RSS

Ancient European DNA and some debatable conclusions

26 Dec
There is a rather interesting paper still in preparation available online and causing some debate.
Iosif Lazaridis, Nick Patterson, Alissa Mittnik, et al., Ancient human genomes suggest three ancestral populations for present-day Europeans. BioArxiv 2013 (preprint). Freely accessibleLINK [doi:10.1101/001552]

Abstract

Analysis of ancient DNA can reveal historical events that are difficult to discern through study of present-day individuals. To investigate European population history around the time of the agricultural transition, we sequenced complete genomes from a ~7,500 year old early farmer from the Linearbandkeramik (LBK) culture from Stuttgart in Germany and an ~8,000 year old hunter-gatherer from the Loschbour rock shelter in Luxembourg. We also generated data from seven ~8,000 year old hunter-gatherers from Motala in Sweden. We compared these genomes and published ancient DNA to new data from 2,196 samples from 185 diverse populations to show that at least three ancestral groups contributed to present-day Europeans. The first are Ancient North Eurasians (ANE), who are more closely related to Upper Paleolithic Siberians than to any present-day population. The second are West European Hunter-Gatherers (WHG), related to the Loschbour individual, who contributed to all Europeans but not to Near Easterners. The third are Early European Farmers (EEF), related to the Stuttgart individual, who were mainly of Near Eastern origin but also harbored WHG-related ancestry. We model the deep relationships of these populations and show that about ~44% of the ancestry of EEF derived from a basal Eurasian lineage that split prior to the separation of other non-Africans.

Haploid DNA
The Lochsbour skull.
The prominent browridge
is very unusual for
Paleolithic Europeans.
The new European hunter-gatherer samples carried all Y-DNA I and mtDNA U5a and U2e.
More specifically, the hunter-gatherer mtDNA lineages are:
  • Lochsbour (Luxembourg): U5b1a
  • Motala (Sweden):
    • Motala 1 & 3: U5b1a
    • Motala 2 & 12: U2e1
    • Motala 4 & 6: U5a2d
    • Motala 9: U5a2
Additionally the Stuttgart Linear Pottery farmer (female) carried the mtDNA lineage T2c1d1.
The Y-DNA lineages are:
  • Lochsbour: I2a1b*(xI2a1b1, I2a1b2, I2a1b3)
  • Motala 2: I*(xI1, I2a2,I2a1b3)
  • Motala 3: I2*(xI2a1a, I2a2, I2b)
  • Motala 6: uncertain (L55+ would make it Q1a2a but L232- forces it out of Q1)
  • Motala 9: I*(xI1)
  • Motala 12: I2a1b*(xI2a1b1, I2a1b3)
These are with certainty the oldest Y-DNA sequences of Europe so far and the fact that all them fall within haplogroup I(xI1) supports the notion of this lineage being once common in the subcontinent, at least in some areas. Today I2 is most common in Sardinia, the NW Balcans (Croatia, Bosnia, Montenegro), North Germany and areas around Moldavia.
I2a1b (which may well be all them) is currently found (often in large frequencies) in the Balcans and Eastern Europe with some presence also in the eastern areas of Central Europe. It’s relative I2a1a is most common in Sardinia with some presence in SW Europe, especially around the Pyrenees. I2a1 (probably I2a1a but not tested for the relevant SNPs) was also found, together with G2a, in a Chalcolithic population of the Treilles group (Languedoc) and seems to be somehow associated to Cardium Pottery Neolithic.
If you want my opinion, I’d think that I2a before Neolithic was dominant, like mtDNA U5 (and satellites U4 and U2e), in much of Central and Eastern Europe but probably not in SW Europe, where mtDNA U5 seems not so much hyper-dominant either, being instead quite secondary to haplogroup H (at least in Western Iberia). But we’ll have to wait until geneticists manage to sequence Y-DNA in several SW European Paleolithic remains to be sure.

Autosomal DNA and derived speculations
Most of the study (incl. the must-read supplemental materials) deals however with the autosomal DNA of these and other hunter-gatherers, as well as of some Neolithic farmers from Central Europe and Italy (Ötzi) and their comparison with modern Europeans. 
To begin with, they generated a PCA plot of West Eurasians (with way too many pointless Bedouins and Jews, it must be said) and projected the ancient Europeans, as well as a whole bunch of Circum-Pacific peoples on it:
The result is a bit weird because, as you can see, the East Asians, Native Americans and Melanesians appear to fall way too close to the peoples of the Caucasus and Anatolia. This seems to be a distorting effect of the “projection” method, which forces the projected samples to align relative to a set of already defined parameters, in this case the West Eurasian (modern) PCA. 
So the projection basically formulates the question: if East Asians, etc. must be forcibly to be defined in West Eurasian (WEA) terms, what would they be? And then answers it as follows: Caucasian/Anatolian/Iranian peoples more or less (whatever the hidden reasons, which are not too clear).
Similarly, it is possible (but uncertain) that the ancient European and Siberian sequences show some of this kind of distortion. However I have found experimentally that the PCA’s dimension 1 (but not the dimension 2, which corresponds largely to the Asian-specific distinctions) still correlates quite well with the results of other formal tests that the authors develop in the study and is therefore a valuable tool for visualization.
But this later. By the moment the PCA is asking and answering three or four questions by projecting ancient European and Siberian samples in the West Eurasian plot:
  • If ancient Siberians are forced to be defined in modern WEA terms, what would they be? Answer: roughly Mordvins (Afontova Gora 2) or intermediate between these and North Caucasus peoples (Mal’ta 1).
  • If ancient Scandinavian hunter-gatherers are forced in modern WEA terms, what would they be? Answer: extreme but closest (Skoglund) to Northern European peoples like Icelanders or Lithuanians.
  • If ancient Western European hunter-gatherers are forced in modern WEA terms, what would they be? Answer: extreme too but closest (La Braña 2) to SW European peoples like Basques and Southern French.
  • If ancient Neolithic/Chalcolithic farmers from around the Alps and Sweden are forced in modern WEA terms, what would they be? Answer: Canarians (next close: Sardinians, then Spaniards).
Whatever the case, there seems to be quite a bit of autosomal diversity among ancient Western hunter-gatherers, at the very least when compared with modern peoples. This makes some good sense because Europe was a big place already in Paleolithic times and must have harbored some notable diversity. Diversity that we may well find to grasp if we only sample people from the same areas once and again.
On the other hand, they seem to cluster in the same extreme periphery of the European cluster, opposed to the position of West Asians, and therefore suggesting that there has been some West Asian genetic flow into Europe since then (something we all assume, of course). 
Using Lochsbour as proxy for the WHG (Western hunter-gatherer) component, Mal’ta 1 as proxy for the ANE (ancient north Eurasian) one and Stuttgart as proxy for the EEF (early European farmer) one, they produce the following graph (to which I added an important note in gray):
The note in gray is mine: highlighting the contradictory position where the other Western hunter-gatherers may fall in because of assuming Lochsbour as valid proxy, when it is clearly very extreme. This was not tested in the study so it is inferred from the PC1, which seems to best approach the results of their formal tests in the WHG vs EEF axis, as well as those of the WHG vs Near East comparisons.
I tried to figure out how these formal tests are reflected, if at all in the PCA, mostly because the PCA is a much easier tool for comprehension, being so visual. Eventually I found that the dimension 1 (horizontal axis) is very close to the genetic distances measured by the formal tests (excepted those for the ANE component, obviously), allowing a visualization of some of the possible problems caused by their use of Lochsbour as only reference, without any control. Let’s see it:

