January 20, 2012

Out of Africa migration was coincident in time with an internal expansion in Africa of mtDNA L3

Ethio Helix mentions today a rather interesting study (potentially, as it is behind a paywall) on the expansion of mtDNA L3 in Africa and its relation with the migration Out of Africa of L3-derived lineages M and N.


Pedro Soares et al., The expansion of mtDNA haplogroup L3 within and out of Africa. Molecular Biology and Evolution, 2011. Pay per view.
Abstract
Although fossil remains show that anatomically modern humans dispersed out of Africa into the Near East ∼100–130 ka, genetic evidence from extant populations has suggested that non-Africans descend primarily from a single successful later migration. Within the human mtDNA tree, haplogroup L3 encompasses not only many sub-Saharan Africans but also all ancient non-African lineages, and its age therefore provides an upper bound for the dispersal out of Africa. An analysis of 369 complete African L3 sequences places this maximum at ∼70 ka, virtually ruling out a successful exit before 74 ka, the date of the Toba volcanic super-eruption in Sumatra. The similarity of the age of L3 to its two non-African daughter haplogroups, M and N, suggests that the same process was likely responsible for both the L3 expansion in Eastern Africa and the dispersal of a small group of modern humans out of Africa to settle the rest of the world. The timing of the expansion of L3 suggests a link to improved climatic conditions after ∼70 ka in Eastern and Central Africa, rather than to symbolically mediated behavior, which evidently arose considerably earlier. The L3 mtDNA pool within Africa suggests a migration from Eastern Africa to Central Africa ∼60–35 ka, and major migrations in the immediate postglacial, again linked to climate. The largest population size increase seen in the L3 data is 3–4 ka in Central Africa, corresponding to Bantu expansions, leading diverse L3 lineages to spread into Eastern and Southern Africa in the last 3–2 ka.

As always, take the age estimates with lots of spices: they are just educated guesses which will vary wildly depending on the assumptions made a priori, like the usually under-estimated time for the divergence of the Pan-Homo genera, etc.

But otherwise the paper appears very much coincident with what Behar and myself (based on his data) could have inferred, including a NW African specificity of L3k, it seems.

Frequency maps by subhaplogroup:

A: L3a, L3i, L3h, and L3x combined, B: L3f, C: L3e, D: L3b, E: L3d


Their reconstruction of the spread of L3:



For many more details, you should visit Ethio Helix' original entry: The Mother of Mothers!

PS- As contrast to Ethio Helix' "too white" (in my opinion) selection of Ethiopian portraits, I have located some real women from non-Semitic Ethiopia, which may give a better idea of how the women of the L3 clan looked originally:

Hamer girl (from the SNNPR):



Gumuz mother (from the Benishangul-Gumuz region):



Nuer woman with their characteristic scarifications (from the Gambela region and nearby South Sudan):


Oromo woman (from Oromia, the largest region of Ethiopia):

80 comments:

  1. I am beginning to sense some sort of a romantic fixation that you have with 'South Western' Ethiopia Maju. I clearly and explicitly said that those sketches, and I emphasize sketches instead of the random pictures that you took the liberty to post, represented women that inhabit the area of the putative origin of L3 as outlined in the subject of the paper, that does off course not mean that the original L3 bearer exactly looked like them, since nobody really knows what they looked like.

    The Area outlined in this paper for L3's origin is approximately bounded in the North by Khartoum, North Sudan, in the East by Awash, Ethiopia, in the South by around the middle of Lake Turkana, Kenya and in the West by Agok, Abyei South Sudan. This is a large area with numerous types of people living in it, but mostly just Nilotes , that is Nhilo Saharan speakers and Afroasiatics. You can not simply just say these people are 'white' looking or 'black' looking as likely all 'black' and 'white' looking people originated in this region, there is nothing surprising about finding people who have a 'white' morphology in this area. As for your back migration hypothesis as to why the pictures I selected are unjustified, you first have to prove what the people who back migrated looked like, you also have to quantify what physical divergence these people had to the people of the area they back migrated to, and remember that it is not called back migration for no reason, those people went back to whence they came from, so in essence they were the same people.

    In addition, before you nonchalantly throw around terms like 'Oromia', and 'Semitics', the woman in the middle of my selected sketches is an oromo lady, have you any clue as to the numerous type of people that live in these regions? They are not monolithic, have been cross pollinating with each other for millenia upon millenia, and the languages they speak bear little to no witness to the genesis of their physical types. Please cease with your tourist-like romantic fixation of certain African peoples that you pick and choose at whim.
    Now as to the paper, I'm glad you found it interesting, to me also it was one of the best genetic papers done on Africa that I have read in quite some time, besides Tishkoff and some of the Cruciani papers. It was well sourced, with tones of references and a great read.

    ReplyDelete
  2. Very interesting paper. I am not sure that the people that left Africa look like anybody currently living, for example, early modern skulls have rectangular eye sockets,unlike anybody living today.

    Personally, I think that we are just beginning to find out about the diversity of our own species, as we further explore the genome. There is no reason to think that as early moderns moved into new areas that these areas were uninhabited.

    ReplyDelete
  3. The mtDNA data need error bars and track records of accurate calibration to be accurate as absolute dates (and in particular, in their ability to distinguish pre-Toba v. post-Toba data when the archaeology is arguing for older range dates), but I agree with you that the relative dating that mtDNA mutation rate dating can provide is more informative and that this therefore does support one big L3 expansion that included Eurasian branches M and N, and that the date of L3 expansion is in the general vicinity of the Out of Africa main wave.

    The internal structure of L3 expansion elicited, absolute date uncertainty aside, also does seem to have some merit to it.

    The notion of one big L3 expansion for the entire clade including M and N also disfavors an "Out of Arabia" scenario in my mind.

    ReplyDelete
  4. "Frequency maps by subhaplogroup"

    Your map C and map E both list L3e. One is obviously wrong. Your map A combines subclades from within several haplogroups. Why? Haplogroups L3i and L3x are subclades of L3e'i'k'x. Haplogroup L3e appears on its own in map C. L3a and L3h are separate major haplogroups. Map B and Map D are of the two haplogroups within L3b'f. L3b looks to have been split by the exapnsion of L3f.

    I saw a reference to this at Phylotree yesterday while looking for something else. The only change made to the tree seems to be the appearance of an L3a1, with a 5 mutation stem. Wonder where it is found. Possibly here:

    "I am beginning to sense some sort of a romantic fixation that you have with 'South Western' Ethiopia Maju".

    From the data Etyopis supplied I suppose Maju should actually now refer to 'West Ethiopia/South Sudan', although that is merely the centre of the expansion.

    "this therefore does support one big L3 expansion that included Eurasian branches M and N, and that the date of L3 expansion is in the general vicinity of the Out of Africa main wave".

    Yes, and the abstract contains the interesting information:

    "The similarity of the age of L3 to its two non-African daughter haplogroups, M and N, suggests that the same process was likely responsible for both the L3 expansion in Eastern Africa and the dispersal of a small group of modern humans out of Africa to settle the rest of the world. The timing of the expansion of L3 suggests a link to improved climatic conditions after ∼70 ka in Eastern and Central Africa, rather than to symbolically mediated behavior, which evidently arose considerably earlier".

    Presumably the 'improved climatic conditions after ∼70 ka' applied beyong Africa as well.

    ReplyDelete
  5. @ terryt, Map C. is for L3e and Map E. is for L3d, it was just a simple typo when Maju re-entered the labels, but as far as the map itself he did not create it, it is part of the paper. The Authors combined L3a,L3i,L3h and L3x, because all of them (especially L3a and L3h) were found strictly in East Africa, where they were all implied to have their origin. While L3e (Map C) is implied to have an origin in Central Africa. And L3f (Map B), is implied by the Authors to have migrated to the sahel and central Africa in the holocene of the Sahara. L3b(Map D) and L3d (Map E) were postulated to start their diversification in Central/west Africa.

    ReplyDelete
  6. "Map C. is for L3e and Map E. is for L3d, it was just a simple typo when Maju re-entered the labels"

    Sure, and corrected, thanks.

    "... as far as the map itself he did not create it, it is part of the paper".

    Of course and I never said otherwise. In fact the whole entry, including our aesthetic disagreement, is based on what Etyopis wrote at his blog (because I do not have access to the paper myself).

    ReplyDelete
  7. @Etyopis:

    I have already discussed the aesthetic discrepancy at your blog, so I won't be redundant here, just trying to incite some debate, as I really could not find your selection of faces too credible.

    Yet, just to be clear:

    "You can not simply just say these people are 'white' looking or 'black' looking as likely all 'black' and 'white' looking people originated in this region"...

