Wednesday, June 10, 2015

10% of Bronze Age Europeans Lactase Persistent


[Lactase Persistence] we find it at most at low frequency in the Bronze Age (10% in Bronze Age Europeans; Fig. 4), indicating a more recent onset of positive selection than previously estimated
Really?  Mathematically, how does this work?  Everyone on planet Earth should be Lactase Persisent using this weird logic, really weird logic.

Were talking about about 400 generations on one hand, on the other hand a crises in the sustainment strategy or infant rearing customs so great as to winnow lactose intolerance almost completely from some parts of Europe (ie. the Bell Beaker parts of Europe).  There simply was not any event or factor that contributed to the modern frequency other than migration, founder effect and the growth of that particular population.

Really, there are two arguments here.  One is positive selection on steroids, which is ridiculous given the fact that non-LP people seem to be having plenty of babies.  The other is continuous or severe culling, and that just doesn't seem to fit anything that is known of the Bronze Age forward.



Blah, blah, blah
"The results for rs4988235, which is associated with lactose tolerance, were surprising. Although tolerance is high in present-day northern Europeans, we find it at most at low frequency in the Bronze Age (10% in Bronze Age Europeans) indicating a more recent onset of positive selection than previously estimate... Our results confirm a low frequency of rs4988235 in Europeans, with a derived allele frequency of 5% in the combined Bronze Age Europeans (genotype probability>0.85)...  Among Bronze Age Europeans, the highest tolerance frequency was found in Corded Ware and the closely-related Scandinavian Bronze Age cultures.  Interestingly, the Bronze Age steppe cultures showed the highest derived allele frequency among ancient groups, in particular the Yamnaya, indicating a possible steppe origin of lactase tolerance."
Algerian herd-women putting up a tent.  What!?

You might remember this "Guess What!? LP Allele Not Under Selective Pressure"  Obviously, the derived allele comes from Central Asia, probably in the Paleolithic.  I would take a wild guess that LP is associated with the earliest users of lactating domesticates, possibly the nanny goat for babies and cow's milk after 1 year of age.


Bronze Age population dynamics, selection, and the formation of Eurasian genetic structure, Allencroft et al.  Nature 522, 167–172 (11 June 2015) doi:10.1038/nature14507 [Link]

56 comments:

  1. To clarify, the very few Beaker samples so far are from Germany. The oldest LP detected to date is from Quedlinburg

    ReplyDelete
  2. I honestly don't see why the hypothesis of extreme selective pressure isn't the correct answer. There are a variety of detailed mechanisms for this that could be at work, but it seems entirely plausible to me, particularly in the 4.2 kya event which was much more serious than the event that brought about Bronze Age collapse, although it isn't nearly as well documented.

    ReplyDelete
    Replies
    1. Extreme selective pressure is certainly plausible in humans, but I think the factors that would contribute to this trait foaming from the Atlantic are not there based on current reasoning.

      For example, what common denominator would explain the frequency of T-13910 in Basque country, the North Sea, the Isles, Baluchistan, Kurdistan, NW China or Mali?

      Similar climates? Similar histories?

      And why sould similar circumstances not be applicable in neighboring Southern or Eastern Europe, China, etc, etc (Obviously varying frequencies)

      Also, from the Bronze to Iron Age there is little movement in frequency from the Polish paper.

      I'll leave one possibility open, that is LP has nothing to do with milk consumption (at least the phenotype) and something else is at play. But, I think the real reason is just passive ancestry proportioned stemming from an Early Holocene, Near Eastern population

      Delete
    2. "what common denominator would explain"

      calories

      all you need are regions where for a certain time cereals didn't grow or grew badly.

      the association of milk with oats - originally a weed that grew in wheat fields but which grew better in northwest Europe - points at the (to my mind) obvious answer = low wheat yields in the north west leading to replacement with milk and oats


      the fact that is spread so dramatically among some populations but only much less so to the south and east implies calories is the answer

      .

      "Also, from the Bronze to Iron Age there is little movement in frequency from the Polish paper."

      The vistula was part of a major trade route from the Baltic to the Black Sea and the Polish Dukes hired lots of Scandi mercenaries so the 80% LP they found in some of those Polish samples could be from that.

      Southern Europe = wheat
      North Eastern Europe = millet
      North Western Europe = milk

      Delete
    3. Yes, but in the North Sea you have fish and reindeer and if you have herds you meat. If you have milk you have cheese. And even if you are not lactase persistent that doesn't necessarily make you intolerant.

      And why would T13910 and other appear with R1b V88 high Fulani in Northern Sudan, Niger and Cameroon but not their neighbors? And why in M73/M269 Western China or in Baluchi but not their neighbors?

      Delete
    4. Most non-White Americans drink Milk. Has it been proven these SNPs make much of an effect?

      Delete
    5. "And why would T13910 and other appear with R1b V88 high Fulani in Northern Sudan, Niger and Cameroon but not their neighbors? And why in M73/M269 Western China or in Baluchi but not their neighbors?"

      That's an interesting question. If R1b stemmed from near the steppe and spread during the copper age then V88 people would have had quite a journey: to Egypt then down to Sudan, across that pastoralist strip south of the Sahara to the gold fields of West Africa then up north again to hook up with BB / Atlantic Megalith in NW Africa.

      People will go a long way for gold.

      Delete
    6. @ Krefter. The answer is no, not yet.

      Correlation doesn't equal causation, and an observed phenotype may or may not have anything to do with a specific allele. For example, there has been an observed correlation between milk consumption and Multiple Sclerosis onset, however, MS, Cystic Fibrosis and Type I diabetes occur mainly within populations that are highly LP and the reason is probably caused by HLA mutations nearly exclusive in Western Caucasians, not milk consumption or LP.