The same PCA as above with a few annotations in magenta and green
While not exactly, the slashed vertical magenta line (median in the dimension 1 between Lochsbour and Stuttgart) approximates quite well the WHG vs EEF values measured in the formal tests. Similarly, the slashed green axis (median in PC1 between Lochsbour and an good looking Bedouin) approximates to a great extent the less precise results of the formal tests the authors applied to guesstimate the West Asian and WHG ancestry of EEFs, which ranged between 60% and almost 100% West Asian (my line is much closer to the 60% value, which seems more reasonable). 
When I tried to find an alternative median WHG/West Asian line, using Braña 2 and the first non-Euro-drifted Turk I could spot (Anatolia is much more likely to be the direct source of West Asian ancestry in Europe than Bedouins), I got exactly the same result, so no need to plot any second option (two wrongs sometimes do make one right, it seems). But when I did the same with La Braña 2 and Stuttgart I got a genuine good-looking alternative median line, which is the slash-and-dot magenta axis.
This alternative line is probably a much more reasonable 50% WHG-EEF approximation in fact and goes right through Spain, what makes good sense for all I know.
Of course the ideal solution would be that someone performed good formal tests, similar to those done in the study, with Braña 2 and/or Skoglund, which should be more similar to the actual WHG ancestry of modern Europeans than the extremely divergent Lochsbour sequence. An obvious problem is that La Braña produced only very poor sequences but, well, use Skoglund instead or sample some Franco-Cantabrian or Iberian other Paleolithic remains.
Whatever the solution, I think that we do have a problem with the use of Lochsbour as only WHG proxy and that it demands some counter-testing. 
What about the ANE component? I do not dare to give any alternative opinion because I lack tools to counter-analyze it. What seems clear is that its influence on modern Europeans seems almost uniformly weak and that it can be ignored for the biggest part. As happens with the WHG, it’s quite possible that the ANE would be enhanced if the sequence from Afontova Gora is used instead of that of Mal’ta but I can’t foresee how much. 
Finally some speculative food-for-thought. Again using the visual tool of the PCA, I spotted some curiosities:

Speculative annotations on the PCA

Most notably it is apparent that the two WHG populations (Western and Scandinavian) are aligned in natural axes, which seem to act as clusters. Extending both (dotted lines) they converge at a point closest to some French, notably the only “French” that tends towards “Southern France” and Basques. So I wonder: is it possible that these two WHG cluster-lines represent derived ancient branches from an original population of SW France. We know that since the LGM, the area of Dordogne (Perigord) was like the megapolis of Paleolithic Europe, with population densities that must have been several times those of other areas. We know that this region was at the origin of both Solutrean and Magdalenian cultures and probably still played an important role in the Epipaleolithic period. 
So I do wonder: is that “knot” a mere artifact of a mediocre representation or is it something much more real? Only with due research in the Franco-Cantabrian region we will find out. 
Advertisements
 

112 responses to “Ancient European DNA and some debatable conclusions

  1. Davidski

    December 26, 2013 at 2:37 pm

    “On the other hand, they seem to cluster in the same extreme periphery of the European cluster, opposed to the position of West Asians, and therefore suggesting that there has been some West Asian genetic flow into Europe since then (something we all assume, of course). “

    It's not West Asian gene flow to Europe, it's North Pontic Steppe gene flow to both Europe and West Asia, along with R1. In other words, ANE + R1.

    This started during the Copper Age, and continued for a while. From the paper…

    “We observe a striking contrast between Europe west of the Caucasus and the Near East in degree of relatedness to WHG. In Europe, there is a much higher degree of allele sharing with Loschbour than with MA1, which we ascribe to the 60-80% WHG/(WHG+ANE) ratio in most Europeans that we report in SI12. In contrast the Near East has no appreciable WHG ancestry but some ANE ancestry, especially in the northern Caucasus.”

    http://img12.imageshack.us/img12/3053/48ms.png

     
  2. Maju

    December 26, 2013 at 5:45 pm

    No. You seem to have completely misunderstood the matter (probably because of wishful thinking).

    The formal ANE scores are low across the board with a record high of 18.7% for Estonians and a record low of 4.1% among Sardinians. Most other populations are around 15%, except Basques which have a slightly lower score of 9.8%.

    What your quote says is that West Asians show no WHG admixture (logical!) but that they do show some ANE admixture also, notably North Caucasians.

    Mixing WHG with ANE is of no help because these components probably correspond to two different ancestral populations, even if their origins may have touched at some point. In essence the ANE component should correspond to Altaian UP (or at most Eastern Gravettian, i.e. last likely contact with Europe before the LGM) and the WHG must correspond to European specific later developments such as Solutrean and Magdalenian (or also Hamburgian-Ahrensburgian in the North).

    All them derive from ultimately West Asian origins (Aurignacoid cultures and later Gravettian originated all in West Asia) and that's why we all cluster as West Eurasians but beyond those pre-LGM common deep origins is where the differences between WHG, ANE and West Asian components begin.

    In the supplemental material it is formally tested the EEF affinities as function of Lochsbour and a Bedouin, with results that vary depending on the other control populations but ranking between slightly above 60% West Asian to almost 100% West Asian, so the WHG component in EEF is not greater than 40% most likely (the ANE component should be negligible, as happens with Sardinians but not directly tested for AFAIK).

    Then, as everyone can see in the simplified triangular plot Europeans align far away from the ANE component between the WHG and the EEF proxies. I question the validity of Lochsbour as good proxy for WHG (too drifted away probably) but otherwise I see no obvious errors in that triangular plot.

    The ANE component could be higher than spotted in the paper (AG2 is closer to Europeans than MA1) but I can't evaluate that with any clarity. The map you link to anyhow is just a measure or relative affinity in which we can see how Palestinians or Syrians score similar to Spaniards or peninsular Italians in ANE (MA-1) affinity, that would make them ~12% ANE, maybe a bit less but clearly above the Sardinian score of 4%, which is represented in black color.

    The right map illustrates the same kind of variance in the WHG-EEF axis and this one obviously excludes West Asians because they would be hyper EEF always (off the scale). But this map says nothing about ANE, only the left one does.

    I see no reason to relate the ANE component with R1 as you want to do, although I guess that some of its influence is indeed owed to Indoeuropean invasions and other flows from Eastern Europe (Uralics, Pitted Ware, antique-medieval slave trade and even the Huns if you wish). But alone it cannot explain R1b for example. I'd rather believe in the Neolithic origin hypothesis for R1b than in the Indoeuropean one, which is absolute nonsense. Now, some R1a correlation in Eastern-Central Europe and even Central Asia… that I can agree to (but not to a simplistic R1-IE correlation, which makes zero sense).

     
  3. Maju

    December 26, 2013 at 6:02 pm

    You correctly has insisted a lot in reading the supplemental material. Well the relevant data for Stuttgart's WHG-WestAsian admixture is discussed in SI-10. Quote:

    “The amount of Near Eastern admixture estimated for Stuttgart can be seen in Table S10.2 and range between 61-98% with estimates increasing as the amount of estimated African admixture in BedouinB increases”.

    The near 60% West Asian admixture figures correspond all to an assumed Bedouin African admixture of 4.2%. For an assumed African admixture in Bedouins of 5.1%, the West Asian admixture estimates in EEF are around 70%. For an assumed African admixture in Bedouins of 7.2%, the West Asian component in EEF would be around 90%.

    Considering my independent estimates (using Braña 2 and a Turkish proxy on the PCA) I would think that the realistic figure would be c. 60%. But that also should imply, I believe, that my alternative line for the EEF-WHG half-way score is approximately correct.

    More in depth it may also imply that, even if Bedouin African admixture is effectively higher than the assumed 4.2%, the excess was anyhow present in the EEF population, as its Canarian affinity suggests (and the Neolithic lineage of African roots E1b-V13 may help to explain).

     
  4. Davidski

    December 26, 2013 at 8:08 pm

    Uralics, Turks and other post-Copper Age easterners brought more Eastern Non-African (ENA) than ANE to Europe. That's why some Northeastern Europeans are outliers today.

    Post-Mesolithic ANE in Europe is linked to Indo-Europeans, and didn't come with any ENA. Moreover, basically all Near Eastern admixture in Europe, except among Sicilians, Maltese and Jews, can be explained by ancestry from EEF, not requiring additional gene flows from West Asia.

     
  5. Maju

    December 27, 2013 at 5:12 am

    “Eastern Non-African (ENA)”

    … is a mere catch-all term with no score anywhere. Some ENA like Amerindians are strongly related to MA-1 and therefore the more real ANE component. In principle they should be treated as various populations being just a convenience term for all non-Africans who are also non-WEA.

    We can't say nothing about such paraphyletic element in any case because it's measured nowhere.

    “Post-Mesolithic ANE in Europe is linked to Indo-Europeans”…

    Maybe but we do not know. We do not even know when it arrived to Europe, we do not know if it was already affecting WHGs, we only have some quite uniformly low scores for this component across the board with records in Uralic peoples (Estonians but also off-the-scale Finnish, Russians, Mordvins). We cannot say much about that other than it peaks in NE Europe, notably in Uralic peoples.