    But some populations in that area have more back-flow genetics from West Asia, that's undeniable, and I do not have any reason to think that West Asians are more repesentative of the massive Eurasian-plus diversity than, say, Melanesians, Japanese or Bengalis. In addition the trait of depigmentation "white skin" in the broadest possible sense was almost without doubt evolved after the migration from Africa, as people needed to raise children in climates with quite low solar radiation: skin color it is one of the very few clearly adaptive traits, so it was for sure not present before th OoA, although it's possible that lighter variants were freer to coalesce in subtropical climate regions like Southern or Northern Africa.

    "The Area outlined in this paper for L3's origin is approximately bounded in the North by Khartoum, North Sudan, in the East by Awash, Ethiopia, in the South by around the middle of Lake Turkana, Kenya and in the West by Agok, Abyei South Sudan".

    I like that description: it's wide but very reasonable and fits quite well with my notion of "Upper Nile" although for the White Nile it is more like "Middle" in fact.

    ReplyDelete
  8. @Andrew: "but I agree with you that the relative dating that mtDNA mutation rate dating can provide is more informative"

    With me? I did not say that. In fact I said that I did not like the 70 Ka estimate because archeology disproves it (must be older).

    ReplyDelete
  9. Btw, nobody has seemingly noticed but the two last women in my little sample show quite apparent sinodonty, a dental trait typical of NE Asia (opposite is sundandonty, typical of SE Asia and most of the World in fact) characterized by having shovel-shaped frontal teeth (and other traits that we can't so easily discern visually). The Nuer woman actually has a quite marked epicantic fold and to me she does suggest certain "proto-Mongoloid" traits overall (however her red hair and generally reddish shade of brown rather hint towards West Eurasian typologies - might be back-migrating genes but may also be ancestral).

    ReplyDelete
  10. "Map C. is for L3e and Map E. is for L3d, it was just a simple typo when Maju re-entered the labels, but as far as the map itself he did not create it, it is part of the paper".

    Thanks for that explanation.

    "The Authors combined L3a,L3i,L3h and L3x, because all of them (especially L3a and L3h) were found strictly in East Africa, where they were all implied to have their origin".

    I have no problem with L3a and L3h being 'Ethiopian'. In fact I'd worked that out from Maju's earlier blog regarding haplogroup L. But L3i and L3x are both subgroups within L3e'i'k'x. Just two of these can be described as 'Ethiopian'. Of the other two you mention that 'L3e (Map C) is implied to have an origin in Central Africa' and Maju originally mentioned 'a NW African specificity of L3k, it seems'. So it looks as though L3e'i'k'x had a wide distribution before it broke up. Was that distribution centred on Ethiopia or are L3i and L3x immigrants from further west? Haplogroups L3b'f and L3c'd'j certainly appear to be centred further west.

    ReplyDelete
  11. L3i and L3x are also found in Arabia peninsula, an area ignored in these maps.

    L3i does look Ethiopian very much, with offshoots in Oman, Yemen and Sudan.

    L3x appears split 50-50 between Ethiopia and Arabia Peninsula:
    ·L3x1 [Ethiopia, Yemen]
    ·L3x2a [Ethiopia, Arabia]

    But then: L3x2b [Algeria], this one is a very long-stemmed lineage however, so I would not consider it too much (as it's seed lineage could well have been wandering in the Sahara, so to say, for many many millennia before coalescing into something worth a name). If we don't consider it, the centroid falls in Eritrea or so, if we do, then in Sudan, rather to the South.

    ReplyDelete
  12. "L3i and L3x are also found in Arabia peninsula, an area ignored in these maps".

    True. But the same subclades of both haplogroups are found in Ethiopia as well, which suggests either that the haplogroups have yet to be fully resolved or that the arrival in the Arabian Peninsula is not associated with the 'original' OoA. No subclades of either haplogroup are exclusive to the Arabian Peninsula.

    "L3i does look Ethiopian very much, with offshoots in Oman, Yemen and Sudan".

    Yes.

    "L3x appears split 50-50 between Ethiopia and Arabia Peninsula:
    ·L3x1 [Ethiopia, Yemen]
    ·L3x2a [Ethiopia, Arabia]"

    With both subclades being found in Ethiopia it is much more likely that the haplogroup had split before separately leaving Ethiopia rather than that the two haplogroups had moved separately into Ethiopia from across the Red Sea.

    "But then: L3x2b [Algeria], this one is a very long-stemmed lineage however, so I would not consider it too much (as it's seed lineage could well have been wandering in the Sahara, so to say, for many many millennia before coalescing into something worth a name)".

    But, as I continue to to try to point out, a long stem is far more likely to indicate a long period spent in isolation rather than a long period spent on the move. The presence of L3x2b in Algeria is far more likely to show that L3x coalesced in the Sahara.

    ReplyDelete
  13. There are many known West Asian lineages which have back-migrated to Ethiopia, surely long ago. Some 25-30% of mainstream Ethiopians' ancestry (Y-DNA and mtDNA). This makes it more difficult to discern where the lineage coalesced if equally split between Arabia Pen. and The Horn and rare enough not to have been studied in depth. A typical case is L6, whose highest basal diversity may be in fact in Yemen (but is rare enough to leave us all wondering and scratching our heads). It can be the same with L3x (or not).

    "as I continue to to try to point out, a long stem is far more likely to indicate"...

    You're wrong! Write a blog and "demonstrate" (if you can) that in your blog. Not in my comments section! If your demonstration is good (that it won't) then I'll grant you the right to claim that. But it won't happen.

    "The presence of L3x2b in Algeria is far more likely to show that L3x coalesced in the Sahara".

    This is the kind of idiocies that you end up spitting around when your basics are so extremely wrong. Quit it, ok? It's annoying and pointless and, most importantly, radically false.

    ReplyDelete
  14. The supplemental data in this paper actually found more L6 cases in East Africa. 2/77 Cases of L6 in Ethiopians. 3/148 Cases of L6 in Somalis, and an additional 2/77 Cases of L6a in Ethiopians. I doubt L6 originated on the non-African side of the Red sea.

    ReplyDelete
  15. Even the supp. material is behind paywall, so I must trust your word acritically. However it's not the number but the phylogenetic diversity what can give us clues about origin. I remember that there was somewhat more of that basal diversity for L6 in Yemen but it is very hard to decide.

    "I doubt L6 originated on the non-African side of the Red sea".

    Not impossible at all. There was surely some L(xM,N) flow into Arabia around the time of the OoA. It's not possible that so much L0 (and rare clades within it) is found for example unless it is a very old founder effect. Clades that bridge the Red Sea like L6 or L3x are less clear but clades like L0d3, shared by Khoisan and Kuwaitis, or L4b, split between Yemen (L4b1) and East Africa (L4b2)... speak, I understand of a time, long ago (OoA period surely), when Africa overflowed into Arabia.

    These clades that bridge the two shores of the Red Sea in such a balanced manner can well have coalesced in Arabia and backmigrated with the other West Eurasian lineages. Or not... only detailed discernment of each clade will give us some answers. It's not a matter of belief.

    ReplyDelete
  16. Maju, check your email @ lialdamizDELETETHIS@gmail.com.

    Could you also explain to me how the 'basal diversity' of the L6 found in yemen is more indicative of its origin there than on the other side of the red sea?

    ReplyDelete
  17. You have to delete the "DELETETHIS" spam protection. Otherwise it will never arrive anywhere.

    "Could you also explain to me how the 'basal diversity' of the L6 found in yemen is more indicative of its origin there than on the other side of the red sea?"

    Going to search for the relevant paper because I have it a bit forgotten. There's a haplotype NJ tree that illustrates the matter and rather supports a Yemeni origin as far as I recall.

    ReplyDelete
  18. O, how silly of me, ok check again!

    ReplyDelete
  19. First I have found is that Behar 2008 data located L6a in Yemen and Egypt (see Fig. S1), and L6b in Yemen and Ethiopia. Unless there is data that contradicts this, this should be enough to support in principle, L6 as a whole being original from Yemen, as both lineages are found there and nowhere else.

    To be precise the sequences in this study were:

    L6a (branch 1): Egypt (n=1)
    L6a (branch 2): Yemen (n=3)
    L6b (branch 1): Ethiopia-Jewish (n=1)
    L6b (branch 2): Yemen-Jewish (n=1)
    L6b (branch 3): Ethiopia (n=1)

    I'll see if I can find something else because I remember an NJ tree from a discussion years ago, also somewhat supportive of a Yemeni origin. But this alone, in absence of any contradictory data, is strongly suggestive of a Yemeni origin. L6b could still be Ethiopian by origin or not (too Jewish, right? - it can also be sample bias, let's not forget Behar is Israeli).

    ReplyDelete
  20. "You're wrong! Write a blog and 'demonstrate' (if you can) that in your blog".

    Surely I don't have to write a blog to demonstrate the fact. I've already pointed out that the vast majority of haplogroups with a long stem almost certainly developed that stem in situ. The very small minority of haplogroups where that is not obviously the case are actually quite easily imagined to fit that most common scenario. It is only your blind commitment to migrating stems in haplogroups M, N, N2 and A that prevents you from seeing the blindingly obvious.