      In other words, there are probably larger things going on that are ancestry driven, not fruit-fly stuff happening after the Bronze Age. Having said that, LP vs. LI is definitely something that is observable. There probably was selective pressure at some point, but my own hunch is that this happened a very long time ago.

      @ Grey
      Until DNA is obtained from SW Egypt, Algeria and Libya it's pretty much moot

      Delete
    7. MS is autoimmune and probably has a quite low ancestry component, but is associated with altitude (higher altitude is greater risk) suggesting a general immune system role which Vitamin D is associated with.

      Delete
    8. "And why would T13910 and other appear with R1b V88 high Fulani in Northern Sudan, Niger and Cameroon but not their neighbors? And why in M73/M269 Western China or in Baluchi but not their neighbors?"

      Are you suggesting, as does Maciamo Hay, that lactase persistence spread with R1b? That might explain Ireland 98% or Scotland, but how is Sweden a non R1b majority country listed at 99% or the Danes at 96%.?

      "If R1b stemmed from near the steppe and spread during the copper age then V88 people would have had quite a journey: to Egypt then down to Sudan, across that pastoralist strip south of the Sahara to the gold fields of West Africa then up north again to hook up with BB / Atlantic Megalith in NW Africa.

      People will go a long way for gold."

      R U Serious? Obviously R1b is old and the V88 branch broke off well before the chalcolithic. R1b only spread into west Europe in the late neolithic to chalcolithic.

      Delete
    9. I missed this one earlier but mind you: Fulani are not high in R1b-V88, that's preposterous! R1b-V88 is high among some Chadic peoples and ill-researched R1b (which could well be largely V88 but uncertain) is high in Sudan. Fulani are not Chadic, nor (with a lesser exception) Sudanese: they are West African by geography (extreme West African by origin: from Senegal) and Nigher-Congo by language. They do however sport some West Eurasian admixture of the NW African kind (Sahrawi-like) but the 13910-T allele is rare in NW Africa (Sahrawis have not yet been studied though).

      Delete
    10. @Anonymous

      Lactase persistence will appear at a low frequency in any mammal species that is lactose intolerant. A large portion of domesticated cats, for example, are lactase persistent. Because house cats were probably domesticated in the Middle East, it may be a sign of convergent selection given the culture of their diariest owners. BTW, populations with strong T-13910 allelic frequency and dairying tradition generally have strong Rhesus D deletion. (the Basque as one example)

      But regardless, the native state in mammals occurs naturally as it did in Paleolithic humans and as it does randomly from Tibet (three additional haplotypes) to Peru (others) at whatever frequency.

      The question isn't when or where T-13910 originated, because it is not relevant and it happened a long time ago. The question is how did this trait reach the frequency in the populations that have it at such high rates. Maciamo's map of the worldwide distribution of R1b and T-13910 nearly correlates to the "T" (no pun intended) and this is because of specific population movements beginning in the 4th millennium and I would suppose that the two are roughly ancestry proportioned. It's a reasonable hypothesis, so yes.

      "That might explain Ireland 98% or Scotland, but how is Sweden a non R1b majority country listed at 99% or the Danes at 96%.?""

      One thing to keep in mind is that most maps are interpolations of small test groups and Sweden straddles a cline so the data is kind of shaky in my opinion. You may well have a high phenotype in Southern Sweden with ethnic Swedes and then dropping off quickly after that. Some of the data may not exclude non-ethnic Swedes either.
      The allelic and phenotype frequency I've seen for Sweden has been all over the place so it's hard to make a quick judgement.

      @Maju

      I meant the percentage of ethnic Fula that are R1b which various among Fulani of different nationalities. I would say that they roughly correlate.

      Delete
    11. Well, excuse my ignorance but I'm not aware of any Fulani population or individual that are R1b. Could you illustrate me?

      Delete
    12. http://dienekes.blogspot.co.uk/2013/03/y-chromosomes-of-pastoralists-and.html

      Y chromosomes of pastoralists and farmers from the Sahel

      Am J Phys Anthropol DOI: 10.1002/ajpa.22236

      Multiple and differentiated contributions to the male gene pool of pastoral and farmer populations of the African Sahel

      Jana Bučková et al.

      The African Sahel is conducive to studies of divergence/admixture genetic events as a result of its population history being so closely related with past climatic changes. Today, it is a place of the co-existence of two differing food-producing subsistence systems, i.e., that of sedentary farmers and nomadic pastoralists, whose populations have likely been formed from several dispersed indigenous hunter-gatherer groups. Using new methodology, we show here that the male gene pool of the extant populations of the African Sahel harbors signatures of multiple and differentiated contributions from different genetic sources. We also show that even if the Fulani pastoralists and their neighboring farmers share high frequencies of four Y chromosome subhaplogroups of E, they have drawn on molecularly differentiated subgroups at different times. These findings, based on combinations of SNP and STR polymorphisms, add to our previous knowledge and highlight the role of differences in the demographic history and displacements of the Sahelian populations as a major factor in the segregation of the Y chromosome lineages in Africa. Interestingly, within the Fulani pastoralist population as a whole, a differentiation of the groups from Niger is characterized by their high presence of R1b-M343 and E1b1b1-M35. MOREOVER, THE R1b-M343 IS REPRESENTED IN OUR DATASET EXCLUSIVELY IN THE FULANI GROUP and our analyses infer a north-to-south African migration route during a recent past.

      Delete
    13. @ Bell Beaker Blogger -

      It does appear [at least superficially] that there is some correlation between R1b distribution and 13910*T distribution, I must agree.