    You talk of this stuff as if it'd be as you say: “X is Y, A is B, Z happened in the W way” without no cautions and quite annoyingly without support of any kind. That way tentative opinions are ascended to the category of dogma.

    I don't like that at all. We can't debate that way. That's like a believer preaching religious dogma: no alternative opinions allowed, no need for evidence. Why do you do that? It's a mere propaganda trick: not serious scientific debate. I know you can do a lot better.

    “Moreover, basically all Near Eastern admixture in Europe, except among Sicilians, Maltese and Jews, can be explained by ancestry from EEF, not requiring additional gene flows from West Asia”.

    This part may be correct, at least I think it makes some sense on first sight.

    But then: why EEFs don't lean towards Turkey: they should be some >60% West Asian, what means Anatolian (with lesser some other affinity) but their tendency is not towards Turkey at all but towards Palestine, Bedouins, Yemen… It's not possible that a mere 10-18% ANE influence can distort the modern European ancestry that much in the PCA, so maybe there's more to West Asianness than what you say.

    Rather than rushing to conclusions, as you seem so willing to do, I get lots of questions. And I think that difficult questions are much better than misleadingly easy answers. Explanations must be produced if the data allows for them but when there is room for doubt it's better to generally remain in the question zone and at most produce clearly tentative hypothesis and not oversimplifying unfounded dogmas.

     
  6. Davidski

    December 27, 2013 at 6:45 am

    There's no need to make this so complicated. The paper doesn't settle everything, because more ancient genomes are required for that. But it gives a very clear chronology of how the European gene pool formed, including two sets of outliers, in the Northeast and Southeast.

    WHG/ANE = Mesolithic foragers

    EEF = Neolithic farmers

    ANE/WHG = Indo-Europeans

    ENA/ANE = Uralics

    Finally, there was West Asian, North African and Sub-Saharan gene flow into Southern Europe.

    I don't see how you can argue with that if you read the paper and looked at all the supplementary data properly. It is what it is, and we'll soon see it confirmed beyond any doubt when more genomes get sequenced.

     
  7. Matt

    December 27, 2013 at 11:32 am

    ANE people living in the Pontic-Caspian would integrate neatly with the Pontic-Caspian Indo-European hypothesis, so I can see why that would be compelling. But I don't really see any way to test that their meta-population was or wasn't also already in Bactria or other adjacent steppe regions prior to the Neolithic – it seems possible to me they already were given the presence of a component with ANE like distribution across West Asia in non-Indo European folk.

    One problem with the f3 statistics for Caucasian / West Asian people here is that the Stuttgart Woman, the reference for farmers who were migrating to that, is already quite mixed with West Eurasian / West European Hunter Gatherer clade, to the proportion of around 56% of her ancestry.

    Some of this may reflect a real pre-Neolithic or early Neolithic population fusion among Near Eastern “farmers” but much of it is probably due to expansion to Europe.

    So the problem is that this sets an artificial floor of how much West European Hunter Gatherer ancestry ancestry is in the West Asian populations.

    If using Stuttgart as a reference for early farmers and the early Near East, West European Hunter Gatherer ancestry in West Asian populations cannot go less than around 50% of their early farmer component!

    This is a real problem since most West Asian populations probably don't have ANY real West European Hunter Gatherer ancestry. They're instead more likely to be almost wholly mixes of Ancient North Eurasian and Basal Eurasian (or ancient Mediterranean, ancient Near Eastern, whatever this ghost population is best termed). As WEHG and ANE are similar this results in reduced estimates of ANE.

    If we look at an example of how this affects Lezgins, they're estimated to be 0.712 Near East using Stuttgart as a reference for Near East and 0.288 ANE.

    This would decompose to 0.313 “basal Eurasian” (or ancient Near Eastern), 0.399 West European Hunter Gatherer, and 0.288 ANE.

    However in reality, it seems likely to me based on admixture that they're instead closer to around 0.686 (0.399 + 0.288) ANE (perhaps with some slight WEHG fraction).

    Another example is why Ashkenazi Jews come out as the “closest” population to EEF at around 0.931 EEF and 0.069 ANE in the comparison panels as well, this despite the fact that we know that Sardinians are a much better proxy for EEF.

    In reality they're probably have more ANE ancestry, but using the mix of WEHG and BE that is EEF sets a level of an obligate WEHG ancestry that prevents their ANE being accurately represented.

    I may crosspost this to bioRxiv, to ask the study authors…

     
  8. Heraus

    December 27, 2013 at 12:00 pm

    About the French dot approaching Basque and “South French” (actually 7 individuals from Béarn) dots, I suppose it matches the aberrant individual from the famous Lyon sample that most studies use.

    You can indeed see that one of the individuals from the Lyon sample possess a clear SW French genomic identity (on the following graph, much more loaded in the pink component and nearly lacking the yellow component) :

    Graph

    We clearly need regional French samples, both ancient and modern.

     
  9. Maju

    December 27, 2013 at 4:49 pm

    @David:

    “The paper (…) gives a very clear chronology of how the European gene pool formed, including two sets of outliers, in the Northeast and Southeast”.

    As far as I can discern it is a mere speculative reconstruction based on nothing measurable nor measured.

    I'm not making this complicated: you are by appealing to non-measured nor measurable hypothetical components like “Basal Eurasian” and “Eastern non-African”.

    “WHG/ANE = Mesolithic foragers”

    WHG: (some) European Epipaleolithic foragers

    ANE: Siberian Upper Paleolithic foragers

    Two different things, two different origins and two different effects across the board.

    And then there were also pre-Neolithic foragers in West Asia, believe me. And they were ancestrally related to the other two groups. They just became farmers at some point.

    “ANE/WHG = Indo-Europeans”

    We do not know for sure. Was there not even a slight element of EEF or West Asian or Uralic admixture in them? In fact we know nothing of where the ancestors of Samara culture originated and how the previous archaeological layers evolved, but we know that the Uralic/Oriental genetic element was already around Eastern Europe in the Epipaleolithic: not just in Karelia and such but also in the much more influential area of Ukraine. Those mtDNA C lineages must have arrived from the Far East (almost certainly in relation with yDNA N1).

    “ENA/ANE = Uralics”

    That's the worst simplification of all. Obviously Uralics are a complex issue and West of the Urals they must have been growingly Europanized since very early. It is notable that the most distinctive marker of Indoeuropean-like autosomal influence in Europe peaks in and around Finland. It may be an artifact or whatever but all analysis since the long gone times of Cavalli-Sforza show that. Is that ANE? Probably not because ANE is quite homogeneous in all European populations. Is it related? Surely but rather obliquely.

    The ENA pseudo-component is not measured anyhow.

    “I don't see how you can argue with that if you read the paper and looked at all the supplementary data properly.”

    If I'm missing something, please point me to the relevant chapter and I'll check it. All I see is WHG (with the many doubts explained above), ANE (also with some doubts) and EEF. They also measured EEFs vs. BedouinB (and Lochsbour) but this comparison is full of issues.

    “It is what it is”.

    It is not the way you describe it in many ways. I don't know how you arrived to this oversimplifying dogma but it is completely wrong, precisely because your insistence on exaggerating the certainty of what you claim to be the truth, which is proper of a preacher, not a scientist.

    I know you can do much better than that dogmatic junk and I have no idea why you always seem to fail into such oversimplifications.

     
  10. Maju

    December 27, 2013 at 5:16 pm

    @Matt:

    The WHG vs EEF measure obviously can only measure (at best) those two European ancestral components (with the caveats I mentioned before). EEF is not and should never be read as West Asian, of course.

    However, according to their estimates EEF (multiple consistent formal tests rel. to Bedouin B), they'd be 60-90% West Asian, depending on the African admixture assumed for Bedouins (although they failed to assume African admixture in EEF, which is somewhat likely considering the the importance of E1b-V13 in all this issue).

    I don't know where you get your figure of 56% but it is below of what the authors estimated in the formal tests.

    “If using Stuttgart as a reference for early farmers and the early Near East, West European Hunter Gatherer ancestry in West Asian populations cannot go less than around 50% of their early farmer component!”

    Let's assume this you say is correct (unsure), what's the problem with that? Don't you see that Europeans invariably cluster separately from West Asians?, quite neatly so. Don't you see that WHGs cluster with Europeans and opposite to West Asians? So at least some WHG must be there.

