup.

    For me it's likely (unless other evidence comes forward) that N2 and N1'5 coalesced more likely in SA than in WEA. However it also looks like they were very small, private, lineages in that time and only found real room for expansion in the WEA colonization. It's the same case as with R (after excluding Eastern R).

    And it's probably the case with pre-N before its expansion in SEA: it was small (small enough to go extinct by mere drift) but eventually found a niche for expansion. Other such lineages surely never made it all instead (though we can't know anymore).

    "I agree with the importance of Iran, but I certainly don't agree that 'it would not make much of a difference'. It makes a huge amount of difference".

    I just do not see the "big" difference. One must have obsessed by some sort of anti-Indian prejudice in order to think that there's a big difference, really. Iran is even listed as part of South Asia in some versions of the definition.

    ReplyDelete
  42. "It must be that I have really short memory because I don't remember any moment since the discovery of MNOPS, when I did not think it was a SE Asian clade".

    Short memory alright. You argued against me for ages. I agree that I did use 'K' because the haplogroup was incompletely known at that stage.

    "But it's something new, something that you could not know a few years or even months ago and something that, if you knew, you should have documented".


    It was not possible to 'know', but it was certainly possible to see that it was 'very likely'. It was just more complicated to explain.

    "If I have been defending that R is a clear case of SA coalescence it is because the evidence pointed that way and no other reason at all".

    No. It was because you were convinced that it must have coalesced in S Asia. You were not prepared to consider any alternative, no matter how briefly.

    "And the only half-decent answer experts have been able to muster is that the migration between Africa and SEA was quite fast, what could well been have aided by the existence of a population which was more specialized in coastal foraging and therefore primitive navigation".

    And what a totally inadequate explanation that is!

    "the coastal model is very much self-evident".

    It is self-evidently rubbish to me. But, of course, it's complicated to explain why it's rubbish. I'll try when I have more time.

    "This also explains why people seem to have crossed the Wallacean straits so fast, because we see very old basal sublineages popping out in Melanesia (and nowhere else)".

    Excuse me. Haven't yoy forgotten about the West Asian basal lineages? And I'd hardly describe the crossing of Wallacea as 'rapid'. How long had humans been in South Asia before then?

    "as the origin of N overall is in SEA"

    Aren't you making a huge assumption there?

    "That N5 is small is exactly what we would expect if N sublineages had very limited room for expansion in the SA passage. And that is precisely what we see, right?"

    We don't see a 'very limited room for expansion in the SA passage'. We see absolutely no expansion across India, only in the west.

    "It's the same case as with R"

    Isn't there quite a huge expansion of R in India? Isn't that why you considered that region to be the origin of R for so long?

    ReplyDelete
  43. "I agree that I did use 'K' because the haplogroup was incompletely known at that stage".

    And my memory is still good enough to recall that I said that I suspected that Oceanian Y-DNA K probably would be eventually gathered in a single haplogroup, that I had considered an expansion of K from Melanesia but that I found it too inconsistent with the rest of the genetic data for Eurasia.

    I was vindicated when MNOPS arose (and some day probably M, S and some of the Oceanian K-other will be discovered to be a subclade of MNOPS, the same that NO is). In Y-DNA we can't have (or rather we never actually have) the certainty we have with mtDNA, so even if I happen to be wrong, I'm right to doubt about such details of the phylogeny.

    Ren, Manju and Ibra stand witness that I once adventured, long before I met you online, before I even started blogging at all, that K might have expanded from Melanesia (it must be filed somewhere in Ren's forum Quetzalcoatl). Of course, I was wrong but I had been there before you came with that idea, so I could evaluate it from BOTH sides.

    "No. It was because you were convinced"...

    And I was convinced for a reason, not out of thin air.

    "You were not prepared to consider any alternative"...

    Actually I was. I have just done it. All I need is evidence.

    So when you find evidence of MP boating or lack of it anywhere, when you find evidence suggesting a radiation of mtDNA N in Altai... tell me. So far there is nothing of all that, so wasting so much keyboard on it is not really worth the keystrokes.

    "Haven't yoy forgotten about the West Asian basal lineages?"

    Not sure what's the point but the WEA basal lineages (under M and N) are two: M1 and X.

    "And I'd hardly describe the crossing of Wallacea as 'rapid'. How long had humans been in South Asia before then?"

    Not long: there is only one CR mutation between M and M29'Q, an Oceania-only lineage. I estimate something between 2.5 and 5 Ka, stretching it to the limit not more than 10-15 Ka (minimum is one generation).

    So it was really fast indeed.

    "We see absolutely no expansion across India, only in the west".

    That's not really true. There's a lot of evidence for lineages involved in that back-migration existing in Eastern India and Bangla-Desh. In this I have to point to Y-DNA: to the existence of abundant P* and also the diversity core of R2 in that area of the Ganges Delta (something about Q too but can't recall). The geography of small mtDNA lineages in SA is not so well known, at least to me, so I have here to resort to Y-DNA.

    If you are so kind anyhow of providing the references for the exact location of all N sublineages in South Asia, I'd appreciate it.

    "Isn't there quite a huge expansion of R in India? Isn't that why you considered that region to be the origin of R for so long?"

    Number of basal lineages, diversity, variance... that's why.

    ReplyDelete
  44. Update on the debate R had a South Asian coalescence almost for sure, even if this one happened (maybe) closer to Bengal than depicted in the map.

    Why? Because as I just wrote elsewhere:

    West and South Asian basal sub-lineages: 10 (of which 6.5 are SA and 3.5 WEA). They are: R1, R2'JT, R3, R5, R6'7, R8, R30, R31 and U.

    East Asian and Oceanian basal sub-lineages: 7, of which 4 are East and SE Asian and 3 are from Near Oceania. They are: R9 (incl. F), R11'B, R12'21, R14, R22, R23 and P.

    So not just this side of Movius Line has the largest basal diversity but South Asia has the largest one of all four regions. It's pretty clear, I think.

    However the center of this scatter may well have been at or near Bengal (Orissa?)

    ReplyDelete

Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (sorry, too many trolls).