    "This is the kind of idiocies that you end up spitting around when your basics are so extremely wrong. Quit it, ok? It's annoying and pointless and, most importantly, radically false".

    You are obviously wrong on this point.

    "Not impossible at all. There was surely some L(xM,N) flow into Arabia around the time of the OoA".

    I agree that it's not impossible. But even your own diagram shows L6 did not diversify until the 37 mutation level. I agree that the molecuklar clock may be broken but surely such a long stem before diversification argues against an emergence from Africa as early as the OoA.

    "It's not possible that so much L0 (and rare clades within it) is found for example unless it is a very old founder effect".

    But not one of those L0 clades is specifically ex-Africa. All have close relations, if not actual members, within Africa. Their arrival outside Africa could easily be as late as the Neolithic.

    "clades like L0d3, shared by Khoisan and Kuwaitis, or L4b, split between Yemen (L4b1) and East Africa (L4b2)... speak, I understand of a time, long ago (OoA period surely), when Africa overflowed into Arabia".

    You may be onto something with this haplogroup, especially as L4b1 has an especially long stem.
    And L4 formed almost simultaneously with L3.

    ReplyDelete
  21. "Surely I don't have to write a blog to demonstrate the fact".

    You do have to do that "demonstration" (if you can) in a distinct space than this blog's comments section, which you are once and again hijacking and clogging with your dead horse.

    So do yourself and the World a favor and write something more or less consistent that WE can debate at YOUR comments section. Thanks.

    ReplyDelete
  22. I must say that the data of this study MAY be supportive of Ethiopian origin for L6. However as they used only HVS haplotypes it's a bit confusing at times, specially in a ill-understood clade as this one.

    Thanks to Etyopis for the data and the conversation.

    ReplyDelete
  23. @ Maju

    the authors do not take into account the Eurasian incursion into Africa and specifically North Africa, that clusters today genetically with West Eurasia rather than the rest of the continent..

    take a look at this Kabyle sample :

    E1b1b1b (E-M81) (47.36%), R1*(xR1a) (15.78%) (later tested as R1b3/R-M269 (now R1b1b2)[9]), J1 (15.78%), F*(xH, I,J2,K) ( 10.52% ) and E1b1b1c (E-M123) (10.52%).

    MtDNA Haplogroups, : H (32.23%), U* (29.03% with 17.74% U6), preHV (3.23%), preV (4.84%), V (4.84%), T* (3.23%), J* (3.23%), L1 (3.23%), L3e (4.84%), X (3.23%), M1 (3.23%), N (1.61%) and R (3.23%).

    what is the percentage of the ancient Aterian lineages that survived the Eurasian replacement ?
    1,5 or 3 % ?
    did the Authors ask themselves this important question ..
    what was North Africa like prior to the back to Africa migration ?
    so,what happened in North Africa in the last 45 thousand years ?
    it s obvious , 90 to 98 % of the ancestral Aterian lineages went extinct !(depends on the regions)..
    this study's conclusion is speculative at best .because it was based on the haplogroupial distribution TODAY ..

    ReplyDelete
  24. What exactly you disagree with the authors re. North Africa?

    They don't even say that there is any Aterian lineage, because for them L3k is as young as 35 Ka (???) and arrived with U6. That L3k may be Aterian is my call but I may be wrong.

    Whatever the case I find difficult to use a Kabyle sample to discern this matter: look in South Morocco because the autosomal signature that looks Aterian is found only there (at least at meaningful levels).

    I find anyhow very intriguing that Kabyles have 12% U(xU6). So far I thought that U5 (never mind other U clades, although an U4 is very apparent in Taforalt ancient DNA) was extremely rare in North Africa. M1, R0 (pre-HV), X... look rather Eastern (Egypt, Ethiopia, Arabia) to me: most NW Africans don't have those lineages AFAIK, and I guess they arrived late to the area (Capsian?). The 10% of Y-DNA F* (G?) is very intriguing as well.

    ReplyDelete
  25. "What exactly you disagree with the authors re. North Africa? "

    we can't be sure that L3 did originate in East Africa,because, North Africa lost forever the vast majority of its old ancestral lineages,based on that they should not draw any conclusion at all ..

    "They don't even say that there is any Aterian lineage, because for them L3k is as young as 35 Ka (???) and arrived with U6. That L3k may be Aterian is my call but I may be wrong."

    L3k is most likely a surviving Aterian lineage ..

    "Whatever the case I find difficult to use a Kabyle sample to discern this matter"
    the Kabyle sample is very representative of the relative majority Coastal Berbers populations,therefore, quite relevant as the U6 people made the region their home (Afalou ), and later expanded to the rest of the Maghreb , as Luisa Pereira et al hypothesised if you recall..

    "I find anyhow very intriguing that Kabyles have 12% U(xU6)"

    you wondered lately if U6 entered Africa as an U* ...I agree ,and the same goes for the European U5..

    I also agree with Terryt on the DE entering Africa around the same time span.
    DE most likely settled in the Nile Delta,a group took a westward route (with U*)..(dabban industry )..
    the second larger group spread along the Nile all the way to Ethiopia ..(the Nilothic Khormusan industry)...

    ReplyDelete
  26. "we can't be sure that L3 did originate in East Africa,because, North Africa lost forever the vast majority of its old ancestral lineages,based on that they should not draw any conclusion at all .."

    No one would be able to make any type of parsimonious inferences with that line of reasoning, if that was case, then one would be able to argue that YDNA haplogroup A originated in Oceania, but has now died out, or YDNA haplogroup B originated with the Japanese, but it has long died out...sorry but that is not how phylogeny works.

    The fact of the matter is that *Modern* North Africans maternal diversity is mostly composed of a few derivative L3 lineages, while *Modern* East Africans carry lineages of L3 that are far more diverse and that are a superset of the lineages found in North Africa.

    ReplyDelete
  27. "we can't be sure that L3 did originate in East Africa,because, North Africa lost forever the vast majority of its old ancestral lineages,based on that they should not draw any conclusion at all .."

    I cannot agree with this assessment. The overall geometry of the L3 scatter points to East Africa, not just that, the overall geometry of the whole human mtDNA scatter points to East Africa, maybe a bit more to the south or maybe to that very same "pump" area of the Upper Nile.

    There is absolutely no indication of a North African origin at any stage. The only loose indication could be that shell ornaments are (as of now) known first to have existed in Palestine and North Africa but these are associated with techno-cultures (Mousterian, Aterian) that are definitely not at the origin of what most of the Sapiens fraction of Humankind are known to have used... in Africa and in Eurasia.

    The technological base is the MSA (Middle Stone Age, which is a bit like a proto-Solutrean without stone blades, a refinement of Acheulean therefore) This is what we see in most of Tropical and Southern Africa and this is what we see in India c. 80 Ka and later.

    Petraglia 2007 in fact found in the typology that, at least the Jwalapuram tech, is closest and overlapping with the Southern African facies of MSA (and not the East African one as expected). More recently, Rose, indicated that the Dhofar group (c. 130 Ka ago and after) is culturally linked to the Nubian techno-culture (of Nubia but with offshoots in Ethiopia and South Arabia) derived from Lupemban.

    There is even more in Arabia (Petraglia 2010 and 2011, Hermitage 2011) that has still to be properly linked to whatever origins in Africa. But whatever the case Northwest Africa (nor Palestine either) does not look like the origin of the main Human expansion out of Africa (and partly in Africa as well, with the L3 lineage).

    I think therefore that your criticism is unfounded and based on some kind of "geographic ethnocentrism" (as you don't even claim lineage continuity for the greatest part) and not on any objective foundation.

    As far as I can tell, NW Africa was prehistorically mostly a destination of migrations and not any source. There are probably minor exceptions, specially in association with Oranian (to West Europe and to Egypt) but that's about it.

    (cont.)

    ReplyDelete
  28. ...

    "L3k is most likely a surviving Aterian lineage" ..

    I'd love to think so but after looking at the sequence again, we are before a lineage that only shows signs of expansion 8 mutations downstream of the L3 node, what is comparable to HV or almost U (9 mut.) in Eurasia. So the expansion signature of this lineage is probably more of the age of the recolonization of West Eurasia from the East, c. 50-40 Ka ago.

    Of course, there's nothing impeding pre-L3k (L3* leading to L3k) to have been in NW Africa since right after the very moment of the L3 expansion as such but we have no evidence in favor (nor against).

    "the Kabyle sample is very representative of the relative majority Coastal Berbers populations,therefore, quite relevant as the U6 people made the region their home (Afalou ), and later expanded to the rest of the Maghreb , as Luisa Pereira et al hypothesised if you recall.."

    In truth I do not recall, sorry. Can you provide links, quotes or at least the date of publishing? What I know on U6 comes mostly from Maca-Meyer 2003 and from that data the highest basal diversity of U6 is in Morocco (followed by Canary Islands and Iberia, although maybe there's some lacking data from Kabylia or other areas).