    “This is a real problem since most West Asian populations probably don't have ANY real West European Hunter Gatherer ancestry.”

    I don't see why this is a problem: the study does not measure West Asians relative to their WHG component (or when it does, they fall off the margins, as expected – some Europeans do to, incidentally).

    “They're instead more likely to be almost wholly mixes of Ancient North Eurasian and Basal Eurasian”.

    “Basal Eurasian” is a chimera! (Chimera: mythical critter made of many different parts). You and David (and some others for what I read in his blog's comments) are reading too much in the hypothetical reconstruction graph, which is almost certainly wrong.

    This we can't discuss on the basis of this study only but it's fairly obvious in non-African genetics overall. West Eurasians are derived from Asians from beyond the Hindu Kush, with at most a tenuous remnant of West Asian OoA survivals. All West Eurasian lineages (excepted N1 and E1b on the yDNA side and whatever correlates within mtDNA) derive from the general Asian lineages: F on the yDNA side, N (and some M) on the mtDNA one. Overall the West Eurasian ancestry looks South Asian and to some extent probably also SE Asian (ancient Oriental).

    That's the true Basal Eurasian thing and is essentially indistinct from the Other non-African polarity (although Native Americans have some West Eurasian ancestry of the ANE type and South Asians were not measured).

    Once the West Eurasian expansion began some 50 or more Ka ago, you can forget about Basal Eurasianness: it's all West Eurasian (with the N1 and E1b partial exceptions). There's nothing (but those two minor elements) that does not originate in Europe, West Asia or Central Asia-West Siberia.

    So how can you negate the West Asian partial ancestry of EEFs? I don't see that being possible at all and must imply some sort of fundamental misunderstanding on your side.

    I don't think you measure Lezgins or Ashkenazim in the WHG-ANE-EEF triangle: they should fall off the edge because they are not essentially European. Just draw the “triangle” on the PCA: you cannot effectively measure anything that falls too much outside it. Can you tell me in which table are they measured such things?, because I believe there lays part of the confusion.

     
  11. Maju

    December 27, 2013 at 5:19 pm

    Looks like you're right, Heraus. Can you tell me if all the usual “French” sample is from Lyon? It's interesting to know that the “South French” sample must read as Bearn Gascons, because these invariably fall extremely close to Basques.

    “We clearly need regional French samples, both ancient and modern.”

    ABSOLUTELY!

     
  12. Maju

    December 27, 2013 at 6:18 pm

    An important clarification on the “Basal Eurasian” thing:

    In SI-10, where they ponder the West Asianness of EEF (Stuttgart), specifically in fig. S10.1 they use a highly technical concept of “Basal Eurasian”, which is nothing but West Eurasian with some African admixture. It is very clear in the model and it should be interpreted not only as effect of the remnant OoA elements but very especially as effect of the E1b influence in the Thessalian Neolithic and derived populations.

    In other words, for a founder effect reason, the European Neolithic component was more African-leaning than most West Asians from past and present. For that very reason it may present some confusing affinity (or quasi-affinity) to Palestinian and Bedouin populations, which almost without doubt were intermediate (in the Epipaleolithic?) for this African component in the European Neolithic, most apparent in the yDNA lineage E1b-V13.

    It's not a too real different composition of Neolithic West Asia but actually a founder effect in Thessaly and by extension in all Europe (more diluted, of course).

    This is also apparent in EEFs clustering best with Canarians, who have a good deal of North African aboriginal admixture, more than any European has.

     
  13. Gui S

    December 28, 2013 at 2:06 am

    The HGDP French sample was indeed collected in Lyon. But after looking at it very extensively individual by individual in various admixture runs, and comparing it with Dodecad and Eurogenes samples, I am quite convinced that it's a good representation of French diversity in general, with perhaps a slight Rhodanian bend. There is indeed an obvious Gascon, but also an individual with recent Ashkenazi ancestry, several who cluster close to the British Isles and appear to be from the Channel coast (from Brittany to Dunkerque), as well as some more clearly mediterranean. Comparison with French Canadian samples, whose ancestry is overwhelmingly Western French also makes the Lyon sample seem a good representation of French variation.

     
  14. Gui S

    December 28, 2013 at 2:06 am

    I am working on some maps of Europe for ANE, WHG and EEF for David at the moment, hope that can add some clarity to the numbers.

    I find myself wondering if there is a possibility that the variation between WHG and ANE was taking the shape of a West/East cline across Northern Eurasia, rather than that of completely distinct entities with a break off point somewhere East of Europe.

    It seems like the pattern of Siberians moving west in waves is quite ancient, almost akin to the meteorology of the Siberian high pushing cold air out every winter. Could it be that the following happened :
    -As a result of the Chinese Neolithic, Hunter-Gatherer yDNA N carriers and associates are pushed North of Manchuria into Siberia by the advent of yDNA O agriculturalists on their lands.
    -Thus interbreeding with and displacing (perhaps due to some technological advantage) the local Siberian “ANE” towards Western Eurasia.
    -As the Neolithic revolution is taking over Europe, the new “ANE” are kept at bay in far Eastern Europe and the Northern Slopes of the Caucasus (where they gain admixture from agriculturalists, thus creating the West Asian component we usually see) and adopt pastoralism at the margins of neolithic civilisation on the Steppe.
    -Something (perhaps a harshening of winters, probably the same thing that has also pushed Huns, Xiongnu, Mongols and Turkics out of their homeland) pushes a subset of yDNA N (most likely Uralic-speaking) carriers West of the Ural. Again interbreeding with and displacing the ANE-WHG Hunter-Gatherers from the Northeast European forest belt.
    -That last event must have upset the demographic balance in the area and probably caused our ancestrally Siberian ANE (but now quite admixed by local HG and some trans-Caucasian agriculturalists) Steppe pastoralists to move West into Europe perhaps bringing with them IE languages…
    -Later the same pattern repeats with Huns, and Turkic speakers, spreading West, displacing and interbreeding with populations in their wake. If we compare the levels of Siberian admixtures in Turkey to the levels of “ANE” admixture in Europe, things seem quite comparable and can even make place for a more ancient “ANE”-like presence in Europe as part of a Eurasian paleolithic cline…

    It also explains the difference between the Caucasus and “Gedrosian” components we've seen regularly pop up. With Gedrosian being an admixture of Middle Eastern agriculturalists and “ANE”-like people from the Himalayan slopes (another place where Turkic speakers would later come to settle) and Caucasus being an admixture of Middle Eastern agriculturalists with drifted and WHG-admixed “ANE”-like people in the Caucasus.

    This leads me to think that contemporary Turks cannot be a good proxy for Neolithic Anatolians. I believe that what comes up as “West Asian” (as opposed to “Southern”/”Middle Eastern”) in admixture tests in Turkey is the result of a more recent flow into Anatolia of ANE-admixed agriculturalists from the Caucasus (perhaps IE speaking, but also likely to speak Caucasus languages, depending on their level of ANE-admixture) back into Turkey (and incidently Greece).

    Although your idea of African-admixed yDNA E1b Thessalians being the cause of that “Southern” component as opposed to the “West Asian” component being much more ancient in West Asia is quite tempting. I am going to reflect on it, as it brings an opposite vision to the model I have in mind.

     
  15. Gui S

    December 28, 2013 at 2:07 am

    I might also make some maps to clarify the model I have in mind, writing it down might make it a bit obscure. I would love to hear what you think, or how it matches with archeological record.

     
  16. Maju

    December 28, 2013 at 2:53 am

    Nice to know that for a fact, I wasn't sure if it was the same sample.

    I would not say that they reflect well French variation because for some reason they seem clearly scored northwards. This should be obvious when we compare this sample with the Bearn sample. They may reflect well the variation of Paris but what about Gascony, Brittany, etc.? France is a very large state which has played a key central role in European prehistory and history. It still does, although in the Ice Age it was much more important, of course.

    Something I often complain is that many samples from small ethnicities like Lithuanians, Slovaks, Basques, Sardinians, Finns, etc. may overshadow the real structure of Europe because the really big states, which hold the bulk of European population are seldom sampled regionally nor represented according to their real (or historically realistic) weight. You can't (shouldn't) just compare in most cases 20 French and 20 Lithuanians, because it's 70 million French and 3 million Lithuanians. You can compare 20 Lithuanians with 20 Basques but better to compare 20 French (representative of all the territory) with 20 or 40 Russians (also territorially representative) in order to get a realistic structure of Europe in the autosomal aspect.That's a very neglected aspect, very especially in the case of France.