      "The question isn't when or where T-13910 originated, because it is not relevant and it happened a long time ago."

      Yes, but we have to consider the long, extended haplotype of which 13910*T is a part.

      Assuming that 13910*T as an allele existed long prior to dairy consumption, and was found in random individuals in many different populations around the world at low frequencies, it would have likely been included in a wide variety of random haplotypes containing various different random alleles.

      Yet at some point in time, in some individuals in some population in some part of the world, dairy farming was starting to be practiced and one of the many possible random haplotypes containing 13910*T was the one which ended up being selected for and came to prominence as the so-called 'European' LCT haplotype.

      A haplotype is like a unique 'fingerprint' - the chances that the whole, entire 11 allele-long haplotype got selected for in exactly the same format in Fulani, Basques, British, Finns, and Udmurts in my opinion are slim.

      I believe that it originated in one population and then spread through asymmetric migration, with subsequent selection.

      13910*T is generally prominent only in those populations which contain the same long haplotype where selection has acted. For all I know the Andaman Islanders could harbor 13910*T as an allele, but is has never been selected for, and they certainly don't harbor the 'European' LCT 11-allele-long extended haplotype.

      So out of those prominent populations containing the unique extended haplotype, we can refer to a paper which has determined the TMRCA for 13910*T in those populations and therefore determine where the haplotype was likely to have been selected for first and subsequently migrated from. And it happens to be the case that the highest TMRCA is in the Fulani, with Basques and Finns/Udmurts tied for second place. And the date is early Neolithic.

      Interestingly, African GM immunoglobulin allotypes are also found in Iberians/Basques plus Volga Urals/Central Asia - this is something I need to look into. And I've already mentioned the HLA haplotypes.

      I also need to dig out that old 'Saami & Berbers - An Unexpected mtDNA Link' paper.
      Guinean Fulani, Algerian Berbers, Italians, Spanish, Yakut, and Saami all possess the same mtDNA marker.

      Delete
    14. @Chris: Alright, thank you. It's R1b-M343 (i.e. unclear what subclade), so most probably is V-88 borrowed from Chadic populations from Cameroon, via the Fulani realm of Adamawa and in general pastoralist mobility and shared lifestyle. It is not a lineage that define the Fulani and, excepted two localities (Zinder and Diffa) it is very minor.

      What you say about following the haplotypes' phylogeny is correct, at least in the outline, however I'd like to know how can you conclude that the Fulani have the oldest or most upstream known haplotype in this set. I'd also like to know if other African populations have been considered, because, you know, Fulani do not exist nor have existed (paleo-)historically in a vacuum.

      Delete
    15. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2253962/

      Scroll all the way down the article and look for 'Table 5'.

      Open table 5.
      http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2253962/table/tbl5/

      POPULATION ALLELE AGES OF THE ALLELE [GENERATIONS]

      Fulani T-13910 245
      Finns, East T-13910 217
      Finns, West T-13910 208
      Basques T-13910 208
      Saharawi T-13910 149
      Moroccans T-13910 82
      'Arabs' T-13910 39

      Ok, I remembered it ever so slightly wrong - 'Finns, East' are in second place behind the Fulani in first place. And Basques and 'Finns,West' are tied for third place.

      "I'd also like to know if other African populations have been considered, because, you know, Fulani do not exist nor have existed (paleo-)historically in a vacuum."

      I know that they share the whole haplotype with Bulala and Saharawi. The Saharawi have harbored the whole haplotype for far fewer generations than Fulani according to my data. But I don't have data for Bulala.

      ADMIXTURE runs which I have seen seem to suggest that Fulani are related to West Africans, Saharawi or Mozabites, and Chadic &/or Sudanic in some combination or other.

      http://www.unzcloud.com/wp-content/uploads/2012/01/fula1.png
      http://blogs.discovermagazine.com/gnxp/files/2012/01/fula1.png
      http://img690.imageshack.us/img690/1290/k6allaf001af080.png
      https://forwhattheywereweare.files.wordpress.com/2012/01/admixture.png ---> THIS ONE IS YOURS MAJU!




      Delete
    16. But that's age estimate based on each population's internal diversity of the haplotypes, not a real phylogeny. It just means that the haplotypes of the populations scoring older ages are more varied. It may be a hint but I'm a bit disappointed at the evidence. I was hoping for a haplotype network of some sort, possibly even phylogenetically sorted.

      For example it's possible that Europeans together have a greater diversity and hence an older joint age estimate than Africans (even if those are only Fulani), and, IF that would be the case, then your conclusion would just be wrong on the very same dataset and the very same logic. Your conclusion would only be partly right in cases in which haplotype diversity in Basque and Finns would be a subset of the haplotype diversity in Utahans (~British+Danish), and even then the age estimate of these and Fulani is close enough to each other than they rather seem both sets derived from a common ancestor rather than one descending from the other.

      Sadly, without any sort of haplotype network it's impossible to judge this issue properly.

      Note: there's a haplotype network but the populations at each node are not detailed. Maybe in the supp. materials? Hmm, there is a haplotype data for some of the populations but it's not organized... it'd need some work but so far the LCT+ haplotype closest to the ancestral state I could spot is Utahan (#97, derived at 2 positions), not Fulani (#27 is derived at 3 positions).

      If you're very interested on the matter (and you seem to be), Chris, I'd suggest that you build a haplotype tree with some patience using that supp. material data. It should be very revealing.

      "I know that they share the whole haplotype with Bulala and Saharawi."

      That would be interesting if true. However my impression from the supp. material is that the Sudanese Fulani and Sahrawi haplotypes do not look particularly related. In many cases the Sahrawi haplotype seems to share markers with Saudi and Mahra Arabs, while there is a characteristic sequence in intron 13 among many Fulani and Gali that is not shared by anybody else. Where are those Gali from? Sudan?, Ethiopia?

      "ADMIXTURE runs which I have seen seem to suggest that Fulani are related to West Africans, Saharawi or Mozabites, and Chadic &/or Sudanic in some combination or other".

      I agree re. Sahrawis + West Africans, can't say about the rest.

      Delete
    17. Erratum: it's not Mahra Arabs, but Mahas from Sudan. Also the Gaali are another population from Sudan, as are the sampled Fulani (who therefore may well be unrelated or only partly related to Fulani in general).

      Delete
    18. Thanks for your comments Maju.

      I will have a go at building a haplotype tree over the weekend as I have some spare time available.

      With so many papers and so much supplementary data to read it's sometimes hard to find the time, what with work and other commitments.

      Delete
    19. Ok there is a much more recent paper than the one I mentioned above [which is from back in 2008].

      The other paper, which I am sure you have seen, is:-

      "Genetic Origins of Lactase Persistence and the Spread of Pastoralism in Africa"
      [Tishkoff et al, 2014]

      http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3980415/

      Scroll down to the 'supplemental data' section at the bottom of that paper.

      Then click to view 'Document S1. Figures S1–S12:'

      On this pdf, go to page 2, and you will see:-

      Figure S1: Haplotype network.
      "Haplotype network analysis based on 97 SNPs, identi ed in the three sequenced regions, and four fast-evolving microsatellites spanning a total of 197.88 kb for 509 individuals (Figure 1). Haplotypes are shown as circles and the size of each circle is proportional to the number of individuals with a given haplotype. We assigned the same weight to SNPs and STRs."

      Delete
    20. My own analysis of the supplemental data from the 2008 paper was very interesting.

      *Derived at 1 position [from ancestral haplotype]:- Iranians [x2]; Ob-Ugrics [x2];
      *Derived at 2 positions:- Sudan Mahas [x1]; Sudan Gali [x2]; Utah Whites[x1]; Levant Arabs [x2]; Ob-Ugrics [x1];
      *Derived at 3 positions:- Saudi Arabia [x1]; Fulani Sudanese [x1]; Baluch [x1]; Morocco [x1]; Levant Arbs [x3]; Kalash [x2]; South Korea [x1];
      *Derived at 4 positions:- Sudan Mahas [x3]; Levant Arabs [x1]; Iranians [x1]; Kalash [x1]; South Korea [x2];
      *Derived at 5 positions:- Saudi Arabia [x1]; Sudan Mahas [x1]; Morocco [x1]; Iranians [x3]; South Korea [x2];

      *Derived at 6 positions - FIRST HAPLOTYPE CONTAINING 13910*T:- Fulani Sudanese [x1];
      *Derived at 7 positions - SECOND HAPLOTYPE CONTAINING 13910*T:- Baluch [x1];
      *Derived at 10 positions - THIRD HAPLOTYPE CONTAINING 13910*T:- Morocco [x1];

      [The skipping others]

      *Derived at 16 positions - FIRST HAPLOTYPE FOUND TO CONTAIN BOTH 13910*T & 22018*A IN LINKAGE DISEQUILIBRIUM [THE HALLMARK OF THE 'EUROPEAN' LP HAPLOTYPE]:- Iranians [x1];

      *Derived at 17 positions - MORE HAPLOTYPES CONTAINING BOTH 13910*T & 22018*A IN LINKAGE DISEQUILIBRIUM:- Levant Arabs [x1]; Morocco [x1]; Utah [x1];

      *Derived at 18 positions - MORE HAPLOTYPES CONTAINING 13910*T & 22018*A IN LD:- Ob-Ugrics [x1]; Levant Arabs [x1]; Saharawi [x2]; Baluch [x1]; Fulani Sudanese [x3]; Utah [x1];

      *Derived at 18 positions - ONE HAPLOTYPE CONTAINING 13910*T BUT NOT CONTAINING 22018*A:- Fulani Sudanese [x1];

      [Many more omitted]


      We also know from Tishkoff et al. (2014) that the 'European' haplotype containing 13910*T and 22018*A together in linkage disequilibrium was also found in Cameroon Fulani and Chad Bulala [number of dervied positions unknown].

      Delete
    21. Now you're providing very interesting information, thanks. I don't have time to double-check but, following your analysis, there is some chance indeed that the 13910-T haplotype could have an African origin (but migrated to Europe via West Asia, what is in itself intriguing). I'd like to see all that in tree-like image for better clarity though.

      Delete
    22. Ok I have to confess I have not produced a haplotype tree before (I am an 'armchair geneticist'); however I am more than willing to have a go at it.

      Any tips for me? Would a median-joining network be best? And which software should one use?

      If you have created one before, let me know how you went about doing it.

      Delete
    23. Can't help you. I only did some of those long ago and all by hand, when trying to understand R1b structure in Europe with the limited data we had prior to 2010. There's probably some sort of software but I do not know that much.

      Maybe you could ask Palisto of Kurdish DNA blog. He has made some of those for the blue eyes allele that were very informative.

      Delete
  3. I honestly don't see why the hypothesis of extreme selective pressure isn't the correct answer. There are a variety of detailed mechanisms for this that could be at work, but it seems entirely plausible to me, particularly in the 4.2 kya event which was much more serious than the event that brought about Bronze Age collapse, although it isn't nearly as well documented.

    ReplyDelete
  4. There are no samples from Western Europe. That's why!

    Why don't you post the map of the samples (fig. 1, top left corner): it reads itself. Instead if you say "Europe" it may induce to confusion.

    ReplyDelete
    Replies
    1. Good point. 10% is a bit misleading in of itself especially with the black hole that is the far West

      Delete
    2. Actually Chalcolithic Basques from the Ebro banks already had more rs4988235-T than any of the samples shown here: 27% overall with as much as 31% in San Juan Ante Porta Latinam.

      Notice anyhow that T allele frequency is not equal to "lactase persistance" because, on this allele alone, it may be distributed unequally through individuals (and that's precisely what we see in the Chalcolithic Basque samples, suggesting two different populations). Also there may be other alleles producing the same effect, just that they are not that famous or even known at all.

      I must say that the recent finding of similar mtDNA pools to those found in ancient Basques and resembling modern ones in the Seine basin, does reinforce the notion of Western Europe being key, not just for this allele but in general for the conformation of modern genetic pools in much of Europe.

      Delete
    3. Based on what's avaible from Eastern Europe and the steppe so far, I'm not buying the meme that this trait developed on the the steppe , where it was 30% at best, then moved to Western Europe where it was cut in half, then became almost 100% through a background scenario that can't be demonstrated.
      That's just to start!

      Delete
    4. Not sure where you got that meme from but sounds labyrinthically nightmarish. It does not even fit with the Asian data available here.

      There was some ~100% T population somewhere in West Europe: the Chalcolithic Ebro data is just the tip of the iceberg and is not even the most modern-like sequence in Basque ancient genetics, rather the opposite, showing a strong "Mediterranean" tendencies, like extra mtDNA K, that is missing in the previous and later sequences.

      There must have in SJAPL and Longar two different stocks: an Atlantic double-T one (minority) and a Mediterranean double-C one (majority). The latter is pretty much your usual Cardium/EEF but the former may be closer to Gökhem and the British milk-loving farmers and therefore also to North France.

      What's the T frequency among Germany's BBs? I'd imagine it's larger than average, as they also seem Western in R1b but I just don't recall how much.

      Delete
    5. I'm not sure to the last question, in the Haak paper, the oldest observed T among any samples was Quedlinburg Man, I believe. The other individuals from that cemetery and the other, I'm not sure.
      Altenmarkt Man, who was also R1b from the Allentoft paper is apparently DF27+ but not ancestral to the Iberian ones. I guess now that makes four Beakers that are so far 100% R1b, but all from Germany.

      I think one important component that needs to be added is dental plaque with the alleles that define LP. If the Yamnayans have a haplotype ancestral to LP yet they lack plaque buildup, then I think the case made by Allentoft gets shakier.

      Delete
    6. "Not ancestral to the Iberian" R1b is not important because the Iberian lineages are only part of a wider patch whose center seems to be in France (rather to the south). I'm not really sure what's up with that Iberian obsession some people seem to have, my emphasis is Western or Atlantic Europe, which includes Iberia (or parts of it) but also France (at least it's Atlantic facade) and other areas such as Belgium, Britain and Ireland, maybe even Rhineland (I don't want to be too specific because the data does not allow for that yet).

      I'm going to use the Jean Manco page on the matter, even if she cheats by misreporting on some issues I hope it does not affect this page. Effectively the T allele seems to be associated with Atlantic Neolithic: it is not only found in the already reported Upper Ebro/Basque military cementeries' aDNA but at similarly high frequencies in:

      1. Gökhem Megalithic (25-50%, the data does not seem to clear, at least for Manco's cherrypicking criteria)
      2. Gotland Pitted Ware (1/19 = 5%)
      3. Qedlimburg (50-100%, again not too clear if the single sample was CT or TT)

      And that is all before 2000 BCE. The first T-dominated sample after these is from Medieval Germany (Dalheim), with 100% T allele (18/18), a handful of royal Swedish remains also fit in this scheme.

      The data is not really clear because there are very few samples from areas where the allele is dominant today (Atlantic Europe, plus Finno-Scandia) but in these cases, in all cases studied, the frequencies were high already in Chalcolithic times and there is good reason to expect that they were even higher in some populations within the huge blank area awaiting research, for example those milk-loving British farmers or coastal Basques (as opposed to Ebro ones), etc.

      Delete
  5. "For example, what common denominator would explain the frequency of T-13910 in Basque country, the North Sea, the Isles, Baluchistan, Kurdistan, NW China or Mali?"

    Well, all of those populations bear disequilibrated HLA haplotypes of recent [<15kya] Sudanese &/or West African origin.
    In fact often the same ones.
    Also, a paper which I provided a link to a while ago [if not on here then on Eurogenes and on Greg Cochran's blog] clearly demonstrated that T-13910 has been harbored by the Fulani [including whatever ancestral population from which they derive] considerably longer than any of the other populations which harbor T-13910 in the world.

    A33-Cw3-B58 haplotype - notable frequencies in: West Africa; Sudan; S.Arabia; Pakistan; NW China; etc. [the haplotype itelf and all of its components found at peak frequencies and diversity of recombinants within Africa]

    Of course almost nobody seems to want to believe that the alleged 'European' lactase persistence allele might have actually originated in and diffused from an African population, hence why the data presented in that paper has been ignored.

    ReplyDelete
    Replies
    1. T-13910 can't really be originally African because it's found only at limited frequencies in NW Africa (not Egypt nor Sudan!) and even more limited ones in R1b-high peoples of the Chad region.. East Africa has some LP alleles but are all locally restricted or Arab in origin (different allele than the European variant).

      Anyhow... which are the frequencies of T-13910 in Kurdistan or Uyghuristan? AFAIK they are not too significant. IMO the T-13910 allele is just another European-centric and very specifically West-Europe-concentrated genetic variant like Rh⁻, R1b-M412 or high frequencies of mtDNA H.

      Delete
    2. T-13910 and the East African [Kenyan, Sudanese and Ethiopian] G-13907 LP allele share the exact same haplotype background.
      Another LP allele, C-13913 [found in Italy, Cameroon, Sudan, Ethiopia, & Saudi Arabian Bedouin], is only 3bp apart from T-13910.
      While both G-13907 and a third LP allele, G-13915 [found in Cameroon, Ethiopia, Kenya, Sudan, and Saudi Arabian and Jordanian Bedouin] are only 5bp apart from T-13910.
      And C-14010, found in Kenya and Tanzania, is a variant close to T-13910.
      In addition, Ethiopia populations harbor all not only all four of the above mentioned LP alleles, but also one more [G-14009].
      The TMRCA of T-13910 was estimated to be higher in Africa in the paper which I gave the link to.
      And to top it all off, there are other genetic markers indicating asymmetric migrations from Africa [green Sahara?] into Europe and Asia within the last 10-12kya.

      But no, of course T-13910 couldn't possibly have spread from Africa into Europe and Asia.
      It can only have originated in and spread from the Pontic-Caspian steppe with Kurgan migrations...right??

      Delete
    3. I was reading this this morning or yesterday evening. Just published:

      "Diversity of lactase persistence in African milk drinkers" Bryony Jones et al. 9 June 2015

      Here's the link, its open access. DOI: 10.1007/s00439-015-1573-2

      So your post is timely.

      Delete
    4. Thanks very much for this. I will have a read of it today!

      Delete
    5. @Chris: I don't understand well what you mean by "haplotype background"? Is there any study (professional or amateur, I don't care) that analyzes that? Does it include the non-persistence allele variants, a necessary control group, as we are all related, particularly West Eurasians? Because I have the impression that you are implying that there was some original LP allele that gave rise to all those variants, yet I have never read such claim before now, and I can think of many scenarios in which haplotypes are "related" just as our general genetics are related.

      "But no, of course T-13910 couldn't possibly have spread from Africa into Europe and Asia.
      It can only have originated in and spread from the Pontic-Caspian steppe with Kurgan migrations...right??"

      Not my stand. My stand would rather be that it is a Western European trait and hence totally unrelated to Kurgans. It can be a fluke or it can be subject to selection, that I can't say as of now, but definitely Western European.

      Delete
    6. From Ingram et al, 2009:-

      Lactose digestion and the evolutionary genetics of lactase persistence

      http://www.cnmd.ac.uk/mace-lab/publications/articles/2009/Ingram_HumGen09_LCT_Review.pdf

      " However, the three newly described SNPs all occur on different haplotype backgrounds from each other (using our old nomenclature: -13907*G, on A, -13915*G, on C, and -14010*C probably on B) (Enattah et al. 2008; Ingram et al. 2007; Ingram 2008; Tishkoff et al. 2007), ALTHOUGH -13907*G IS ON THE SAME HAPLOTYPE AS -13910*T."

      So -13910*T [Africa & Eurasia] is on the same haplotype as -13907*G [Africa only].

      Delete
    7. So there are the three long lactase persistence haplotypes [designated as: A, B, & C].
      Of these, haplotype A is found in Africans in 2 forms: one version which includes -13910*T; and one version which includes -13907*G.
      While in Eurasians, haplotype A is found, but only in the form which includes -13910*T.

      My claim, which I fully expect everyone to reject, is that both forms of haplotype A arose originally within Africa. And of the two versions of haplotype A, it was the 13910*T version which entered Eurasia from Africa via asymmetic migration(s) and was then subsequently selected for.

      Of course, somebody might wish to claim that haplotype A first originated in Europe in only the form which includes 13910*T. And then this version subsequently entered Africa via asymmetric migration(s), and a second version of haplotype A which incorporated the new 13907*G variant then subsequently arose and was selected for there.

      Delete
    8. Correct me if I'm wrong, but I understand that the only conclusion that can be derived from your quote is that the first set belongs to a West Eurasian genetic background, even if the 13907-G haplotype is strictly African (but Afroasiatic with a Red Sea coastal distribution centered in Eritrea, see: http://forwhattheywereweare.blogspot.com/2014/03/lactase-persistence-genetics-in-africa.html), and other African LP haplotypes are not.

      There's no implication I can discern that these phylogenetic clusterings exist to the exclusion of non-LP haplotypes, so...

      As for Western Europe, I only mean it in relation to 13910-T, because it is Western Europe where it is dominant today and where it probably coalesced (or at least first became fixated) based on still fragmentary aDNA evidence.

      Delete
    9. The 'A' haplotype is found in the same long format in West Asia, Africa, and Europe.

      The only difference is that East Africa has the 13907*G LP allele contained within this distinctive haplotype.

      Whereas Europeans, Algerian Mozabites, Palestinians, Bulala from Chad, Baggara and Fulani from Cameroon, Arabs and Fulani from Sudan, Ethiopians, and a small percentage of Saudi Arabians have the 13910*T allele contained within the haplotype.

      So the real question is, did haplotype 'A' originate with the 13910*T allele contained with it, or with the 13907*G allele within it.
      Which is derived from which?

      I have no problem accepting that the whole haplotype could have originated in West Asia rather than Africa.

      But I remain far from convinced that the presence of haplotype 'A' in both forms in the various African and West Asian populations mentioned above is explained by a back-migration from Europe [and Greg Cochran actually believes that this is the case].

      Delete
    10. Eritreans, Ethiopians and other NE Africans are like 50% West Eurasian in their genetic pool.

      My question is whether there is a distinction between LP and non-LP haplotypes. And you have provided no information on that. My impression is that there is not such distinction and that the clade you insist so much on ("haplpotype A") is as such non-informative therefore: it is just the generic West Eurasian background, including both LP and non-LP subtypes.

      Delete
    11. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1234910/

      Lactase Haplotype Diversity in the Old World

      Abstract

      Lactase persistence, the genetic trait in which intestinal lactase activity persists at childhood levels into adulthood, varies in frequency in different human populations, being most frequent in northern Europeans and certain African and Arabian nomadic tribes, who have a history of drinking fresh milk. Selection is likely to have played an important role in establishing these different frequencies since the development of agricultural pastoralism ∼9,000 years ago. We have previously shown that the element responsible for the lactase persistence/nonpersistence polymorphism in humans is cis-acting to the lactase gene and that lactase persistence is associated, in Europeans, with the most common 70-kb lactase haplotype, A. We report here a study of the 11-site haplotype in 1,338 chromosomes from 11 populations that differ in lactase persistence frequency. Our data show that haplotype diversity was generated both by point mutations and recombinations. The four globally common haplotypes (A, B, C, and U) are not closely related and have different distributions; the A haplotype is at high frequencies only in northern Europeans, where lactase persistence is common; and the U haplotype is virtually absent from Indo-European populations. Much more diversity is seen in sub-Saharan Africans than in non-Africans, consistent with an “Out of Africa” model for peopling of the Old World. Analysis of recent recombinant haplotypes by allele-specific PCR, along with deduction of the root haplotype from chimpanzee sequence, allowed construction of a haplotype network that assisted in evaluation of the relative roles of drift and selection in establishing the haplotype frequencies in the different populations. We suggest that genetic drift was important in shaping the general pattern of non-African haplotype diversity, with recent directional selection in northern Europeans for the haplotype associated with lactase persistence.

      Delete
    12. "Construction of the LCT haplotype network reveals the relationships of the common haplotypes to each other and to the root haplotype and helps in the interpretation of the present-day distributions by emphasizing that the A, B, and U haplotypes are not directly connected to the root and must have been generated by intermediate haplotypes that are now rare or absent from the non-African populations.

      The sub-Saharan African samples show much greater haplotype diversity than other populations (tables ​(tables11 and ​and2)2) with many extra haplotypes not seen in the other population samples. Analysis of the unweighted means of LCT haplotype-class frequencies shows that the four common haplotypes and their simple derivatives (haplotypes A, B, C, U, O, D, N, and F) comprise 89% of total haplotype diversity in non-Africans, but only 51% in sub-Saharan Africans. Both the San and Bantu have high frequency of haplotypes (30% and 28%, respectively) that are the intermediate steps between the four common haplotypes, but these are almost absent in non-African populations (⩽3%). This pattern of higher diversity in Africans has been observed at other loci and has been interpreted as evidence for an “out of Africa” model for peopling of the non-African continents (Ayala and Escalante 1996) though the role of selection and population size in altering diversity within populations is debated (Harding et al. 1997; Jobling et al. 1998)."

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

      Delete
    14. Thanks Chris. I look forward to reading this.

      Delete
  6. Also:-

    A29-Cw16-B44 haplotype - notable frequencies in: West Africa; Algeria; Eastern Iberia; Basque Country; Western France; Western British Isles. [the haplotype itelf and all of its components also found at peak frequencies and diversity of recombinants within Africa].

    ReplyDelete
  7. If there had been so much selection on light eyes and skin plus lactose persistence it would show up in linkage disequilibrium, but it doesn't so it didn't happen.

    You can keep sampling eastern and central europe neolithic a million years and won't find where NW european people came from because they were already there/ Even if there was some megamigrations, certainly r1b DNA was where it is now by then.

    But I think they know that, they have zero interest to find the truth, just to make some lame political statements.

    ReplyDelete
  8. "If there had been so much selection on light eyes and skin plus lactose persistence it would show up in linkage disequilibrium, but it doesn't so it didn't happen."

    The lactase persistence haplotype found in Europeans is very extended and very disequilibrated, and given its frequencies I don't doubt that it was selected for. The thing which I have been quibbling over is the geographic origin of the haplotype. I don't think that it necessarily originated in Europe. Same haplotype shared between Fulani and Basques, but Fulani have a higher TMRCA for the active allele contained within the haplotype [13910*T]; plus there seems to be another variant of the haplotype in Africa which contains the 13907*G derived allele [no such diversity in Europe].
    Given that Basques also share a disequilibrated HLA haplotype with tropical western Africans, and this haplotype has diverse recombinants in Africa but not in Basques, Iberians, or French, I think that this suggests gene-flow northward to the Basque region and not vice-versa, and I also believe that the lactase persistence haplotype did the same.
    People are saying that the presence of the so-called European LCT haplotype in Fulani must be due to the Romans or the Vandals in North Africa; I am totally not convinced.

    "But I think they know that, they have zero interest to find the truth, just to make some lame political statements."

    Agree with this statement.

    ReplyDelete
    Replies
    1. Can you document the presence of the European LCT allele in Fulani? My data does not report it so far.

      Anyhow, IF there was a selection for the West European LCT allele, it must have happened before 3000 BCE. I must insist on the Upper Ebro sites issue (never mind Gokhem): http://forwhattheywereweare.blogspot.com/2012/01/caught-in-act-lactose-intolerant-and.html

      What we see in San Juan Ante Porta Latinam and Longar (military cemeteries dated to c. 3000 and 2500 BCE) are two distinct populations: one with the T allele fixated and another one with the C allele fixated (almost no CT individuals, underlining a clear segregation of both populations before death). The latter is dominant and fits well with all other Neolithic findings (save Gokhem) in being lactose intolerant (apparently), the former however has the T allele fixated and that is super-important: it means that selection (if that's the cause) had already acted in them driving the T allele to fixation (alternatively it's a random founder effect or fluke fixation by mere drift, fixated T in any case).

      So there is no need to appeal to selection after those dates but a realistic scenario can well be that of expansion of the T-fixated population from (I believe) Atlantic Europe eastwards.

      This scenario would also explain other genetic traits that are most concentrated in the Western coastal region of Europe, including haploid structure, be it Y-DNA R1b or mtDNA H1 and H3.

      Another piece of evidence in favor of an early Atlantic fixation for the T allele is the massive consumption of dairies in Neolithic Britain (and its persistence up to the Danelaw, which is when love for fish is reintroduced in the island). See: http://forwhattheywereweare.blogspot.com/2014/02/neolithic-peoples-from-britain-and.html

      Delete
    2. Never mind: I found the reference myself: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3031279/ (see table 1).

      That is fascinating because it seems to imply (my opinion) that somehow they have inherited the trait from the Paleolithic flows into NW Africa at the Oranian genesis (Fulani have like 1/3 of NW African "Eurasian" ancestry, most similar to Sahrawis specifically). In other words: there is a good chance that the T allele originated in Paleolithic populations of Europe (however it has yet to be found among them, I reckon, but also among the much better sampled Neolithic farmers), otherwise it is nearly impossible to explain its presence among the Fulani.

      Delete
    3. NW Africans have only low frequencies of the European LCT allele. I wonder if Sahrawis or Mauritanians have it (seems that Tuareg do not). That would simplify the process of transmission a lot.

      Delete
    4. @Chris Davies

      "I think that this suggests gene-flow northward to the Basque region and not vice-versa"

      Just wondering why you think northward and not southward.

      Delete
    5. @ Marnie:

      I'll start with the HLA haplotypes.

      Those that have been in European populations a long time are highly equilibrated; those that are more more recent are highly dis-equilibrated.

      HLA haplotype: A29-Cw16-B44 [shorthand].

      = A*29:02-B*44:03-C*16:01-DRB1*07:01-DQB1*02:01 [longhand].

      Peak European frequencies:-
      *Gipuzkoa Basque 10.00% [#1 haplotype in Basques]
      *Spain Catalonia Girona 7.50%
      *Spain Ibiza 6.10%
      *Arratia Valley Basque 5.30%
      *Spain Murcia 5.10%
      *Spain Majorca & Minorca 4.40%
      *Spain Minorca 3.90%
      *Northern Ireland 2.90%
      *Spanish migrants in Germany [N=1107] 2.78%
      etc. etc.
      [Eastern Europe = 0%]

      The haplotype and versions of it are found in Maghreb, Cameroon, Zambia, South Africa Zulu, etc. [probably many others, but Africa needs better sampling - we have plenty of data from Eurasia alredy].

      The haplotype and its allelic components are found at their greatest frequency and diversity of recombinants in tropical African populations.

      Let's look at the alleles contained within this haplotype [class I alleles] - peak world frequencies:-

      A*29:02
      *South Africa Natal Zulu 11.00%
      *Cameroon Yaounde 10.40%
      *Zimbabwe Harare Shona 9.60%
      *Cameroon Bamileke 8.40%
      *Cameroon Sawa 7.70%
      *Morocco Chaouya 6.80%
      *Tunisia 6.70%
      *Cameroon Beti 6.30%
      *Zambia Lusaka 5.80%
      *Kenya Luo 5.30%
      *Uganda Kampala pop.2 5.10%
      etc. etc.

      B*44:03
      *Cameroon Beti 10.90%
      *South Africa Natal Zulu 10.50%
      *Cameroon Bakola Pygmy 9.44%
      *Ghana Ga-Adangbe 8.78%
      *Cameroon Bamileke 8.40%
      *Central African Republic Mbenzele Pygmy 7.26%
      *Zimbabwe Harare Shona 5.80%
      *Uganda Kampala 4.70%
      etc. etc.

      C*16:01
      *Mali Bandiagara 28.30%
      *Burkina Faso Fulani 21.40%
      *Burkina Faso Mossi 17.90%
      *Ghana Ga-Adangbe 17.18%
      *Senegal Niokholo Mandenka 16.80%
      *Burkina Faso Rimaibe 16.00%
      *Equatorial Guinea Bioko Island Bubi 10.50%
      *Zambia Lusaka 10.00%
      *Kenya 8.40%
      *Zimbabwe Harare Shona 8.00%
      *Tunisia 7.20%
      *Morocco Chaouya 7.10%
      *South Africa Black 6.60%
      *Cameroon Beti 6.00%
      *Sudan Mixed 5.00%
      *Uganda Kampala 4.90%
      *Cameroon Bamileke 4.50%
      *Kenya Luo 4.50%
      *Kenya Nandi 4.40%
      etc. etc.

      [Source: www.allelefrequencies.net]

      So my reasoning was, if HLA haplotypes have moved northward, eg. from Fulani or proximal populations towards Basque Country, and Basques happen to share the exact same lactase persistence haplotype with Fulani, then why could the LCT haplotype not also have moved northwards rather than southwards.

      Moreover, another paper which I presented a link to recently demonstrated that the TMRCA for the active allele contained within the LCT haplotype, 13910*T, is greater in Fulani than in Basque [though both give a Neolithic time frame].

      Delete