    What I know is that there are certain distinctions and that in my Admixture exercise, the anomalous component (very distant from both Eurasian and Tropical African components) that I consider as probable Aterian remnant is found mostly among South Moroccans (c. 15%), with all other populations having very low amounts of it (<1.5%).

    It's possible that this component has survived without leaving much of a haploid lineage signature, as these tend to get fixated, by mere drift, at the numerically dominant lineages.

    "you wondered lately if U6 entered Africa as an U* ...I agree"...

    Well, it's almost necessary that it was that way.

    "... ,and the same goes for the European U5"..

    Maybe but I'm not so sure if U5 can be considered a European or rather West Asian lineage by origin.

    Something I wonder is that, if the Dabban industries impact was so limited as it appears (we'll see), maybe the first relevant West Eurasian input came from Europe and not West Asia, and did so in the Oranian genesis. I say because the West Asian genetic signature is quite weak, specially in haploid lineages. Just pondering - I can't decide, really.

    ReplyDelete
  29. "cannot agree with this assessment. The overall geometry of the L3 scatter points to East Africa, not just that, the overall geometry of the whole human mtDNA scatter points to East Africa, maybe a bit more to the south or maybe to that very same "pump" area of the Upper Nile. "

    the same thing was said about the E haplogroup's origins..

    "There is absolutely no indication of a North African origin at any stage. "
    Ifri n Ammar , Jebel Irhoud , Smuggler's cave and the A1b lineage..and that is good enough..
    "In truth I do not recall, sorry. Can you provide links, quotes or at least the date of publishing? What I know on U6 comes mostly from Maca-Meyer 2003 and from that data the highest basal diversity of U6 is in Morocco (followed by Canary Islands and Iberia, although maybe there's some lacking data from Kabylia or other areas). "

    you are right it was Maca-Mayer 2003..

    "What I know is that there are certain distinctions and that in my Admixture exercise, the anomalous component (very distant from both Eurasian and Tropical African components) that I consider as probable Aterian remnant is found mostly among South Moroccans (c. 15%), with all other populations having very low amounts of it (<1.5%)."
    I must insist that that component can be found mostly in the Moroccan Atlas range highlanders..this girl is from Khenifra -Middle Atlas / central Morocco..

    http://img515.imageshack.us/img515/4306/413081048702ea139f9ko5.jpg

    "Honestly I'm more concerned by the fact that such a large (double) sample as the Egyptian one shows no distinct personality at all, appearing as a tripartite mixture of Arabs, Iberians and Ethiopians... how come?!"

    they don't have mountains in Egypt !

    ReplyDelete
  30. "the same thing was said about the E haplogroup's origins.."

    Indeed. And the same conclusion must be reached.

    IMO E is CF'DE's and DE's (main) survivor in Africa. This is indeed peculiar (because two branches are found out of Africa and only one in Africa), what probably feeds the imagination of Dienekes and such... but all the rest of the data strongly suggests (East) African origins. And anyhow D. is highly biased (clearly racist even if he tries to dissimulate) what certainly does not add weight to his ideas in this matter.

    "they don't have mountains in Egypt"

    But it is a very distinct region, protected by deserts and able to support high population densities in a very small area, population densities that should have kept immigration from the nearly empty neighboring regions at bay (extremely diluted).

    And anyhow, why so much "European" component and why so little "North African" one?

    I think we just have not found the Egyptian distinct clusters. It probably deserves a study with its own series of samples (from various provinces), plus some neighbors as controls. One possibility is that the sample everybody is using is not really representative (for example an Alexandria sample might be much "more European" than a Teban one). There must be more (but that same thing I said of Italy and nothing so far).

    ReplyDelete
  31. "But it is a very distinct region, protected by deserts and able to support high population densities in a very small area, population densities that should have kept immigration from the nearly empty neighboring regions at bay (extremely diluted). "

    I don't think that Egypt could retain its deep ancestry better than the Ethiopia or Northwest Africa,specially during periods of severe drought , Egypt could have been repopulated several times in the past..by its immediate neighbours who happened to be of a partial Egyptian ancestry (Levant , Arabia ,Nilothic corridor.. ).

    "And anyhow, why so much "European" component and why so little "North African" one?"
    the"European"component or Anatolo-Iranic is largely Neolithic folks who were attracted to the Nile which is not that surprising..
    the Northwest African component shrank dramatically during the Neolithic period,not only in Northeast Africa,the Mediterranean islands,some areas of Southern Europe like the Balkan ,but,also in Northwest Africa itself , the west Asian component and the southwest Asian and to a lesser degree the East African came along with the Capsian wave that made it all the way to west Iberia ,but, I think that there was a larger European late Neolithic back flow that affected non-Europeans from Morocco to Northern Pakistan.

    "I think we just have not found the Egyptian distinct clusters. It probably deserves a study with its own series of samples (from various provinces), plus some neighbors as controls. One possibility is that the sample everybody is using is not really representative (for example an Alexandria sample might be much "more European" than a Teban one). There must be more (but that same thing I said of Italy and nothing so far). "

    you are right,more samples are needed , the Nile Delta region is much closer to Levant than Upper Egypt..perhaps a distinct Egyptian cluster may show up elsewhere ..
    by the way I was looking at some pictures of the western siwa Berbers oasis , what strike me was the great phenotypical diversity of its people, ranging from the unmixed armenoid , Berberid,to the Nubian..etc the Negroid admixture is unevenly distributed which is suggestive of a recent introduction ..overall I think that in this oasis of 23,000 souls,like other regions of NA ,the endogamous practices are the rule rather than the exception..
    by the way,this article is about an Algerian study about marriages among blood relatives in the country..
    (I couldn't find the original study in French)
    http://www.magharebia.com/cocoon/awi/xhtml1/en_GB/features/awi/features/2007/11/14/feature-03
    the regions of the southern part of the Algero-Tunisian border have the highest percentage of clan inbreeding,it is not very surprising that the Tunisian sample was inbred..

    ReplyDelete
  32. The data produced does not seem to support an Anatolian neolithic flow. Razib said he'd check that but haven't done so far (AFAIK). I already said that the Fst distances and such did not suggest Anatolian/Caucasian component origin.

    Just to be sure I'll make an Admixture run - hopefully today. It seems an intriguing question. I hope that without many outlier populations anyhow, the Egyptian sample will anyhow converge towards a single cluster at some K level.

    "pictures of the western siwa Berbers oasis"

    My concern is that isolated oases are too small and demographically fragile: almost any change, either internal or external may radically change the composition of the population. Internal drift will tend to fixate certain lineages and even phenotype traits, while short but intense periods of trade or conquest can easily alter the overall composition. I'd trust more a comprehensive sample of nearby Cyrenaica instead.

    "this article is about an Algerian study about marriages among blood relatives in the country.."

    That's interesting and certainly bad for their genetic pool's health. On the other hand it used to be a common practice also in Europe not so long ago: Darwin for example was partly drawn to study fitness and evolution because he and his wife (cousins) were descendants of a long line of marriages within the family and he was persuaded that was behind the poor health of his daughter, who eventually died.

    My grandmother from Balmaseda (a Castilian-influenced area) ranted often that girls in her town were given in marriage to wealthy uncles. That's a reason she disliked everything rural. I'm not sure how common was this practice although I'm pretty sure that it was rare to non-existent in other more genuine Basque areas, where it's clearly documented that people married either by love or "penalty" (euphemism for marriage because of pregnancy) or both (often the pregnancy couple were consolidated couples).

    Similarly the Kalash marry only willingly (what presumably means seldom in the family) and yet remain a very distinct population. The alleged inbreeding comes here not from marriage in the family but from marriage within the ethnic group for many many generations. It's not unhealthy that we know, unlike repeated cousin marriage (in fact there was some study in Denmark suggesting that somewhat close, but not too close, marriages were the most fertile - although the significance can be questioned, I guess).

    ReplyDelete
  33. Must say that I'm running right now Admixture with a sufficiently varied West Asian sample and, effectively, the Spanish component almost (but not quite) vanishes, as the Kurdish, Palestinian, etc. components show up and claim their share.

    However the Egyptian samples are still strangely divided among all these West Asian (and African: Berber, East African) influences, what I find very unexpected really. I would have expected more one-sideness and hence I hope that the deeper runs reveal something else: an Egyptian-specific component (plus something else).

    What I mean is that my expectation at this point (with such an even tripartite West Asian affinity at K=6, running K=8 at the moment), that it smells more like an old offshoot from undifferentiated UP West Asians than a true admixture.

    I'll tell you when the Egyptian component shows up, because I'm almost sure it will.

    ReplyDelete
  34. Done: some interesting findings, including an Egyptian specific "Eurasian" component and an Arab-Egyptian very distant (Fst) component, which may be another OoA remnant.