     
  17. Maju

    December 28, 2013 at 3:38 am

    “I find myself wondering if there is a possibility that the variation between WHG and ANE was taking the shape of a West/East cline across Northern Eurasia, rather than that of completely distinct entities with a break off point somewhere East of Europe.”

    Impossible to say without aDNA from Eastern Europe and the like. But in any case WHG and ANE are defined for being different clusters of Paleolithic sequences, so the distinction is clearly there.

    “It seems like the pattern of Siberians moving west in waves is quite ancient”…

    Epipaleolithic at most, I'd say. Earlier the flow seems rather from West Asia (and maybe in some cases also Europe) into Siberia. The only doubt I have is regarding Kostenki and its U2 sequence, which may indicate one such early minor flow from Central Asia.

    The yDNA N1 flow is almost certainly from Epipaleolithic onwards in the case of Europe and it came via the taiga area, which was earlier (Ice Age) almost certainly uninhabited. Rather than pushing peoples westwards it seems that the N1 clan went around the pre-existing peoples through the Far North. Their distribution does not suggest that they pushed others around, at least not often.

    It's simpler to think that most ANE advanced with Indoeuropeans but then, why don't we see a stronger gradation? Maybe the lower frequencies are just an artifact? Maybe the ANE component partly reflects the Eastern European influences in Siberia, such as Gravettian? In this last case it's possible that the basic levels of distributed European affinity are mostly reflection of the partial Europeanness of ANE peoples themselves. Can't say with any certainty.

    … ” contemporary Turks cannot be a good proxy for Neolithic Anatolians.”

    Better than Bedouins for sure. Anatolia has suffered several invasions since the Neolithic but it was also always a densely populated area, so the impact of each of the invasions should be weak, although cummulative. However the invasions arrived from different corners, so the accumulation of weak effects should cancel each other to some extent.

    The Caucasus-Gedrosian component makes all the sense to me anyhow: it's Zagros Neolithic, as opposed to Eastern Mediterranean Neolithic (which came in two flavors: Palestinian and Anatolian). Today we'd say “Iranian” but then it was more like Elamites, Hurrians, Sumerians and such. West Asia was not homogeneous in the Neolithic but it had at least those three distinct regions: (Southern) Anatolia, Palestine (and some other Levant) and Zagros (with expansion to Mesopotamia), they met around Gobekli Tepe (although in origin this megalithic site was still Mesolithic). Then Elamites were clearly a different group than Sumerians and there were probably other ethnic nuances like the difference between PPNB and the CAPC. We can't risk oversimplifying Neolithic West Asia: it was a complex of several cultures. Some of those distinctions we still see today quite neatly, like the two quite apparent West Asian autosomal components (N & S), related to the geographical cline between yDNA J2 and J1, which is almost certainly from the Upper Paleolithic. They overlap but the duality persists.

    “Although your idea of African-admixed yDNA E1b Thessalians being the cause of that “Southern” component as opposed to the “West Asian” component being much more ancient in West Asia is quite tempting”.

    Thanks. I find easier to think in terms of founder effect in this case, rather than imagining mass replacements in West Asia after Neolithic. After all lineages like yDNA J2 do not come from outside that region.

     
  18. Maju

    December 28, 2013 at 3:39 am

    Sure. I'd like to see those maps in any case.

     
  19. Maju

    December 28, 2013 at 10:12 am

    Just saw them. Nice maps.

    I have some lesser concerns like moving Lyon to Central France (not the true data point) or the equally dubious ref. for England near York (shouldn't be further south, maybe in London?) Does somebody know if the HGDP Spanish sample is from Madrid or Valencia or what? Given the size of these states, different positioning to their true geography may introduce important errors in the clines.

    Also the lack of scale may lead to the wrong assumptions about the meanings. For example, in the ANE map Sardinia appears as “white”, what may be understood as zero, when it is in fact 4.1%. In that same map the differences appear very much aggrandized: even if most populations range very close (12-18%) in absolute values, they are shown with very different colors. Personally I'd use just four gradients: <5%, 5-10%, 10-15% and 15-20%.

    Otherwise it seems a good reproduction of the apportions suggested in the paper, which, as I mentioned before I do not agree with too much, especially in the use of Lochsbour as proxy for WHG without further control testing. But those problems are from the paper, not yours.

    As for the color tripartite map, it really ends up looking bipartite. I would suggest using a two color WHG-EEF scale and marking the presence of minority ANE with some other means than a third color that ends up looking pretty much invisible, for example dotted lines (a single dotted line indicating the <10% ANE pockets should suffice).

    Self-reference for further discussion:
    http://i40.tinypic.com/2dt8zmb.jpg (WHG)
    http://i44.tinypic.com/9gzvyc.jpg (EEF)
    http://i42.tinypic.com/bjd3sn.jpg (ANE)
    http://i40.tinypic.com/9tlw76.jpg (color composite)

     
  20. Kristiina

    December 30, 2013 at 9:15 am

    “Rather than pushing peoples westwards it seems that the N1 clan went around the pre-existing peoples through the Far North.”

    I think that they took different routes (N1b took a more northern route than N1c-E) and mixed with people carrying R1a and I1 lines. Within N1c there are at least two different flows to the west, earlier one and later one that affected only the Urals area. It is also possible that the Balkanic N is the earliest N in Europe that reached Europe first.

     
  21. Marnie

    December 31, 2013 at 10:32 am

    🙂

    Someone needs to hand these authors a map of Europe during ice ages:

    http://www.ngdc.noaa.gov/mgg/topo/pictures/GLOBALsealeveldrop110m.jpg

     
  22. Grey

    January 2, 2014 at 3:18 am

    I do think the LGM is big in this.

    imo

    1) Hyperborean parental population in a crescent shape from europe through north asia to siberia.

    2) Hyperboreans split into at least two child populations (and pushed south) by the ice, those child populations spreading out north again afterwards – i'll call them WHG and EHG.

    3) First farmers somewhere in eastern Anatolia
    3a) Small numbers of first farmers move north towards the steppe and mix with the EHG to eventually become the Kurgan guys.
    3b) Much larger numbers of first farmers move west into central Europe forming LBK mostly displacing the WHG to marginal terrain and the peripheries where they undergo a HG to herder transition.

    4) Something weakens the LBK leading to
    4a) firstly a herder over-run from the west into the center (funnelbeaker)
    4b) a second herder over-run this time from the east into the center (unetice)
    4c) bell beakers involved in this process as well but as a caste of miner/metalworkers like the african blacksmiths rather than as full displacing type populations

    5) All four populations WHG/EHG, Bell Beaker, LBK survivors merge in central europe to become the ancestors of Hallstadt, La Tene etc spreading out from there in multiple directions.

    Leading to
    Western Europe: WHG + first farmers + EHG (in indo-european form)
    Eastern Europe: EHG + first farmers
    South Eastern Europe: first farmers + WHG

     
  23. eurologist

    January 2, 2014 at 1:37 pm

    Marnie,

    Yes,
    As I have mentioned numerous times before, I would like someone to explain to me how SE Europeans could have been substantially different from W Anatolians, at the beginning of the Neolithic. It is virtually impossible.

     
  24. Marnie

    January 2, 2014 at 10:24 pm

    Agree, especially since the Bosphorus is estimated to have opened only about 7,600 years ago.

     
  25. Maju

    January 2, 2014 at 10:30 pm

    What do you mean by “substantially different”? 8800 years ago (or 8500 with pottery) the evidence of agricultural settlements in Western Anatolia is as thin as a graphene layer (effectively zero). See: → http://context-database.uni-koeln.de/index.php

    Most relevant maps:
    http://context-database.uni-koeln.de/img/P4B.jpg
    http://context-database.uni-koeln.de/img/P4C.jpg

    There's a major gap between that area (Fertile Crescent) and Thessaly. It may be because of gaps in our knowledge but the gap is clearly there and recent research in Thrace and such has not changed the situation (Thessaly is still oldest farmer country of Europe and many hundred kilometers west of the nearest Anatolian (Cappadocian) site.

    It is very possible that Thessalian Neolithic was seeded by a coastal migration with peculiar founder effects (notably E1b-V13) which seem original from further south.

