Monday, July 10, 2017

Armenian confirmation bias


Current Biology recently published a paper by Margaryan and Derenko et al. titled Eight Millennia of Matrilineal Genetic Continuity in the South Caucasus. I wasn't going to bother calling out the authors on their, unfortunately I have to say, rather dubious claim, but then I saw this ScienceNordic article enthusiastically attempting to drive home their misguided point, so a few words are now in order.

“It’s basically the same female population in the region over the past 8,000 years. It’s very surprising considering the many waves of migration and cultural shifts,” says lead-author Ashot Margaryan from the Centre for GeoGenetics at the National History Museum of Denmark, University of Copenhagen.

Genetics have remained constant for 8,000 years in world’s melting pot

I'm at a loss as to why Ashot Margaryan is very surprised. I'm not even mildly surprised. Why? Let's take a closer look at what we're dealing with here:

- the authors sequenced just 52 full mitogenomes to represent 8,000 years of prehistory and early history in the South Caucasus

- they lumped all of these sequences together into an "Ancient" sample set as if they were from a single time slice (I know, pretty crazy)

- they then ran a few complex models on this neither here nor there sample set, and concluded that it resembled the maternal gene pool of present-day Armenians.

Well, duh, present-day Armenians are more or less the end product of the population history of the last eight thousand years in what is now Armenia and surrounds. Is anyone still as surprised about this as Ashot? Surely not.

Obviously, the problem here is that the authors have mistaken their none too surprising outcome to mean that the South Caucasus has not experienced any major upheavals in its maternal gene pool over the past 8,000 years, which, if actually true, would indeed be very surprising, and even shocking.

But the haplogroup assignments of the 52 mitogenomes are reported in the spreadsheet here, and just by eyeballing these results, I can tell that they suggest an influx of foreign ancestry, probably from the Pontic-Caspian steppe, to the South Caucasus after the Early Bronze Age (EBA). Note, for instance, the presence of what appear to be typically steppe haplogroups U4a, U2e1e and U5a1b in the samples dated to the Middle Bronze Age (MBA), Late Bronze Age (LBA) and Early Iron Age (EIA), respectively.

Citation...

Margaryan and Derenko et al., Eight Millennia of Matrilineal Genetic Continuity in the South Caucasus, Current Biology 27, 1–6 July 10, 2017, DOI: 10.1016/j.cub.2017.05.087

26 comments:

  1. Neolithic British mtDNA is basically the same as modern British mtDNA. But of course we know there's hardly any Neolithic British derived mtDNA in modern Britain. 52 ancient mitogenomes from Armenia can't prove or disprove mtDNA continuation in Armenia.

    @Everyone,

    It's time people realize the mitochondrial chromosome is the least ancestrally informative chromosome in DNA. And this is coming from a guy who has collected 30,000 mtDNA samples. There is no mtDNA equivalent of R1b P312 or J2. Usually the mtDNA of populations in the same large region is pretty much exactly the same.

    A 100 mtDNA samples from ancient or modern people can't rarely ever tell us very much.

    ReplyDelete
  2. Yeah, but the methods in this paper are off the wall considering what they're testing and what they're claiming.

    Crazy.