    ReplyDelete
  35. "So do yourself and the World a favor and write something more or less consistent that WE can debate at YOUR comments section. Thanks".

    Surely I don't have to do that, especially when you can write:

    "after looking at the sequence again, we are before a lineage that only shows signs of expansion 8 mutations downstream of the L3 node, what is comparable to HV or almost U (9 mut.) in Eurasia. So the expansion signature of this lineage is probably more of the age of the recolonization of West Eurasia from the East, c. 50-40 Ka ago".

    So in this case you agree with me that a long stem indicates a period confined to a single region. Selective interpretation of the evidence? If you are prepared to claim migration during the formation of a stem Dalouh could easily be correct. L3k could have originated anywhere.

    "Of course, there's nothing impeding pre-L3k (L3* leading to L3k) to have been in NW Africa since right after the very moment of the L3 expansion"

    Far more likely to have involved some sort of L3e'i'k'x* rather than L3*, of course.

    "IMO E is CF'DE's and DE's (main) survivor in Africa. This is indeed peculiar (because two branches are found out of Africa and only one in Africa)"

    E is the odd one out here of the whole four, and there is no evidence at all that C,F and D were ever in Africa. You are indulging in exactly what Etyopis warned against, 'if that was case, then one would be able to argue that YDNA haplogroup A originated in Oceania, but has now died out, or YDNA haplogroup B originated with the Japanese, but it has long died out'.

    ReplyDelete
  36. "Far more likely to have involved some sort of L3e'i'k'x* rather than L3*, of course".

    Sure. In any case I still think long stem and distant (upstream and downstream) nodes mean migration while in stem stage (small non-expanding lineage). Just because it's simpler: you can even divide the plausible route by the extant CR mutations and estimate where each happened.

    Re. CF'DE, I trust more mtDNA (and even upstream Y-DNA structure: A clades, B) than the anomaly of this case. IF mtDNA or something else (archaeology for instance) would support a back-migration scenario Dienekes-fantasy-style, I would have to consider it of course. But with only Y-DNA... discard.

    Y-DNA has spoiled you and so many others. Y-DNA is much more subject to wild unpredictable fluctuations and is often not informative enough, sometimes apparently contradicting everything else. That's a problem of the type of clue: Y-DNA (HLA haplotypes are also "extremist" for different reasons).

    ReplyDelete
  37. "In any case I still think long stem and distant (upstream and downstream) nodes mean migration while in stem stage"

    I've been looking back at your series of blogs on mt-DNA L, and I'm convinced of the opposite. I do agree that a long stem indicates a, 'small non-expanding lineage', but one isolated in a particular region over a long period.

    For example the basal L0 clade, L0d, has a stem 10 mutations long from L0. L0d is Khoisan and it probably developed somewhere within the Khoisan geographic range at the time (perhaps further north that Khoisan are found today). Other L0d haplogroups also have a long stem, almost certainly developed within Southern Africa: L0d2a, L0d2b, L0d2c and L0d3. These haplogroups are unlikely to have developed their stems while on the move.

    Other L haplogroups:

    L0f1 (18 mutations from its immediate ancestral haplogroup) found in South Africa

    L0a3 (13 mutations) Chad
    L0a4 (11 mutations) Kenya
    L0b (10) Ethiopia

    all three most likely to have developed in situ. A couple that you will like:

    L0f2b (16) Oman
    L0k (22) Yemen and the Khoisan (separate haplogroup in each)

    Turning to L1:

    L1b (20) Sahel and Ethiopia
    L1c (10) Central/West Africa
    L1c1b (9) Gabon
    L1c6 (16) Gabon
    L1c1c (17) Fula

    L5 (24) probably became isolated in Ethiopia.

    And so on.

    In all cases the long stem is most parsimonioulsy explained as being the product of a long period spent in a single region.

    "Re. CF'DE, I trust more mtDNA (and even upstream Y-DNA structure: A clades, B) than the anomaly of this case".

    But we know that mt-DNA and Y-DNA can be surprisingly independent. If you insist always on associating the two it will occassionally lead you down blind alleys.

    ReplyDelete
  38. I don't think what you say makes any sense. The default option is to draw a line between upstream node's and downstream node's locations and divide the line in equal segments, as many as (CR) mutations between the nodes.

    IF there is reason to think otherwise (other than mere uncertainty), it must be demonstrated reasonably.

    Also I think that in most such cases, the logical conclusion is to expect that the expanded upon arrival to a region that offered them a favorable niche. This is totally contradictory with your idea of having them "waiting" in destination, as if the destination was not yet ready for them.

    It could be in some hypothetical case, I guess but as rule it makes no sense.

    Your many examples are pointless because in no case you explain why you think what you think. You just imagine so but that's not any valid argument.

    "But we know that mt-DNA and Y-DNA can be surprisingly independent".

    For me they are incredibly coupled. The illusion of independence is in most cases just that: an illusion. I find rather easy to fit Y-DNA within mtDNA scatter patterns in most cases.

    ReplyDelete
  39. "The default option is to draw a line between upstream node's and downstream node's locations and divide the line in equal segments, as many as (CR) mutations between the nodes".

    True. But in most cases you will find that both ends of the line look most likely to have been the same place. So the stem will have developed in that place, not while the ancestors of the lineage were on the move.

    "IF there is reason to think otherwise (other than mere uncertainty), it must be demonstrated reasonably".

    The coincidence of 'origin' and 'destination' is surely all the evidence we need. Surely that is the respeonse to your, 'I don't think what you say makes any sense'.

    "Also I think that in most such cases, the logical conclusion is to expect that the expanded upon arrival to a region that offered them a favorable niche".

    It is probably the case that expansion equates to being able to exploit a newly offered niche. But in most cases that is likely to be because of climatic variation rather than migration. If the niche is relatively unfavourable surely that is the period most likely to produce severe drift. And remember that the Paleolithic population was nowhere near as huge as it later became, so the opportunity for drift within isolated populations was much greater. That appears to be exactly what we see in pre-L3 Africa.

    "Your many examples are pointless because in no case you explain why you think what you think".

    OK.

    L0f1: L0f as a whole is found in Ethiopia, Oman and South Africa, but it must originally have coalesced in a single region. L0a'b'f became widespread at 23 mutations so the ancestral haplogroup is unreliable as indication of origin. However its relation L0d is concentrated amoung the Khoisan, so it is reasonable to assume that at least some L0 was present at an early period in Southern Africa. Consequently L0f1 most likely has both its 'origin' and its 'destination' in Southern Africa. Most likely in a confined region within it until about the 35 mutation level.

    L0a3: L0a became spread from Guinea Bissau to Ethiopia and Kenya so it is reasonable to assume that L0a3 coalesced somewhere within that region. Chad is not unlikely. So again both its origin and its destination coincide.

    L0a4: The same argument holds here. Kenya is both its origin and its destination.

    L0b: This time Ethiopia is both its origin and its destination.

    L0f2b: L0f2a is Ethiopian and L0f2b is from Oman. In this case I'm more than happy to accept your argument that L0f2b's stem developed on the move. However that seems to be contrary to what you constantly claim for this haplogroup. You claim it is an ancient extra-Africa haplogroup. My suggestion supports that claim. L0f2b's ancestral haplogroup may have arrived in Oman with the OoA, at 23 mutations, where it became isolated and experienced a considerable period of drift.

    L0k: L0k1 in the Khoisan and L0k2 in Yemen. However L0k2 didn't coalesce until 32 mutations so its ancestral haplogroup probably arrived in Yemen at that time. It then became isolated and experienced drift of 8 mutations. Again origin and destination coincide.

    ReplyDelete
  40. L1b: has a long stem before it diversifies at 31 mutations through much of Africa across the Sahel. Its relation L1c appears to be Central/West African but L1b could easily be a more easterly haplogroup, although perhaps it coalesced amoung the people who became the Fula. There is no need to postulate the stems development while the haplogroup was on the move.

    L1c: But L1c's origin and destination again look most likely to coincide. It underwent a long period of drift before being able to expand. This expansion does seems to be through adaptation to a new environment, but not involving migration. Simply adaptation to a neighbouring one. L1c1a entered the nearby tropical rainforest from the region where L1c1 had been confined during its period of drift. It expanded rapidly as the Pygmies developed.

    L1c1b: Most likely Gabon is both its origin and its destination.

    L1c6: The same applies. Gabon is both its origin and its destination.

    L1c1c: The people who became the Fula is both its origin and its destination.

    L5: Both the basal L5 haplogroups have a long stem, but presumably they coalesced in separate but neighbouring regions. L5c may have coalesced in the Sudan and entered Egypt and Ethiopia at the 28 mutation level. L5a expanded at the 39 mutation level but there is no reason to assume the very long stem developed while its ancestors were on the move.

    In fact no convincing argument can be put forward for any of these haplogroups having developed their stem while on the move. The same goes for all the other long-stemmed L haplogroups. All the haplogroups with long stems look to have coalesced within the region its ancestral haplogroup colonised.