    The Ice Age sea levels do not matter, first because the Ice Age was long gone, second because with or without Bosporus/Dardanelles that area was always a very practicable crossing point but also a funnel.

    The problem is not what you can imagine from physical maps, but what we can understand from actual human socio-cultural geography.

     
  26. Maju

    January 2, 2014 at 10:35 pm

    I don't understand much of what you say here, beginning with the totally misleading anachronistic mythical term “Hyperboreans”.

    In Central Europe there must have been important Chalcolithic flows from the West in any case. Indoeuropeans alone cannot explain the current genetic pool. Maybe you should start thinking in other mythological terms: “Atlanteans”. Just a guess.

     
  27. andrew

    January 3, 2014 at 1:00 am

    Is there some reason, such as French specific privacy laws for the poor sampling in France?

     
  28. Maju

    January 3, 2014 at 2:06 am

    I know mostly from our Gascon friend Heraus but not enough to discern the exact legal and bureaucratic obstacles. As you may know several countries such as India forbid the export of human DNA and this may be the core of the French legislation.

    But French red tape can be utterly annoying, more so as there are nationalist ideological reasons to oppose the “butchering” of the Republique on regions and ethnicities. French national ideology is quite extremist in opposition to all forms of difference: everyone is equally French, no peculiarities allowed or almost. The few studies performed have suggested a strong diversity within the Hexagon and that's something that many may feel threatens the artificial essence of France as unity. At least that's the impression I have.

    An example of utterly annoying red tape for no reason in France was the recent long struggle of organic farmers to get nettle purine legalized again as fertilizer. The state had arbitrarily decided to forbid it and was stubborn in that position for many years. It was eventually legalized again after a long environmentalist struggle but there seems to be no such impetus re. human population genetics.

    So researchers and private companies alike seem to face all kind of legal trouble but I don't really know the details in depth, just Heraus complaining all the time. Ask him.

     
  29. Maju

    January 3, 2014 at 2:13 am

    “… the Bosphorus is estimated to have opened only about 7,600 years ago”.

    Even if true (under some debate), it would not have affected the Aegean significantly nor the ability to cross between Anatolia and Thrace. The Marmara Sea already existed and crossing those extremely narrow straits can be done even with Neanderthal navigation skills very easily. It's not any kind of natural barrier at all and there is no archaeological evidence of people crossing through that path either.

     
  30. Grey

    January 4, 2014 at 9:58 pm

    “I don't understand much of what you say here”

    I think WHG and ANE will turn out to be WHG and EHG i.e. both descended from the same ANE population split by the ice.

    “In Central Europe there must have been important Chalcolithic flows from the West”

    That's the other part i'm saying.

    “Maybe you should start thinking in other mythological terms: “Atlanteans”.”

    Yes, Funnelbeakers and Bell Beakers imo.

     
  31. Maju

    January 5, 2014 at 2:50 am

    “I think WHG and ANE will turn out to be WHG and EHG i.e. both descended from the same ANE population split by the ice.”

    As I understand it, they both descend (in essence) from from the same West Asian population but ANE went to Central Asia and Siberia, and WHG ancestors went to Europe. Eastern Europe could have been settled from both origins but the main origin should be from Europe.

    Anyhow your hypothesis does not explain why the ANE affinity is homogeneous across the European and even all the West Eurasian geography, much more than the WHG one. No numerical data is provided for West Asia but, judging from the colors in the Extended Data Figure 6a, only some populations in Palestine and Yemen are at the relatively low levels of ANE affinity of Sardinians (when compared with EEF). Map EDF 6c shows that, when compared with Lochsbour, West Asians are extremely much more akin to ANE than Europeans (logical because their negative affinity is only rel. to Lochsbour WHG, not ANE).

    The study does not attempt (nor certainly can) study ancient admixture in West Asia (no local HG samples) but something that becomes clear in these visual analysis is that, while WHG defines Europe vs. West Asia (maps B and C), ANE and EEF affinities do not.

    We would probably need to make more extensive comparisons, with more ancient and modern samples (missing North Africa, Siberians), before we can reach to any well structured conclusion but, whatever the case for me it is clear that ANE has no particular preferential relation to Europe when considered vs. West Asia and hence it must be considered a population of general West Eurasian stock (with some Far East admixture probably), neither more akin to Europe nor to West Asia, but specifically Central Asian instead.

    “Yes, Funnelbeakers and Bell Beakers imo”.

    Vale. I could not understand that in your previous comment.

     
  32. Grey

    January 5, 2014 at 7:32 am

    “for me it is clear that ANE has no particular preferential relation to Europe when considered vs. West Asia”

    I agree. I think it will turn out to be ancestral over a very broad crescent shape stretching from Europe to Siberia and then became even more broad after that population was pushed south by the LGM.

    However what i'm guessing is ANE only *appears* to be more eastern because when the first farmers expanded out of their epicenter larger numbers (as a proportion) moved into west and central europe than the east.

    so
    Stage 1) ANE
    Stage 2) ANE -> WHG & EHG
    Stage 3a) WHG & lots of first farmers
    Stage 3b) EHG & far fewer first farmers

    (Quite possibly wrong but hopefully clearer at least.)

     
  33. Grey

    January 5, 2014 at 9:11 am

    More wild speculation and i apologize if it's annoying but one of the things that most struck me when i first started getting into this is how the R1b distribution looked like a blood splatter diagram with the spray pattern centered somewhere on an intersection south of Ireland and west of Brittany.

    If you drew a vertical line down from the center of Ireland and a horizontal line across from the center of Brittany the intersection would pretty much sit right on the LGM coastline.

    http://www.ngdc.noaa.gov/mgg/topo/pictures/GLOBALsealeveldrop110m.jpg

     
  34. Maju

    January 5, 2014 at 12:23 pm

    WHG individuals seem less leaning to ANE than modern Eastern and even NW Europeans in the PCA, so if anything there's been some increase in the influence of ANE after the Epipaleolithic, not the opposite. ANE reflects an intrusive element in Europe just like the West Asian component (partly reflected in EEF).

    Additionally EEFs seem a simple West Asian + WHG mix, with negligible ANE relevance, so the excess ANE almost certainly arrived after Neolithic, what should be read as Indoeuropean influence. This seems corroborated by the fact that, other than Sardinians, Basques are the lowest scoring in ANE affinity (barely 10%, half that of Estonians).

    The question that seems to fuel your speculations may be the fact that Sardinians score even lower than Basques (less than half in fact) and that Sardinians seem a good proxy for EEFs. I'm not really sure how to explain this but either Basques have got some IE-like influence or EEFs scored particularly low in ANE affinity, as do for example Palestinians or Yemenis today. That's why the authors use the “Basal Eurasian” conjecture, I guess, but I'd rather explain it with North African admixture, something that seems very obvious in the fact that EEFs were vector for the spread of Y-DNA E1b-V13 in Europe.

    This “unexpected” African affinity of EEFs subtracts from basal ANE affinity, which should be otherwise relatively neutral for all West Eurasians, with lesser clines relative to geography or whatever. Alternatively or complementarily, some of ANE-affinity weight could reflect Gravettian roots, shared by all European HGs (West and East equally), as well as Siberian ones but whose impact in West Asia may have been less uniform.

    In other words: variable affinity with ANE does not need to reflect immigration from Siberia, it may reflect relations with peoples more or less akin to Siberian HGs since the beginning. Lower ANE affinity can for example be caused by greater North African affinity, while greater ANE affinity can be caused by mere Eastern European (Indoeuropean, Uralic) flows.

    “… hopefully clearer at least”.

    I understand what you suggest but not why. There's no reason to think of flows from Siberia into Western Europe in the Paleolithic, none at all. Aurignacian and Gravettian have West Eurasian roots, while post-LGM cultures are all locally rooted in Western Europe. If there was some admixture with actual ANEs in the Paleolithic, it must have been an Eastern European phenomenon only (and Eastern Europe did not influence Western Europe back then).

    So I understand that the issue is best explained by “anti-ANE” African (or just Palestinian) extra affinity in EEFs and by later “pro-ANE” IE (and Uralic) migrations.

     
  35. Grey

    January 5, 2014 at 11:43 pm

    “There's no reason to think of flows from Siberia into Western Europe in the Paleolithic”

    Yeah i'm thinking the opposite. I think there were multiple routes out of Africa and one (not the main one) was Atlantic coastal: west coast of Africa – Iberia – Western Europe – Scandinavia – North Eurasia – Siberia in a crescent shape.