    ReplyDelete
  3. Of course I ridiculized this pretension, and could do more, but then you say that I am not respectful with all these said scholars:
    There is also an mt K1a1b1e like mine but with the extra mutations 114T 189G 16093C. I have the extra mutation 11204C, but basal subclades are found in Tuscany. There is someone who does think that this haplogroup came from Caucasus?
    3. HQ176413(Italy) FTDNA K1a1b1e 14-SEP-2010
    A73G A263G 309.1C 315.1C C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8860G G9055A T9698C G9932A A10398G A10550G T11204C T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16224C T16311C T16519C
    An American woman, linked to my mother not more than 5 generations ago, hasn't the mutation in 11204, thus demonstrating that also my Tuscan line was basal only a few generations ago, and other basal subclades are in Tuscany and in Sicily from Pisans emigrated to Sicily after the Florentine conquest of Pisa at the beginning of the 15th century.
    Anyway the pathway of this Armenian K1a1b1e is from Italy for the mutation in 16093C:
    5. JQ704056 Behar K1a1b1e 07-APR-2012
    A73G A263G 315.1C C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8521G A8860G G9055A T9698C G9932A A10398G A10550G T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16093C T16224C T16311C T16519C
    7. JQ706057(Italy) Behar K1a1b1e 07-APR-2012
    A73G A263G C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8860G G9055A T9698C G9932A A10398G A10550G T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16093C T16224C T16311C T16519C
    Armenian K1a1b1e is only 3400-3200 years old, i.e. very likely belonged to the Indo-European component in the Armenian pool come from the Balkans (29% of the Armenian autosome).
    Mine is a sample as to an inquiring about Y and mt DNA should be done: not on haplogroups percentage, but each haplotype in its pathway.
    Harold, I am writing something to me about this subclade, in Italian because I wasn't going to post that. Of course I make some hypotheses for looking for experiments, which only may confirm or disprove them. It's a shame that others take them not for what they are (hypotheses at the light of the reason, but that need proofs) but as a signal of my nationalism.
    In K1a1b1e this sample gets C114T:
    6. JQ704207(Ireland) Behar K1a1b1e 07-APR-2012
    A73G C114T T146C A263G 315.1C C497T C522- A523- A750G T1189C A1438G A1811G A2706G A3480G G3705A A4769G C7028T A8860G G9055A T9698C G9932A A10398G A10550G T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16224C T16311C T16519C
    3. HQ176413(Italy) FTDNA K1a1b1e 14-SEP-2010
    A73G A263G 309.1C 315.1C C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8860G G9055A T9698C G9932A A10398G A10550G T11204C T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16224C T16311C T16519C
    Possiamo immaginare che da K1a1b1, con 114 in eteroplasmia, si ebbe la mutazione G9932A. Due linee si separarono: una rimasta in Toscana/Italia, e l'altra emigrata verso l'Irlanda, forse con gli agricoltori di 7500 anni fa verso l'Iberia. Nell'aplotipo iberico/irlandese si fissò C114T, mentre in quello italiano no. E' quindi da immaginare che i K1a1b1e inglesi siano arrivati posteriormente dall'Italia, anche forse in epoca romana.

    ReplyDelete
  4. Il mio aplotipo è basale, poiché la mutazione T11204C non è presente in una linea americana da Stàffoli (Pisa) e legata alla mia non più di 5 generazioni fa.
    L'altro marker in eteroplasmia era 16093, anch'esso fissato in
    5. JQ704056 Behar K1a1b1e 07-APR-2012
    A73G A263G 315.1C C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8521G A8860G G9055A T9698C G9932A A10398G A10550G T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16093C T16224C T16311C T16519C
    7. JQ706057(Italy) Behar K1a1b1e 07-APR-2012
    A73G A263G C497T A750G T1189C A1438G A1811G A2706G A3480G A4769G C7028T A8860G G9055A T9698C G9932A A10398G A10550G T11299C A11467G A11470G G11719A G11914A A12308G G12372A C14167T C14766T T14798C A15326G A15924G T16093C T16224C T16311C T16519C
    e l'aplotipo basale è ancora in Italia (JQ706057).
    Il K1a1b1e trovato nel DNA antico in Armenia, con T16093C fissato, deriva quindi dall'altro aplotipo italiano o proveniente dai Balcani. Questa migrazione può essere parallela e collegata con quella di alcuni aplotipi di Y R-L23 presenti posteriormente a Samara (Russia). 7500 anni fa può essere una data probabile.
    Nell'aplotipo armeno si era fissato anche C114T.
    The H15a1 sample, 7900-7600 YBP, is in line with Behar 2012b ages

    H15a1 6136.3 2389.3

    and is a basal clade
    21. MF362769(Armenian) Margaryan H15a1 29-JUN-2017
    T55C T57C A263G 309.1C 309.2C 315.1C A750G A1438G A4769G T6253C A8860G T11410C C14953T A15326G

    H15 has been found in Iran and Yemen
    1. KC911292(Iran) Derenko H15 20-NOV-2013
    T55C A56T T57C A263G 315.1C A750G A1438G A4122G A4769G T6253C A8860G G14384A A15326G T16092C T16357C T16519C
    2. KM986579(Yemen) Vyas H15 21-MAR-2015
    T55C T57C T131C T152C A263G 309.1C 315.1C A750G A1438G G3693A A4122G A4769G T6253C T7609C A8860G C9731T A15326G T16092C T16519C