    ReplyDelete
  41. "But in most cases you will find that both ends of the line look most likely to have been the same place".

    Not sure if "most" is the word but in some cases yes.

    "So the stem will have developed in that place, not while the ancestors of the lineage were on the move".

    The move in this case equals (or rather nears) zero, so you can say both.

    But the really interesting thing is that this non-migrant subclade represents a local evolution of the mother node (which eventually vanishes, at least if old enough): sometimes there are several but sometimes only one (typically only one, even if there's another a valley further away). I think we can infer also some interesting conclusions about this but I do not think I have chewed on the matter enough to share, specially with an argument twister and ideological fighter like you.

    "The coincidence of 'origin' and 'destination' is surely all the evidence we need".

    No. The "mother haplotype" must eventually evolve into something else. Also we don't have enough precision to infer "location" but at very loose regional level.

    But enough. Seriously: write a blog. This is not the space to expose your hypothesis with such insistence and detail once and again: it's just a comment section and I would not like to have to act as moderator against you too.

    ReplyDelete
  42. "Also I think that in most such cases, the logical conclusion is to expect that the expanded upon arrival to a region that offered them a favorable niche".

    I've had a time to consider what is wrong with that statement. Any population that was unable to undergo haplotype diversification through ecologically limited population numbers is also unlikely to be able to migrate very far because of the same ecological limitation. In other words if ecology limits their ability to increase in numbers it is also likely to limit their ability to move very far.

    "Also we don't have enough precision to infer 'location' but at very loose regional level".

    But we can make very informed guesses from the distribution of related haplogroups.

    "Not sure if 'most' is the word but in some cases yes".

    See if you can find an example where the origin is definitely different from the destination. There are a few (very few) possibilities but they cannot by any means be considered definite examples.

    "But the really interesting thing is that this non-migrant subclade represents a local evolution of the mother node (which eventually vanishes, at least if old enough): sometimes there are several but sometimes only one (typically only one, even if there's another a valley further away)".

    I very much suspect it is not uncommon for separate clades to coalesce in adjacent valleys. Even L0 and L1''6 could well be an example of such a separation. If in fact your comment 'this non-migrant subclade represents a local evolution of the mother node' is correct it should then be the case that the derived haplogroup with the longest stem has remained where the mother haplogroup first coalesced. The members of the haplogroup that managed to 'escape' diversified whereas the ones who stayed behind continued to endure drift.

    "I think we can infer also some interesting conclusions about this but I do not think I have chewed on the matter enough to share, specially with an argument twister and ideological fighter like you".

    We seem to be at last getting somewhere. I have been considering the above possibility for some time, but it doesn't always hold. Other factors are obviously involved.

    "The 'mother haplotype' must eventually evolve into something else".

    Not necessarily so. If the population is reasonably large mutations that might give rise to minor haplogroups will be swamped by the 'mother' haplogroup. Mother haplogroups will generally only be replaced in relatively small isolated populations where drift is a major factor. However as my genetics lecturer said, 'any population not infinitely large will suffer drift. And no population is infinitely large'.

    "But enough. Seriously: write a blog. This is not the space to expose your hypothesis with such insistence and detail once and again: it's just a comment section and I would not like to have to act as moderator against you too".

    There are already too moany blogs out there. I think it is better to share ideas in a single blog.

    ReplyDelete
  43. "In other words if ecology limits their ability to increase in numbers it is also likely to limit their ability to move very far".

    I don't see it that way: if the environment is (a) poor or (b) crowded, it's normal that you move in search of something better. In case (a) you do move in small numbers, in case (b) maybe in larger ones.

    "See if you can find an example where the origin is definitely different from the destination".

    Many: whatever the origin you imagine for N, R, M or any other major haplogroup there's always a number of direct descendants that are far away from that origin. I don't know even how you may ask that question.

    "'any population not infinitely large will suffer drift. And no population is infinitely large'".

    Technically yes. But many populations are sufficiently large for drift to be negligible.

    So the tendency in larger populations will be (in mtDNA) stagnation of the dominant lineages (for example H1 and H3 in West and SW Europe respectively).

    "There are already too moany blogs out there. I think it is better to share ideas in a single blog".

    No. Blogs are primarily for personal (or sometimes collective but restricted) writing, what you mean is a forum. If you wish to create and manage a forum there are spaces in which that can be done for free, just as with blogs.

    ReplyDelete
  44. "I don't see it that way: if the environment is (a) poor or (b) crowded, it's normal that you move in search of something better".

    But you have to have somewhere to move to. If it is physically possible to move then members of the population will have already moved. Human population increase is the main pressure on migration. By the time the ecology becomes critical it's usually too late to move. Migration routes are already cut off.

    "Many: whatever the origin you imagine for N, R, M or any other major haplogroup there's always a number of direct descendants that are far away from that origin".

    You're getting confused again. We're concentrating here on haplogroups with a long stem. The presence of haplogroups with relatively short stems is easily explained by migration during times when population numbers are expanding. Hence the 'star-like' expansions often associated with sets of haplogroups with short stems. Sure, many haplogroups later develop a long stem once they become isolated after such an expansion. Several SE Asian haplogroups fit that scenario.

    "Technically yes. But many populations are sufficiently large for drift to be negligible".

    Correct. And such populations do not develop a long stem. Daughter haplogroups do not replace parent haplogroups so often in such a population.

    "Technically yes. But many populations are sufficiently large for drift to be negligible".

    Exactly. That's what I'm saying.

    ReplyDelete
  45. "We're concentrating here on haplogroups with a long stem".

    There we are again going in circles. I see no point to focus on long stem (as if short stemmed haplogroups could migrate faster and more farther away than long stemmed ones, that's a ridiculous idea!) and anyhow it's just crazy: your claims are untenable, your arguments make no sense and you are unable to write your own article(s) on the matter but once and again hijack unrelated threads to debate that itch of yours.

    That's trolling.

    ReplyDelete
  46. "There we are again going in circles".

    You're the only one going round in circles. I have always been concentrating on haplogroups with a long stem. You keep changing the subject.

    "as if short stemmed haplogroups could migrate faster and more farther away than long stemmed ones, that's a ridiculous idea!"

    Any stem, long or short, develops once the haplogroup's ancestor has arrived in a particular region. Stems never develop while a hapolgroup is on the move. For example M spread through much of South and Southeast Asia as M. It then diversified into the many regional variations that survive today. Same with N, L3 etc. I defy you to find any examples that contradict this scenario.

    ReplyDelete
  47. "Stems never develop while a hapolgroup is on the move".

    That is a totally unfounded claim. Bye.

    ReplyDelete
  48. "I defy you to find any examples that contradict this scenario".

    In your blog. I'm tired of going in circles and of you hijacking every other entry for this.

    ReplyDelete
  49. "That is a totally unfounded claim".

    Just one example, and you'll convince me. As it stands it is your claim that is totally unfounded. And, by the way, I'm just trying to help you where you are obviously blinded by your pre-existing beliefs.

    ReplyDelete
  50. I don't hope to convince you Terry but the easiest example at hand is L3, with all short-stemmed subclades in Africa and the long ones in Asia, far away.

    But I know well it's useless.

    End of the discussion in this blog.

    ReplyDelete
  51. "I don't hope to convince you Terry but the easiest example at hand is L3, with all short-stemmed subclades in Africa and the long ones in Asia, far away".

    You're ignoring the latest paper regarding haplogroup N that you mentioned in another post. M and N did not develop their respective stems while on the move. L3 made it out of Africa and two branches underwent drift in some region just outside that continent for some time before they were able to expand further. You can't use M and N as examples of what you're claiming.

    And I've just gone back and had a look at your diagrams of some time ago and notice several haplogroups with long stems. Have a look for yourself. The many with long stems all almost certainly coalesced where they are found today. None formed stems while on the move. The same goes for almost every other mt-DNA haplogroup. The only reason you are prepared to claim the stems in one or two problem haplogroups have developed while on the move is to make those inconvenient haplogroups fit what you already believe.

    ReplyDelete
  52. "You can't use M and N as examples of what you're claiming".

    I can indeed: they are archetypal and all you say is gibberish to me.

    You are forcing me by means of your stalking insistence to temporarily set up comment pre-approval and YOU, Terry Toothill, are banned from commenting until you make up your own homework and publish it online.

    ReplyDelete
  53. Hello There: Okay, so, my (adopted) son was born in a town called Shone in the Hadiya area of the Southern Peoples Region, Ethiopia. (Born into an Ethiopian family, yes, within a cultural/ethnic group of long standing in the area.) Anyway, and here's the point: My son's maternal haplogroup is L0f2b. According to 23andMe. So, I've been trying to figure out what this means, by googling around . . . and your blog here is one of the only sources I've found that talks about the L0 situation.

    Does it not seem interesting that an Ethiopian would have L0f2b ---? Or is this just, you know, a mom imagining that her kid has "special" DNA?

    BR

    Thanks! I enjoyed reading this conversation (the parts I could understand).
    PS: The 23andMe website said this: " . . . Some scholars have suggested that L0f may have arisen in Ethiopia, then moved south with Cushitic-speaking herders who settled in Tanzania about 4,000 years ago. But because mitochondrial DNA is maternally inherited it could also be that the haplogroup originated in Tanzania, and the Cushitic migrants were primarily men who took their wives from local L0f-bearing populations. Because very few Ethiopians with mitochondrial DNA belonging to the L0f haplogroup have been found, those competing hypotheses have been difficult to test."

    ReplyDelete
    Replies
    1. Hi, Bess. It is at least interesting for our discussion above.

      You see: African population genetics are not as well researched as European or, say, East Asian ones, more so once we consider that the diversity and complexity of the continent is very high.

      Using the limited all-Africa sampling of Behar 2008 it was argued above for example that L0f2b was only known in Oman (while its sister clade L0f2a is indeed found in Ethiopia) but now, thanks to you, we know better in this detail, as your son, whose origin is in that fascinating multiethnic region of Southern Ethiopia, has it.

      This obviously challenges what we were saying above about the origin of L0f2b and suggests that it spread from Ethiopia to Oman after the formation of the haplogroup and not before.

      It's just a detail maybe but good to know.

      As for the "molecular clock" age estimate of just 4000 years ago, I do not believe it at all. It is true that L0f has a rather long stem (cf. PhyloTree) of 9 coding region mutations (+ 5 control region ones) from its "mother" node L0a'b'f, but the derived sublineages L0f1, L0f2a and L0f2b also have very long stems downstream of it, so it makes absolutely no sense to claim such a young age for all L0f.

      A very rough (revised and improvised) guesstimate of my own could be some 40,000 years ago for L0f (with a large error margin: it could be older, probably not much younger - but depends of who is making the judgment, of course: no rocket science in this as of now).

      What I see now is that your son's L0f2b is indeed a very rare haplogroup, which is described in PhyloTree as "proposed" (mutations are written in cursive) with the only reference of Behar 2008. So, barring whatever other findings in the commercial DNA testing sector (of which I'm not aware), I'd say that your son is one of just two persons known to have that particular lineage on Earth, the other being the Omani sampled by Behar.

      Pretty special yes, and of some informative value for population geneticists: it confirms that the haplogroup does exist and it expands the geographical understanding of that particular lineage and also of the parent haplogroups L0f2 and L0f (both small lineages). In my understanding, it also helps to emphasize the importance of SW Ethiopia in the deep (pre-)history of Humankind, showing that there is still a lot to research in that region, which has also provided us with the oldest known specimen of Homo sapiens (Omo 1, dated to c. 195,000 years ago).

      So, would I be you, I would feel morally legitimated to share that information with anyone potentially interested, from specialized forums to professional geneticists with a focus in Africa.

      Delete
  54. So cool! Thank you very much for the information, Maju.

    First off, I am happy that I wasn't just being overenthusiastic mom. ;-)

    Second, I could certainly point a researcher in the direction of grandmother and aunts from my son's birth family, who would have the same haplogroup. (We have contact.) If someone is looking for L0f2b in Southern Ethiopia, I think I know where to find it.

    When I first got the results from 23andMe, I wrote to a researcher who has been examining results from adoptive families from Ethiopian -- for their own research, I gather -- but I haven't heard back yet. So then I wrote to a professor named Tishkoff, who seems to have done research on the L0f groups. Again, I haven't heard back. So . . . that's why I wrote to you. And I really appreciate it that you replied!

    We'll see what happens next. Maybe those two researchers will reply sooner or later.

    THANKS AGAIN!

    ReplyDelete
    Replies
    1. I can't say because I haven't ever reported on any such rare lineages showing up here or there, somewhat unexpectedly. But I do imagine that those authors and research teams with an African focus (mtDNA especially, not many) must be interested in knowing this detail. Now, when they do research they are most likely wanting to make random "blind" samples (or reuse them from previous studies), so the info, as far as I understand is more like a note on a map saying: "SNPP region: interesting for research, having for example this very rare lineage so far not found elsewhere in Africa (just Oman)". Probably there are many other such "notes" on that and other places and together they may serve the researchers.

      Now, whether researchers show interest or not I can't say... it may depend on what mood they woke up the particular data they get your info, you know. I just say that the info is quite interesting and fills a blank on our knowledge, and that, as such, should be of some interest to researchers with an African or Global mtDNA focus.

      Some of them are: Doron M. Behar, Danielle A. Badro, Karina M. Schlabusch and Pedro Soares. These are lead researchers with links to their key papers, where you should find the "contact author", i.e. the one dedicated to public relations, not necessarily the lead researcher, which should interested in your info, especially if they are planning further research.

      Anyhow, a most important site where you should share your son's (as well as others') mtDNA results is mtDNA community. This is database is managed by the same researchers who built and maintain the PhyloTree (the standard mtDNA phylogenetic reference) and AFAIK is a major reference for data mining. I don't know too many details, so you better explore it on your own.

      Delete
    2. Oops! MtDNA Community link: http://www.mtdnacommunity.org/

      Delete
  55. Hey Bess,

    Your son's L0f lineage is not rare everywhere, as was revealed in a recent thesis from Tishkoff's lab. What is now clear is that mtDNA L0f in general is found at high frequencies in hunter-gatherer populations living to the south of Ethiopia. It is also found in Cushitic-speaking populations from that area, who have a long history of interaction with hunter-gatherers since their migration from Ethiopia a couple of thousand years ago. L0f2 was found to be particularly frequent in hunter-gatherers from northern Kenya, whereas L0f1 was more common in southern Kenyan hunter-gatherers.

    mtDNA L0f in studies on Ethiopians is generally found at trace frequencies or absent, with its limited presence probably being related to limited interaction with areas to the south. But, who knows, since your son is from the southern regions, he may descend from an unstudied group with a higher frequency of this lineage.

    So basically, from the quote you posted above, the recent thesis supports the following scenario:

    "But because mitochondrial DNA is maternally inherited it could also be that the haplogroup originated in Tanzania, and the Cushitic migrants were primarily men who took their wives from local L0f-bearing populations."

    Except you could replace "Tanzania" with Rift Valley hunter-gatherers from Kenya and Tanzania.

    ReplyDelete
    Replies
    1. Great reference, Lank. I was not fully aware of it. Still, they only seem to have tested for L0f2 and not for downstream mutations, so it may still have some relevance (assuming anyone is interested in that level of detail).

      Per PhyloTree references, L0f2a, L0f1a and L0f2b (and even for upstream lineages within L0f) only Behar 2008b shows up. This may mean that nobody other than him, has tested for downstream lineages within L0f2. His sample produced two identical L0f2a among 19 Ethiopians with L(xM,N) (so apparently not as rare as you suggest: maybe 5%), another distinct sample of L0f2a in the USA (n=93 L(xM,N) individuals) and then one L0f2b among 10 Omanis with L(xM,N), as mentioned before.

      An intriguing element here is how the downstream structure of L0f2 is distributed geographically. It is somewhat possible that all or most of Hirbo's Kenyan-Tanzanian L0f2 belongs to a single subclade, maybe the "American" one in Behar's paper. Knowing that in greater detail would be quite interesting if we want to properly describe this haplogroup and speculate on its origin.

      As of now we know that L0f2 has an Eastern African distribution (from at least Sudan, Ethiopia, Kenya and Tanzania) with at least some cases in Arabia peninsula (although these seem derived). We also know that the lineage seems totally absent in the Western, Southern and Northern parts of the continent. But we do not know its distribution structure by subclades, which would have relevance in order to explain the lineage's "history", i.e. if it migrated from Ethiopia-Sudan to Kenya-Tanzania or vice versa.

      As happens with L4, which seems to have migrated from the Upper Nile or Red Sea regions to Tanzania, where it's only found as a single subclade, this may also be the case with L0f.

      Delete
    2. So you think 19 Ethiopians from Behar 2008 outweigh all of the other work on Ethiopian mtDNA? Think again; there is a lot more mtDNA data on Ethiopians. Starting with Kivisild 2004, with hundreds of Ethiopian samples and just 1 example of L0f (Amhara). That study includes some Gurage BTW, neighbors of the Hadiya Zone, with no L0f.

      This study with some comprehensive East African mtDNA data does report fairly high frequencies of L0f in Daasanach (found in Kenya, Sudan and the southern outskirts of Ethiopia), but in some of the other Omo Valley groups it is lower, almost absent in some. The high variability of L0f in this region, as opposed to e.g. L0a is probably a result of the culturally dominant (Omotic/Cushitic/Nilotic) groups assimilating previously structured groups of hunter-gatherers.