    I think that base population was split into two segments by the LGM (and pushed further south including into Caucasus / West Asia) and that some of the eastern half of that segmented original population came back west later as part of the Indo-European thing (but some of it was already there).

     
  36. Maju

    January 6, 2014 at 4:55 am

    “I think there were multiple routes out of Africa”…

    Almost certainly not.

    “and one (not the main one) was Atlantic coastal: west coast of Africa – Iberia – Western Europe – Scandinavia – North Eurasia – Siberia in a crescent shape.”

    Nonsense. Don't get offended but it's total unfounded nonsense. There are so many wrongs in that idea that I don't deem it worth explaining them.

    But I'll tell you something about maps: they are not the reality. For example Marnie's Ice Age map shows reasonably well the extension of the coasts in the LGM (not before nor after, just around that cold peak) but it does not show temperatures, ice shelves (covering most of the route you want to imagine) or Neanderthal presence.

    “I think that”..

    “I think” is not a good explanation, sorry. “I think” without factual support is not better than the “I believe” of some poor soul brainwashed by preachers.

    Science works the other way around: (1) let's see what we can get to know, (2) let's put that together in one or several most parsimonious and coherent explanations, (3) let's contrast them with the facts, (4) let's continue researching…

    There are absolutely no facts in support of your conjecture. You have provided not one, just “I think that”…

     
  37. About Time

    January 6, 2014 at 3:36 pm

    Hi Maju, Nothing to add about Paleolithic stuff (mad props to you for synthesizing and presenting so much Paleo information over the years – we are all still miles behind your knowledge and there are no substitutes).

    My one quibble is that yes, Palestinians and Yemenis score low in ANE relative to Europeans. But, these pops are much closer to ANE than they are to WHG, so there is some kind of relationship (maybe indirect somehow in terms of what's modeled in the paper?).

    Also, Iranians and Lezgins score especially high in ANE vs WHG ratio (like Palestinians/Yemenis but stronger), in addition to strong Z scores for direct MA1 admixture (unlike Palestinians/Yemenis).

    It's shown in the Extended Data Figure 6(C) at http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-1.pdf

    I agree that ANE looks like something non-European that came back into Europe and especially northern areas.

     
  38. Maju

    January 6, 2014 at 7:52 pm

    “… Palestinians and Yemenis score low in ANE relative to Europeans. But, these pops are much closer to ANE than they are to WHG, so there is some kind of relationship (maybe indirect somehow in terms of what's modeled in the paper?)”.

    Good question. I think that it simply reflects relative affinity to either. As West Asians have no WHG ancestry, they score comparatively high in the opposite vector, which is similarly akin to all. So basically Europeans look “dark blue” (meaning strong WHG affinity rel. to ANE) and West Asians look “yellow” (meaning not strong affinity to either). Notice that the highest value is still negative (-2), so they are probably slightly closer to WHG than ANE in fact. Similarly all populations are closer to EEF than to ANE but more or less so (map A, same scale).

    Instead in map B there is positive score to WHG (+4), what means that some populations are more akin to WHG than to EEF.

     
  39. Marnie

    January 6, 2014 at 8:16 pm

    “So I wonder: is it possible that these two WHG cluster-lines represent derived ancient branches from an original population of SW France. We know that since the LGM, the area of Dordogne (Perigord) was like the megapolis of Paleolithic Europe, with population densities that must have been several times those of other areas.”

    More “reindeer games”, Luis.

     
  40. Maju

    January 6, 2014 at 8:26 pm

    Meaning what?

     
  41. Grey

    January 7, 2014 at 5:07 am

    “but it does not show temperatures, ice shelves (covering most of the route you want to imagine) or Neanderthal presence.”

    People moved out of Africa before the LGM so the route I am imagining is before the things you mention blocked the route. The question then is what effect would the expansion of the ice have on that *pre-existing distribution* i.e. could a related population have been split into separate refuges and pushed south and secondly what re-expansion paths existed after the LGM peak when the sea levels rose.

    “Science works the other way around: (1) let's see what we can get to know”

    I accept that but an Atlantic coastal route which was largely restricted to the coast has the problem that the old coast is now underwater so any evidence would have to be looked for where the coast *was*.

    “There are absolutely no facts in support of your conjecture.”

    Sardinia and Scotland. They don't support my conjecture but they don't support the ANE as Siberian either.

    #

    (On a completely separate note could the Slavic slave trade from the Crimea have effected the ANE proportions around the eastern med.)

     
  42. Maju

    January 7, 2014 at 6:27 am

    “People moved out of Africa before the LGM so the route I am imagining is before the things you mention blocked the route”.

    But then also the coasts would be similar to modern ones. Ice sheet and sea level are inversely correlated.

    … “the problem that the old coast is now underwater so any evidence would have to be looked for where the coast *was*”.

    Two major problems:

    1. As mentioned above, if there was no ice sheet, the coast was also higher, similar to modern, maybe even further inland.

    2. Why would people so strictly adhere to the coast. We should expect some groups traveling further inland or even the same groups moving camp seasonally or whatever. They must have left some sort of evidence in any case and nope.

    “Sardinia and Scotland. They don't support my conjecture but they don't support the ANE as Siberian either”.

    ANE is Siberian by definition: it is defined by a Central Siberian specimen (MA-1) or at most by two (incl. AG-2, not really used in the tests). Their genetic influence in the West is very homogeneous (with very few exceptions all populations range ~15%, 3 points up/down) and that homogeneity includes West Asia or most of it.

    IMO at least part of that affinity is mere default affinity because of shared origin in Aurignacian/Gravettian. The “negative affinity” is what needs most explanation here and IMO this is best explained by (North) African influence or, if you wish, as the authors hint to albeit in imperfect manner, by a “Basal Eurasian” element (or Arabian/Palestinian OoA remnants, which would tend to Africa in any comparison).

    Speculatively it is possible that the Hamburgian-Ahrensburgian late UP population of the North Sea had slightly more ANE component than Magdalenian ones. That seems apparent from the position of Scandinavian Epipaleolithic samples in the PCA but that would be about it: less important variation because of ancient founder effects. The origin of Hamburgian is not well understood, so it's not impossible that they had some obscure Eastern influences.

    “… could the Slavic slave trade from the Crimea have effected the ANE proportions around the eastern med”.

    You never know but I doubt it is a major element. It rather seems to me that the SE Europeans go secondary influences from Eastern Europe or West Asia after the initial Neolithic, slanting their affinity. Alternatively the SW European (Cardial) Neolithic had important founder effects affecting their overall ANE affinity (or maybe both).

    The reality is that we have too few references for deep judgment. We would need more ancient samples, preferably more relevant ones. All we have now is a generic draft and many open questions. It's better than nothing but clearly not the last word on the matter.

     
  43. Maju

    January 7, 2014 at 11:47 am

    That's a very complex speculation with many issues.

    First, the paper is PPV, so I do not know where and very especially WHEN the reindeer remains belong to in the “Eastern Cantabrian Region” (i.e. Southern Basque Country). If you know the dates, you don't provide them in the entry. I would imagine for what I know that the reindeers appeared in the coldest spell, i.e. around the LGM, not much before and not much later. In general the evidence rather supports other foods and totemic animals (first: deer, later: bison) as dominant, so the presence of reindeer in the LGM probably has little relationship with the Epipaleolithic remains discussed here and in your post.

    On the other hand it is possible that it holds some relationship between core Solutrean in the Franco-Cantabrian region in the LGM and apparent offshoots in Hungary and later Poland (i.e. maybe some FCR hunter-gatherers migrated to Central Europe before Magdalenian was consolidated and maybe these had some symbiotic relationship to reindeer or other tundra-like ecology.

    But since then what we see in the FCR and surrounding regions is:

    1. Development and expansion of Magdalenian (seemingly more oriented to aquatic resources and with important totemism of the bison).

    2. Development of early “Epimagdalenian” cultures such as Azilian (Laminar Microlithism).

    3. Possible migrations from roughly “Northern France” with the Geometric Microlithic industries (Tardenois).

    I would emphasize the latter because it may explain the alleged “Nordic” genetic affinity of La Braña, which in some aspects look immigrants (notably deer fang ornaments). Then the Neolithic arrived (essentially contemporary of La Braña site, so maybe they were being pushed around by Neolithic expansion as well).