    i.e. the oldest diffusion of the haplogroup, but H15a is European and also H15b, with diffusion of H15b1 and H15b2 both to Europe and Caucasus.
    With these data it is difficult to say that H15 was born in Middle East, Caucasus or Iran rather than Europe. Of course only the aDNA of the oldest subclades could say more. It is also difficult to say that H15a1 was born in Armenia and it is there since then, but of course H15a1a, H15a1a1, H15a1b may have had an expansion from Caucasus.
    CRS A263G A750G A1438G A4769G A8860G A15326G
    H15 T55C T57C T6253C
    1. KC911292(Iran) Derenko H15 20-NOV-2013
    T55C A56T T57C A263G 315.1C A750G A1438G A4122G A4769G T6253C A8860G G14384A A15326G T16092C T16357C T16519C
    2. KM986579(Yemen) Vyas H15 21-MAR-2015
    T55C T57C T131C T152C A263G 309.1C 315.1C A750G A1438G G3693A A4122G A4769G T6253C T7609C A8860G C9731T A15326G T16092C T16519C
    H15a T11410C
    H15a1 C14953T
    H15a1a C11143T
    H15a1a1 C16366T
    H15a1b 44.1C T146C G11914A C16184T
    H15b T3847C
    H15b1 A15715G
    H15b2 T8699C T13818C
    Of course to say that KC911292 and KM986579 are H15* is only an artifice due to the lacking of a third sample with some of the three mutations (A4122G T16092C and T16519C) they have in common. These samples belong to a possible H15c come to Iran and Yemen after a long journey we don't know from where.

    ReplyDelete
  5. This will be a shock: but I agree with everything Davidski posted here. This study employed questionable methodology, at best.

    It raises an important question too, in academia, which I don't think is discussed frankly often enough. The lead author of the study, as well as its biggest promoter, are themselves of Armenian heritage. It seems that there are powerful psychological and ethnocentrist reasons why scientists of a certain heritage would argue for continuity and "pureness" in their region.

    As an aside, if you look at many of the papers that argue for big Northern and Southern Italian differences, they are authored by Northern Italians. We see this type of confirmation bias a lot.

    ReplyDelete
  6. @ moreisbetter
    "As an aside, if you look at many of the papers that argue for big Northern and Southern Italian differences, they are authored by Northern Italians. We see this type of confirmation bias a lot".

    I find this adfirmation only stupid, for many reasons:
    1) differences between Northern and Southern Italy there are, due to different histories, and also to the high extension in latitude
    2) I'd say also because due to two different refugia: one in the North (Villabruna) and the other in the South, above all Sicily (many mt hgs in common with other Mediterranean Islands)
    3) many scholars found differences not between North and South but between the Adriatic and the Tyrrhenian parts, separated from the Apennine
    4) I exclude that Italian scholars are "racist" between North and South of Italy
    5) who is plotting for separating Italians are people from out of Italy, and you should know why
    6) fortunately amongst the many accusations I suffered, no one said that I, from Tuscany, thus from North, didn't consider with great interest and very important for my theory the data of the whole Italy.

    ReplyDelete
  7. Honestly that paper reeked of nationalistic sort of elements,they kept on mentioning not having Turkish / East Eurasian Mtdna. Some of those Mtdna lineages are clearly from the northern steppe .

    ReplyDelete
  8. There were North Caucasian Huns which were Christianized by Armenians around ~500-600 AD. There were also Christianized Tatars 1000+ AD. I've seen a lot of Armenians they can range from Chechen to Dark Skinned Arabs.

    ReplyDelete
  9. mtDNA has very low information content. But... it is the easiest and cheapest to sequence from both modern and ancient samples.

    And the scientific groups most concerned with money... also usually have the lowest quality work. It's a feedback cycle.

    But on the other hand, it is a totally different kind of information, and it is what it is. You can't make the sequencing better than it. It's just a fact.

    ReplyDelete
  10. This paper is at least as much about Nagorno-Karabakh as mtDNA.

    https://en.wikipedia.org/wiki/Nagorno-Karabakh_Republic

    ReplyDelete
  11. @moreisbetter
    You raise an interesting point, although it is not universal that individual scientist favor models amplifying their personal heritage. Case in point: My colleague, Peter Underhill, is of Polish ancestry--his name is an Anglicized version of under the foothills (something like Podgorny--maybe Davidski knows the Slavic name), yet Peter proposed that R1a1-M417 originates in Central Asia/India. Also, I'm African-American and I'm interested in the Near East/Greek Neolithic/India-IVC. So more cosmopolitan scientists may surpass the biases of their ancestry.