      At 23andMe, I'm sharing with a Somali who's reported L0a'b'f*, but who actually shares some mutations with L0f. So the lineage is basically pre-L0f. L0b also is found at trace frequencies in mainly southern Ethiopia, Kenya and Tanzania. IMO L0a'b'f as a whole probably traces back to this general region, which seems to be the oldest habitat of our species, with the ancestors of L0d and L0k heading south at some point.

      Delete
    3. Not that they "outweight" but that they weight similarly.

      Importantly in the Kivisild paper, the SNNP Region was not studied, as was not either the also interesting Gambela Region.

      The Boattini study is pay-per-view, so I do not know what was sampled exactly, but for what you say it should be interesting to know, because you are mentioning yourself high frequencies in some groups and lower in others (absence in none), and underlining the high diversity of L0f in that region of SW Ethiopia, which is also very diverse in other aspects.

      What you mention about a Somali with pre-L0f is also very interesting indeed.

      I don't see reason for disagreement.

      Delete
  56. This comment has been removed by the author.

    ReplyDelete
  57. Thanks, Lank and Maju.

    I'll visit that MtDNA Community this evening.

    I'd read -- as best I could -- that stuff from the Tishkoff L0f study, on her blog. But I am not quite sure I follow the theory of how, if L0f potentially originated in Kenya or Tanzania, it would account for there being L0f2b people in the highlands of Ethiopia.

    So, Lank, is what you're saying this?: That it's possible some Ethiopian migrants, a few thousand years ago, came down from Ethiopia (speaking Cushitic) and came to the L0f-originating area in Tanzania -- and/or Kenya -- leaving behind the Cushitic language, and descendents with L0F haplo in the Rift Valley . . . but that other men among these Ethiopian migrants returned home to the highlands of Ethiopia with wives who were Kenyan/Tanzanian L0F. . .? and they left descendents in this area of Ethiopia with L0f2b?

    I guess that sort of makes sense.

    Tho it makes you wonder how the L0f2b ended up in Oman. . . . . Unless, I guess, some men from Oman did the same: Traveled to Southern Ethiopia, OR to Kenya/Tanzania, and brought home wives. . .

    PS: Since the people in Tanzania who are L0F speak a language that originated in Ethiopia (Cushitic), doesn't it seem simpler or more likely that L0f originated in Ethiopia, and spread to Tanzania/Kenya AND to Oman? Rather than the other way around?

    Anyway, this is great fun to learn about!! So GIANT THANKS, Y'ALL!

    ReplyDelete
    Replies
    1. So, Lank, is what you're saying this?: That it's possible some Ethiopian migrants, a few thousand years ago, came down from Ethiopia (speaking Cushitic) and came to the L0f-originating area in Tanzania -- and/or Kenya -- leaving behind the Cushitic language, and descendents with L0F haplo in the Rift Valley . . . but that other men among these Ethiopian migrants returned home to the highlands of Ethiopia with wives who were Kenyan/Tanzanian L0F. . .? and they left descendents in this area of Ethiopia with L0f2b?

      The history of the minor L0f in Ethiopia is probably more complicated than that, reflecting more ancient interactions with populations high in L0f, I would bet.

      Tho it makes you wonder how the L0f2b ended up in Oman. . . . . Unless, I guess, some men from Oman did the same: Traveled to Southern Ethiopia, OR to Kenya/Tanzania, and brought home wives. . .

      L0f is also present in East African Bantu populations, who were highly involved in the Arab slave trade.

      PS: Since the people in Tanzania who are L0F speak a language that originated in Ethiopia (Cushitic), doesn't it seem simpler or more likely that L0f originated in Ethiopia, and spread to Tanzania/Kenya AND to Oman? Rather than the other way around?

      Not really, no. L0f is generally rare in Ethiopians, who represent the Cushitic ancestors of some Kenyans/Tanzanians. Instead, L0f is frequent in the hunter-gatherer populations in Kenya/Tanzania. They may speak Cushitic languages today, but it is generally acknowledged that they are "language shifters" on the basis of their distinct genetic profile, hunter-gatherer lifestyle (as opposed to the pastoral Cushites), and sometimes even with a tradition of recently having spoken a different language. The reason why Cushites have been so influenced by these hunter-gatherers (and vice versa) is that they were among the oldest migrants to the region; they were there before the now more dominant Bantu and Nilotic groups arrived.

      Delete
    2. "L0f is also present in East African Bantu populations, who were highly involved in the Arab slave trade".

      Maybe. But something I gathered when reviewing Behar 2008 back in the day was that the lineages reported in Arabia (and also North Africa, different ones) were mostly not the most common ones in the Tropical African regions that fed the slave trade. In fact they were pretty rare like this L0f2b, what made me think that they are actually residuals of much older migrations, possibly the OoA itself.

      Delete
  58. PS: This study -- by Tishkoff, Gonder, Mortensen, etc -- says that while L0d and L5 are "rare" outside Tanzania, that L0f is "absent" in any African region outside Tanzania:

    http://mbe.oxfordjournals.org/content/24/3/757.full.pdf

    ReplyDelete
    Replies
    1. Very interesting and right out of the oven today, thanks!

      I understand that those claims must be considered always within the dataset of the study, which for what I can see, it is very much focused on Tanzania. At least in fig. 3, Tanzanians are the only East Africans considered, all the rest being from other macro-regions. As we have seen before, L0f is actually spread through East Africa (in, at least, Sudan, Ethiopia, Kenya and Tanzania) but seemingly absent from other subcontinental regions. I take this study as confirming this, not contesting it.

      Delete
    2. @Bess: Notice that for Gonder et al., the age of L0f is of more than 90,000 years ago, pretty different than the mere 4000 years you mentioned before from a 23&Me source.

      Delete
  59. Thanks so much for all the amazing info, you guys! It's been great getting a bit more educated about this stuff. . . . I'm grateful . . .

    ReplyDelete
    Replies
    1. Thanks for sharing the interesting result. If you want to share at 23andMe, send me (Medri Bahri) an invite. :)

      Delete
  60. Great, I will, Lank, thanks!

    ReplyDelete

  61. Importantly in the Kivisild paper, the SNNP Region was not studied, as was not either the also interesting Gambela Region.

    The Boattini study is pay-per-view, so I do not know what was sampled exactly, but for what you say it should be interesting to know, because you are mentioning yourself high frequencies in some groups and lower in others (absence in none), and underlining the high diversity of L0f in that region of SW Ethiopia, which is also very diverse in other aspects.


    Out of Behar's 2 Ethiopian L0f samples, one is from Kivisild, and the other is an Ethiopian Jew, among whom L0f is rare (2% in another dataset).

    You can read a summary of the data from Boatinni's study at Ethio Helix's blog. L0f was absent in Hamer and Ongota, and rare in the larger sample of Dawro (1.5%). These are all groups with evident, ancient roots in the Omo Valley. As you cross into Kenya, you find L0f in 6% of Rendille and 19% of El Molo, both Cushitic-speaking groups. As you head south, L0f becomes increasingly common, reaching 25% in the Iraqw and 32% in the Burunge. These last two are Cushites from Tanzania.

    Maybe. But something I gathered when reviewing Behar 2008 back in the day was that the lineages reported in Arabia (and also North Africa, different ones) were mostly not the most common ones in the Tropical African regions that fed the slave trade. In fact they were pretty rare like this L0f2b, what made me think that they are actually residuals of much older migrations, possibly the OoA itself.

    These parts of East Africa are severely understudied, particularly for high resolution mtDNA. We don't even know which sublineages of L0f are found in the groups studied by Hirbo, beyond the tripartite split into L0f1, L0f2, L0f3. And the Ethiopian L0f samples are too few. I highly doubt it traces back to OOA. This extremely diverse (southern Ethiopia + some Kenyan/Tanzanian groups) region should be studied before speculating about an ancient origin of Arabian L(xM,N) lineages. Also see mtDNA L6, found at a high frequency in Yemen and uncommon in most Ethiopians, but now we know that it's actually common in Omotic groups from southwestern Ethiopia.

    ReplyDelete
  62. BTW, 23andMe's 4 kya figure is in reference to the southward migation of Cushites, not the age of the haplogroup.

    ReplyDelete
  63. which tribe belongs L1c1d mtdna haplogroup?

    ReplyDelete
    Replies
    1. Behar 2008 reported it among the Igbo, in Gabon and among the Khoisan (Bushmen or Khoikoi remnants like the Griqua) → http://leherensuge.blogspot.com/2010/03/reviewing-mtdna-l-lineages-notes-l1.html

      This does not necessarily mean it is not found elsewhere, but at least in those places/populations. African population genetics is still in need of wider and more detailed research.

      Delete
  64. This comment has been removed by the author.

    ReplyDelete
  65. worship this blogger super informative and enlightening

    ReplyDelete
  66. Maju Thanks for providing me with this information useful! 'm mtdna L1c1d

    ReplyDelete
    Replies
    1. Thanks to you Silvia for your kind comments.

      Delete

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).