    Finally I see no particular reason for your “transposition” of the axes, if anything the need of such transposition in your graph says that your conjecture is wrong, because you need to alter the data in order to reach to your conclusion, what, let's be frank, it's outright cheating. Suddenly, with no other explanation than your subjective need to arrive to certain conclusions (wishful thinking), you imagine axes-clusters that are not in the data but which you treat as if they were real. It doesn't make any sense.

    I'm not sure if I am onto something, as you say, with the issue of the axes (I clearly expressed that I was being highly speculative and had all kind of doubts), but certainly not what you imply after all those arbitrary “transpositions”. You have taken my speculation and bastardized it beyond recognition. The axes converge where they do, not where you wish them to. If that convergence means anything at all or is just a meaningless artifact, I can't say at this point but what I'm certain is that your “transposed” axes only exist in your imagination.

    I feel a bit outraged, I must say.

     
  44. Grey

    January 7, 2014 at 1:02 pm

    “ANE is Siberian by definition”

    Yes, my mistake. I meant “solely” Siberian.

    .

    “Why would people so strictly adhere to the coast.”

    As climate zones and ecosystems are partially latitude driven then a population adapted to one latitude band that wants to expand north or south has to adapt to the different climate zone whereas a population who are already adapted to coastal food-getting require less adaptation to move north or south as long as they stick to the coast.

    I don't want to get too hung up on the OOA aspect. The main thing is i think it will be a general case that expansion along coasts will be faster than inland if the inland route crosses climate zones or obstacles because a coastal sea food getting population will have to adapt less.

     
  45. Maju

    January 7, 2014 at 1:24 pm

    Well, whatever the case, you are speculating with a totally invisible population for dozens of millennia, just because. I can't but dismiss that as the wildest baseless speculation, sorry.

     
  46. Marnie

    January 7, 2014 at 4:34 pm

    Hi Luis,

    Gotta go to work, but will respond tonight.

    🙂

    Marnie

     
  47. Marnie

    January 8, 2014 at 6:39 am

    Swimming Cows, Last Glacial Maximum Parisien Reindeer . . . what will she think of next?

    🙂

    I added some references to my blog post. You can check them out if you would like. I did skim through them and it does appear that, more or less, the Paris Basin was continuously occupied by reindeer hunters during the Last Glacial Maximum.

    “On the other hand it is possible that it holds some relationship between core Solutrean in the Franco-Cantabrian region in the LGM and apparent offshoots in Hungary and later Poland (i.e. maybe some FCR hunter-gatherers migrated to Central Europe before Magdalenian was consolidated and maybe these had some symbiotic relationship to reindeer or other tundra-like ecology.”

    A quick look at various papers seems to support this.

    Regarding bison, horses and other animals, its definitely possible. However, only reindeer hug and thrive on and near glaciers.

    “Finally I see no particular reason for your “transposition” of the axes, if anything the need of such transposition in your graph says that your conjecture is wrong, because you need to alter the data in order to reach to your conclusion, what, let's be frank, it's outright cheating.”

    No. The ancient DNA data needs to be transposed because it hasn't been Neolithically shifted like most of the other samples. During the Neolithic, most of the other modern DNA samples have genetically shifted toward Greece and Anatolia. So to put the ancient DNA in the correct geographic position, it has to be “Neolithically transposed.”

    I say most, because, probably, certain groups such as the Basques have shifted less. So maybe we have to transpose the Basques along with the ancient DNA. That would put the Basques in the correct geographic position. I'm not sure about this, but I think the shift away from Anatolia for certain genetically isolated groups should be pointed out.

    If that were the case, then with transposition, the Basques would begin to nicely line up with the La Brana-Loschbour axis. Luis, I always new you were meant to be a reindeer hunter.

    “You have taken my speculation and bastardized it beyond recognition.”

    I haven't done anything, just drawn a few lines on the plot, tried to reconcile the ancient DNA with their actual geographic positions and recent archaeological findings.

    While you're being outraged, I'm going to try to read the details in some of these papers.

     
  48. Maju

    January 8, 2014 at 8:33 am

    “… only reindeer hug and thrive on and near glaciers”.

    Maybe but, other than mountain areas, there were no glaciers over here even in the LGM. In fact the usual maps of glaciation typically show the approx. max. extension of ice at the LGM, which reached to central England, all Poland and Germany East of the Elbe – also the Alps. Even in the Seine area there were no glaciers, although for what I know it was a arid steppe-tundra environment, not too good for human inhabitation, which concentrated in the loess steppe-tundra ones in the northern latitudes of Europe (Rhine-Danube and Dniepr-Don areas). Maybe the reindeer were particularly able to exploit such dry environments (and surely many others in that cold spell).

    I posted some good quality LGM ice cover maps in this entry: → http://forwhattheywereweare.blogspot.com/2012/03/ice-free-pockets-in-ice-age-scandinavia.html (originally from Don's Maps: http://www.donsmaps.com/icemaps.html).

    In pop culture the Ice Age looks like Antarctica but that's not an accurate notion at all. It should look more like, say, Russia, with greater dominance of an ecosystem now rare known as steppe-tundra (→ http://www.esd.ornl.gov/projects/qen/carbon11.html). This ecosystem was harsh but had variedly degrees of habitability judging from actual human presence:

    1. Dry steppe-tundra (Seine basin, Balcans): almost uninhabited
    2. Loess steppe-tundra (Rhine-Danub, Dniepr-Don): inhabited at low densities
    3. Mixed ecology (Franco-Cantabrian region): densely to very densely inhabited

    Additionally the Mediterranean was dominated by forests (European trees found refuge there), what apparently limited inhabitation, although this one was highest in the LGM at least in Southern Iberia (Gravetto-Solutrean).

    “No. The ancient DNA data needs to be transposed because it hasn't been Neolithically shifted like most of the other samples”.

    What?! That's precisely why we can compare it with modern samples. Some are closer and others farther to either WHG or EEF. You get the modern samples and compare aDNA with these (PCA format) or you get some of the aDNA and compare moderns with them (formal tests). The results are similar for the WHG/EEF bipolarity in both procedures.

    “During the Neolithic, most of the other modern DNA samples have genetically shifted toward Greece and Anatolia.”

    That is not what EEFs say. Something great about this study and the PCA in particular is that we do have actual ancient farmer DNA (several samples) and they look more like Canarians or Sardinians than anything Anatolian.

    In fact, when compared with WHGs and modern West Asians (no ancient autosomal DNA from West Asia yet), they do not tend to Anatolia but rather Palestinian/Bedouins. Why? The authors used the “Basal Eurasian” hypothesis to explain it but I actually think it has to do with the peculiarities of the Thessalian Neolithic and its African genetic affinity (most obvious in Y-DNA E1b-V13 but maybe also in G2a, which is important in parts of Palestine).

    Whichever explanation you prefer, EEFs do not tend to Anatolia, so the Anatolian tendency of modern SE Europeans (Balcans and Italy) needs of another post-Neolithic explanation (maybe a combo of Halaf-like Grey Pottery “Pelasgians” and Indoeuropean Chalcolithic invaders from Eastern Europe).

    “I say most, because, probably, certain groups such as the Basques have shifted less.”

    Not just Basques but also Iberians, Sardinians, Gascons… everything SW European seems still strongly attached to that first EEF wave (with variable WHG admixture).

    To do what you want you would have to, first of all, shift EEFs, who do not show any “Anatolian tendency”. But what you actually need to shift is your Anatolian origin paradigm for the European Neolithic, which is probably wrong, at least partly wrong.

    “I haven't done anything”…

    Oh yes, you did! You are just imagining things, very wildly so.

     
  49. Marnie

    January 8, 2014 at 9:06 am

    Here's an up to date map of Europe at the LGM:
    http://www.diercke.com/kartenansicht.xtp?artId=978-3-14-100790-9&stichwort=modern%20man&fs=1

    Regarding EEF, Sardinians, Iberians, Gascons, yes, the are somewhat isolated, like Basques.

    However, much of the Neolithic came from Anatolia and the Danube (not Sicily).

    “Oh yes, you did! You are just imagining things, very wildly so.”

    Must be my hunter-gatherer ancestry, but in any case, I am going to read the reindeer papers.

     

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
%d bloggers like this: