Thursday, April 9, 2015

Guess What?...LP Allele Not Under Selective Pressure.

Beaker blog already knew this.  Check Bernard Secher's blog comments [here].  I think he rightly lays T-13910 at the feet of the mysterious Beakers (or their immediate antecedents).

Moroccan question marks march across Chalcolithic Western Europe

So first of all, let me say that the "no selective pressure" conclusion is not what the paper said; that's what I wrote in the title because, even without ancient DNA, there are walls of constraints which are moving in the wrong direction.  I pre-concluded it.

We are inundated with a steady drumbeat of crazy selection arguments for T-13910 because this imparts some kind of super-human benefits to its bearers even though it takes only fifteen minutes to separate whey from curd.  In order to show rapid selection you'd need to show that the frequency has been increasing steadily over time or has increased through a series of punctuated bottlenecks.

You could have looked at China, Africa or the Middle East to have figured out that it has been a very long time since this adaptation has been under any kind of positive selective pressure.

Short Horn Cattle
In a previous post "The Bracyceros and the Bracycephlics"  I talked a little about the emergence of the Short Horn (Dairy Cow) in Europe about the time of Beakers.  The Short Horn probably entered Europe from North Africa and if you go further back it probably originates in Baluchistan.

I have some other stuff to post, hopefully will get to some of it this weekend.

Hunting for the LCT-13910*T Allele between the Middle Neolithic and the Middle Ages Suggests Its Absence in Dairying LBK People Entering the Kuyavia Region in the 8th Millennium BP. Witas HW, Płoszaj T, Jędrychowska-Dańska K, Witas PJ, Masłowska A, et al. (2015)  PLoS ONE 10(4): e0122384. doi:10.1371/journal.pone.0122384 [Link]


  1. Why do you say it was not under selective pressure? I have not clear myself if there was or not such pressure, so I'm very intrigued about your argumentation.

    Anyhow, what I do have clear is that focusing on cows is anachronistic: the main milk source in Europe traditionally were goats (cows were only for the well-off and anyhow mostly used for traction and meat, rather than milk, goats instead live on almost nothing and give plenty of milk year-long, unlike sheep, whose dairy produce is seasonal only). In fact I've argued that, if there was a selective pressure for the quasi-fixation of the T allele, this must have been social inequality and poverty, against which goat milk provided surely an affordable nutritional paliative ("acorn bread and goat milk", the traditional pauper's food over here). See:

    Also, as I commented on your linked other entry, African cows are radically different from European ones. As complement I will mention here that the European lactase persistence variant is rare even in NW Africa:

    1. I imagine that LP occurs very occasionally in any mammal, in fact I suppose the earliest proto-mammals would have been LP.

      At least one human in prehistory was T-13910 and it probably descended naturally within a population of Paleolithic hunters at some frequency.

      As you said, it could have well been positively selected by people milking goats. It could have happened with people weening their children off nanny goats in the Northern Zargos or thereabouts in the Early Holocene since that is where the ibex was domesticated. It also may explain the frequency in Pakistan. Or it could have been cows or both.

      The problem with the super-duper selection argument is that lactase persistence is not only not-universal, it hovers in unimpressive frequencies among populations where it should be universal (Egypt, SE Europe)

      Also, there hasn't been any major subsistence crisis in Western Europe in the last 5k years

    2. Basically, I think this study shows some mild fluctuation (up and down) within one region with mtdna indication probable continuity (for the most part)

      Those numbers would probably flatten out with more data. So it appears to me that LP frequency is a product of immigration in this area and it has remained stable

    3. " there hasn't been any major subsistence crisis in Western Europe in the last 5k years "

      Besides my argument of social inequality and land accumulation since the Chalcolithic potentially causing major nutritional challenges to the vast majority of the population, who by definition were poor farmers, I find that since historical records began there have been many major famines in Europe ( Often these famines killed huge chunks of the affected population, from 15% to 41%, some were continent sized:

      And this is only since records began some 1600 years ago. As you can see in the list the occurrences grows as time passes just because historical records are better. So I see absolutely no justification for your claim. In theory at least, nutritional challenges, which were common and repeated, can well make the difference between life and death depending on whether you can drink goat milk or not. There's no time to make cheese, a costly process, you have to wet your bread, acorn bread possibly or oatmeal, in the milk and gulp the mix. If you can't, if doing that makes you diarrheic, you may be doomed.

      I don't know if there was or not an actual selection. The data IMO is not yet conclusive, but there is a serious potential for it, especially when we understand that milk from goats was not any luxury but common in the Medieval and hence also surely Proto-historical diet of much of Europe. It's not about calcium nor much less vitamins, it's about something much more basic: fat, sugars (lactose) and proteins. If your only other food was some cereal (bread, oat meal or the equivalent acorn bread) then you seriously needed those proteins, as well as the extra caloric intake of fats and sugars.

      I agree that in the case of Cuyavia, the data seems to support a N-S divide with only minor fluctuation, so it's suggestive of selection (or even regular admixture) having playing no role. But it's a small region.

    4. I've seen figures for lactase persistence in Irish of 96-98%. Given the subsistence nature of much of the population there and the many famines alongside the huge importance, historically, of pastoralism. I would argue it was positively selected for there.

  2. "We are inundated with a steady drumbeat of crazy selection arguments for T-13910 because this imparts some kind of super-human benefits to its bearers"

    Does LP in Arabian Bedouin impart super-human benefits to its bearers or does it allow a pastoralist alternative in a region where farming is not viable?

    LP in Europe was clearly massively selected for at a certain point in time and the simplest explanation is the same as the Arab one.

    What is actually happening is people are blinding themselves to the obvious because they dislike the selection argument in case the reason for the selection is x when the reason is more likely to be y.

    1. What this paper shows is that within one region over time there doesn't appear to be selective pressure (recently).

      Now the fact that Western Europeans are uber dairiest would imply that one is dependent on the other, but if all Beakers were already LP (1/1 or 5/5 so far) then in the last 5000 years it would appear the modern allele frequency is ancestry proportioned.

    2. But what's "all Beakers"? I would say that El Mirador or (Southern) Basque Chalcolithic samples were related to to Beakers. There's no such a clear cut Beaker/non-Beaker definition over here as you know, that's something that only happens in Central Europe, and even there there are many cases like Kromsdorf where there are doubts.

    3. Or what about Treilles?! Treilles was extremely Bell Beaker yet zero LP and zero Atlantic genetic influence we can discern.

  3. As I have discussed previously with Maju over at Eurogenes Blog a few weeks ago [literature was cited]. Fulani have harbored T-13910 LP allele for by far the greatest number of generations of any population. Basques and British on the one side of Europe are next; and Finns & Udmurts on the other side of Europe have harbored the allele for a similar number of generations to the Basques. By comparison, 'Berbers' have not harbored the allele for so many generations. My opinion is that T-13910 spread into Western Europe from North Africa with megalithic culture rather than Bell Beaker culture, from a small number of carriers connected to Saharan pastoralists, and related to present-day Fulani. I think that Maju is right, and that the allele may be more associated with an earlier pastoral culture based, at least partly, on locally-domesticated ovids/caprids [with some African influence] plus pre-Bronze Age imported African bovids [as attested to by mtDNA studies in cattle]. Probably La Hoguette culture is related to this. Another genetic marker worth considering is the HLA haplotype A*29:02-B*44:03-C*16:01. This haplotype is clearly of West African origin. It still exists there, diversity of all three alleles is higher in that region, and haplotypes of C*16:01 are nodal and more diverse there. The haplotype is present in North Africa, but looking at Europe one can see that it peaks in the Basques, and is also at high frequency in Catalonians, Murcians, and Balearic Islanders. It is also at a lower, but still significant frequency in different regions of Portugal. From Eastern Iberia/Western France the haplotype appears to have spread northwards, and is notable in Ireland and Western Britain. There has also been some spread towards the Low Countries/Northern Germany/Denmark. Considering the range of its spread but its limited equilibration I would estimate that it likely entered Europe prior ro 5kya.

  4. What about T-13910 in Finns and Udmurts at an early date? I suspect that the paper "Saami and Berbers, an unexpected mitochondrial DNA link" is connected with this. The populations which are mentioned are Fulani of Senegal; Algerian Berber; Italian; Spanish; Saami; and Yakut. But they suggest that the mtDNA connection between these populations can be attributed to repopulation of Europe and North Africa from the Franco-Cantabrian refuge <15kya. I'm not sure that I necessarily agree. If the mtdna link of Fulani->Algeria->Italy/Spain->Saami is connected to the T-13910 link of Fulani->Basque & Fulani->Finns/Udmurts, then I am inclined to consider gene-flow via separate routes in the Neolithic: Sahara->NW Africa->Iberia/W. France->W. British Isles; & Sahara-> NE Africa->Levant->Caucasus->Finns/Udmurts; &/or: Sahara->NE Africa->Levant->Anatolia->Balkans->[Basques/British in the west + Finns/Udmurts in the east].

    1. Great comment. One thing I will add regarding Eastern Fennoscania and Karelia is that this area is much more complicated that simple HG's. Karelia man was certainly part of an immigrant population from the Volga. Palynology suggest some in Karelia cultivated domestic crops around 5,000 b.c. although it has so far failed in the thin at archeological record. Also, metallurgy much in this area much earlier than expected around 4,000 b.c. I suspect it may be much earlier than that.

      I'm interest in how this allele relates in Western Pakistan. Where does this population fit with the others?

  5. Thanks.

    "I'm interest in how this allele relates in Western Pakistan. Where does this population fit with the others?"

    You mentioned Baluchistan.

    I have full HLA data from Pakistan Baloch and Iran Baloch. I can confirm that these populations have a number of haplotypes shared with West Africans, Sudanese, and West Asians. I won't list them all, but I can confirm that some of the prominent haplotypes are clearly of African origin and not vice versa.

    For instance, A*33:03-B*58:01-C*03:02 - very prominent in Pakistan Baloch and Iran Baloch, but also found in Gambia; Cameroon; Sudan; Oman; UAE; Pakistan Burusho/Kalash/Pathan/Brahui; and southern China. The alleles in this haplotype are more diverse in West Africa and haplotypes of A*33:03 and B*58:01 are nodal and more diverse there also.

    Also worth considering is that Y DNA haplogroup B has been found in Iran, Pakistan, Afghanistan, NW India. And I spotted it among a large sample of Chinese Americans [Xu et al 2014].

    I can pick one haplotype at random from Iran Baloch - A*74:01-B*81:01-C*08:04 [1.70%].

    Now look at where the alleles are at peak global frequency and diversity:-

    Peak world frequencies: Guinea Bissau; Burkina Faso; Cameroon; Kenya; Uganda;
    Peak world frequencies: Cameroon; Kenya; Zimbabwe; South Africa;
    Peak world frequencies: Mali; Cameroon; Kenya; South Africa;

    There are several such examples in the data.

    1. Actually, I meant T13910. It shows three frequency peaks, the third one appears in the Lower Indus and Baluchistan.
      I'm interesting in the directionality/age in this region if the data is available.

    2. Looking back, it seems that I made an error and the East Finns may have harbored T-13910 slightly longer than the Basque.

      Here is the relevant link:-

      "Age Estimates of TMRCA for LP H98 T-13910 in Global Populations"

      Table 6
      Using Rho Method (years [±SD])
      [Upper Bound]
      Fulani 10,735 (±1,193)
      Utah Whites [Predominantly British-descended] 9,736 (±1,082)
      Finland East 9,508 (±1,494)
      Finland West 9,113 (±1,013)
      Basque 9,108 (±1,012)
      Udmurt 8,820 (±1,386)
      Baloch 7,543 (±838)
      Sindhi 7,134 (±793)
      Brahui 6,615 (±735)
      Pathan 6,125 (±962)
      etc. etc.

      Lower Bound
      Fulani 6,134 (±682)
      Utah Whites [Predominantly British-descended] 5,563 (±618)
      Finland East 5,433 (±854)
      Finland West 5,207 (±579)
      Basque 5,205 (±578)
      Udmurt 5,040 (±792)
      Baloch 4,310 (±479)
      Sindhi 4,077 (±453)
      Brahui 3,780 (±420)
      Pathan 3,500 (±550)
      etc. etc.

      And here is the link to the main paper:-

    3. Perfect, thanks.

      Basically, it seems that T13910 can be broken up into four regional groups.

      (1. Baltic to White Sea corridor) (2. Baluchistan/Afghanistan/Indus) (3. Sahel (represented by Fulani and others)) (4. NW Europe extending SE)

    4. The T-13910 allele is part of a haplotype.

      This haplotype includes the T-13910 and A-22018 variants.

      The same haplotype background is shared by several populations, including Europeans, Fulani, Baggara from Cameroon, Bulala from Chad, Mozabit from Algeria, and Palestinians:

      T-13910 also shares the same ancestral allelic background as the East African G-13907 allele:-

    5. I noticed earlier that A-22018 appears to be the dominate LP type of Northern India and Northern China.
      I also read a paper (2013) concerning possibly more unique alleles associated with the Tibetan plateau.

      It could all be a simple case of convergent evolution if you buy the cultural evolution premise of LP. I'm sort of curious that these sort of alleles don't spontaneously appear in populations like Eskimos, but rather seem to find their way back to animal domestication events of the Near East connected either by the silk road or the African savanna.

  6. I also think that Y DNA haplogroups L and T might be relevant. Haplogroup T connects Africa with Pakistan, but Pakistan also has sister haplogroup L. Previously it appeared that Africa lacked L. But a few days ago I looked at some Family Tree DNA data from the Sudan Project and happened to notice that L was present in that data. T of course is found in numerous African populations in West, Central, Eastern and Southern Africa.

  7. Quote: "We are inundated with a steady drumbeat of crazy selection arguments for T-13910 because this imparts some kind of super-human benefits to its bearers even though it takes only fifteen minutes to separate whey from curd"

    But you first have to learn how to separate whey from curd which I imagine took a lot longer.

    1. →

      They were smarter than you seem to think. Cheese production is very old.

      However cheese is still risky for lactose intolerant people, unless it's cured for a very long time (varies a lot depending on individual person and individual cheese). That takes more than just 15 minutes. Cheese, particularly mature cheese, is therefore a more expensive product than raw milk or its immediate derivates (fresh cheese, curdled milk, butter, cream or more modernly yogurt, which seems a Turkic invention) and any cheese-lover knows that sad fact.

      On the other hand many genetically lactose intolerant people seem able to develop at least some level of tolerance. It's not all genetic. However if a population is generally lactose intolerant, it should tend to consume much less dairies than another population that is tolerant. I say "should" because the most milk-drinking region of Spain is also the most lactose-intolerant: Canary Islands, so there are things that don't always fit well with the theory (I wonder if it's just because of massive affluence of Nordic tourists however).

  8. By the way something I recently noticed is that lactose intolerance only seems to manifest in the late teens, what means that even lactose intolerant children are actually lactose tolerant until quite a late age. With a max. Neolithic lifespan (per burials) of c. 40-45 years, that's more than 1/3, almost 1/2 of life. Maybe is not that important whether fully grown adults can or cannot drink milk.

    1. Interesting point. It sounds somewhat like pigmentation traits which increase expression from birth and then slowly peter away after maturity. Obviously, people with the lactose intolerant alleles are able to break lactose down at birth and then it slowly fades in childhood.
      Within the frame you mention, it would sound as if LP was originally selected against older children. As you suggest, "adulthood persistence" may be a red herring.

    2. Well, I'm founding my opinion on a single case I know, who is now 19 and only experienced intolerance symptoms since a year ago (that I know) or few (that I don't know). Yesterday I suddenly realized it's rather a late onset.

      As for selection, I presume that producing lactase is somewhat costly and before Neolithic there was absolutely no reason to have it beyond X age, so initially the allele which saved a minor cost was probably selected for, even if weakly so. Also lactose tolerance seems to favor obesity (but not mere overweight), that can be another contra but not in times of famine, of course, when those extra pounds may save your life in fact.

    3. Lactose intolerance in mammal children is rare since milk's primary purpose is to feed mammal children. If you were an adult living in an area where other sources of protein, fat and/or carbohydrate were rare then it would probably be important whether or not you could consume milk and dairy products.

    4. interesting thought - in a live fast, *reproduce young*, die young environment an early advantage might be the critical advantage.


      "I'm founding my opinion on a single case"

      Did they recently change how they took milk e.g. did they used to take it with cereal and more recently take it on its own or anything like that? I just wonder if there are certain foods that partially negate the effects of intolerance.

    5. The guy had to stop having milk altogether, as well as dairy cream, because it got him sick. I advised him a bit: yogurt, cured cheeses... Yogurt is generally safe because the bacteria eats all the lactose but cheeses depend on the type and the person (not everyone's intolerance is the same). I also advised him to do medical testing because intolerance can be confused with allergy, which is more dangerous.