    ReplyDelete
  12. The paper is about trying to paint Armenians as the most autochthonous people in the region.
    Had Azeri scientist ran this he would have gotten similar "results" albeit with the "conclusion" that its who have enjoyed full ethnic continuity

    ReplyDelete
  13. Please don't discus or speculate about the people posting in the comments here. Take it elsewhere, like e-mail.

    ReplyDelete
  14. @ Davidski

    You are right. The blog is yours, but you, like YFull with me (not counting all the other blogs and fora which banned me) cannot censure us and not the others. Of course Roy King is an African-American from the mother side, but I suppose he has also an Y... and ak2014b has good arguments. Otherwise only the "truth" of these persons is ligitimate, and never ours. I think I have good arguments also from a scientific point of view. If the book of a great phisicist is traslated into English, read that (Carlo Rovelli, L'ordine del tempo), and all its exergos from Horatius. We had Horatius, Seneca and many others in those times...

    ReplyDelete
  15. From the paper:

    "Coalescence-based analyses suggest that the population size in this region rapidly increased after the Last Glacial Maximum ca. 18 kya."

    There's a major problem with this blind use of the term "LGM".

    Evry updated climatologist and ice-age-geologist knows that the Eurasian populations of people and other terrestial mega-fauna was heavily reduced during the LGM. But they also know that a later - and FAR worse decimation - happenned during the Younger Dryas, 13-12.000 years ago.

    The big bottleneck and the consequent "re-start" of the surviving species from palearctic Eurasia - such as bears and boars, caprines and Caucasians - occurred NO more than 12.000 yrs ago.

    According to well documented facts brought forth by several scientific diciplines.

    Which means that some 98-99% of all y-dna-lines and mt-pools from paleolithic Eurasias went EXTINCT - and has no DIRECT influence on the human genome of the mesolithic and neolithic Eurasia.

    Whatever "coalescence-time" some geneticians wants to advocate.

    Geneticians who want to be viewed as straight and accountable scientists better get their professional premisses and preliminary context right.

    Otherwise their network and discriminant analysis may end up as biased blatter - turning "principal components" into tools of manipulation.

    ReplyDelete
  16. @Gioiello. This thread is about confirmation biases and how they played into this paper, and I raised the fact that they are often present, which results in bad science. The facts remain that Europeans are an extremely homogenous grouping of people. Extremely homogenous. Italians, as a peninsular people, are an extremely homogenous subset of that extremely homogenous grouping of people. Yet there are so many papers on Southern Italian "otherness" that it has become something of a meme on the Internet. Why is that? Why so much ink spilled?

    I have found that biases (perhaps subtle, perhaps subconscious, and perhaps political) infiltrate academic studies on DNA. And the discussion of this topic is essential, because this discussion is about the validity of the SCIENCE.

    Here, as others have pointed out, there may be an agenda behind this paper. That calls into doubt the validity of its findings, which Davidski has pointed out, already seem quite stretched. The title of this thread is about confirmation bias, right?

    Surely you have noticed a large number of papers by Ashkenazi authors noting the deep ties that Jews have to the Middle East. Now while I 100% agree with that, I don't think that anyone out there doubts that those papers are about Israel.

    This paper is likely about Nagorno Karabakh.

    As for the Italian theme, Northern Italians have a bit of a chip on their shoulders, because they encounter people who stereotype all Italians based on the actions of a few uneducated southern Italians. I find both sad. First, the stereotypes, and second, how it motivates northerners to emphasize how different the groups from their nation is. I can find hundreds of papers on this theme. Yet maybe 1 for Sweden; 1 for Germany; 1 for France, etc. Those are big nations, with the same clines. Yet the confirmation bias of certain researchers comes into play way more for the Italian thing.

    It is up to all of us who care about learning to identify these confirmation biases in any paper, and to call them out.


    ReplyDelete
  17. I didn't want to call it out but there is definitely/absolutely a Armenian confirmation bias in many recent DNA studies. Reminds me of the early Yamnaya papers claiming the Near Eastern part of Yamnaya would be Armenian like while there were literally half a dozen population that fitted better as proxy for this admixture.

    ReplyDelete
  18. mooreisbetter

    "It raises an important question too, in academia, which I don't think is discussed frankly often enough. The lead author of the study, as well as its biggest promoter, are themselves of Armenian heritage. It seems that there are powerful psychological and ethnocentrist reasons why scientists of a certain heritage would argue for continuity and "pureness" in their region."

    This, saying that modern day Armenians or surrounding populations inherited allot of ancestry from people living in the same region 8000 years ago is one thing, but claiming there is a continuity, as if no sign of genetic change of the landscape, which is obviously bogus, is another.

    ReplyDelete
  19. continue..

    there is absolutely no modern population that can claim for a continuity of it's ethnicity for over 6000 years, let alone 8000 years.

    If you say population A has inherited allot of ancestry from people living in the region 8000 years ago that is correct.

    But as some people correctly suggested here to me it looks like some individuals within the scientific landscape try to paint their own ethnicity as the most ancient in the region for obvious political reason. I have seen studies made by Turkish or Iranian scientists show a similar tendency.

    I know for example that there are tensions between Azeris and Armenians.


    ReplyDelete
  20. @ mooreisbetter

    Of course I completely agree with you that many papers (perhaps pretty all) are biased, I didn't agree with what you said that Italian geneticists from Northern Italy were biased about Southern Italians. That isn't true, above all for the fact that pretty all the Italian geneticists are leftists, and "racism" isn't the ideology of leftists. Of course what you said is true: "Yet there are so many papers on Southern Italian "otherness" that it has become something of a meme on the Internet. Why is that? Why so much ink spilled?", but the explanation about that is complex, and to say that it is due to Northern Italian geneticists is a semplification and not true. You could find the answer in my thousands of letters written so far, but they are in blogs mostly out now (aren't you asking why?). Of course I don't agree with your statement that "Surely you have noticed a large number of papers by Ashkenazi authors noting the deep ties that Jews have to the Middle East. Now while I 100% agree with that, I don't think that anyone out there doubts that those papers are about Israel". I think that this question has much to do also with the former one, and very likely isn't the less important.
    About the Armenian mt I think having given some samples as to how we may disprove those statements. I did that above all about my mt K1a1b1e, but also H15 and I could do for many other haplogroups, and who knows me should grasp the link also with the Y I think coming from Italy or Western Europe and not Caucasus or Samara, but we have had enough theories so far: I am waiting that these scholars test, as I said, Tyrrhenian Italy from 12000 to 5000 years ago. We'll speak again then.

    ReplyDelete
  21. George

    Most probably You didn't read the paper, otherwise You would note this part.

    ""one would expect to observe a somewhat similar pattern of matrilineal genetic continuity in other parts of this region, i.e., Georgia, Azerbaijan, and Armenian Highland (partially modern day Eastern Turkey and North-West Iran). Future studies should, however, prioritize to expand the sampling of both modern and ancient populations in the whole region to uncover the geographic and temporal extent of this genetic continuity signal.""

    http://www.sciencedirect.com/science/article/pii/S0960982217306954?_rdoc=1&_fmt=high&_origin=gateway&_docanchor=&md5=b8429449ccfc9c30159a5f9aeaa92ffb

    Do this mean that they claim that Georgians and Azeris are also the most autochtonous?

    ReplyDelete
  22. mooreisbetter

    ""This paper is likely about Nagorno Karabakh.""

    Woow. And Olalde's paper is about leftism and promigrationistic politics in Europe. :)

    http://eurogenes.blogspot.am/2017/07/ashg-2017-crowdfunding-drive.html

    ReplyDelete
  23. The big problem is that the paper doesn't prove 8,000 years of total, or even near total, maternal/mtDNA continuity in Armenia.

    It's an interesting question who was reviewing this paper and how they missed this problem.

    ReplyDelete
  24. Davidski

    **It's an interesting question who was reviewing this paper and how they missed this problem.**

    I also ask myself this question. Certainly I don't agree with their conclusions. And I see nothing pro-Armenian in their claims that cultural and linguistic shifts occurred in SC without genetic shifts ( in this case mitogenes ).

    But in general I am not very surprised. After all Elhaik managed to publish his very questionable paper about Druzes in Nature journal. And in Hoffmanova paper they introduced the idea of independent invention of farming in Aegean just based on 2 mitogenomes ( K1c ) that we now learn were widespread in Mesolithic Balkanes in purely WHG context.

    ReplyDelete
  25. Btw Reich lab also has samples from same periods ( Neolithich, Chal, EBA) Afaik they will publish high quality genome wide data. So we can compare their conclusions to this one.

    This team is also currently working on extracting genome wide data. And as Ashot Margaryan said in that interview their conclusions can change.

    But one thing is sure. The comparison of Neolithic versus Moderns will have very different dynamics in South Caucasus compared to Anatolia, Levant, Iran.

    It is quite possible that the closest pops (by FST) to SC Neolithic will be the same populations that live today there or in neighbourhood. Wee need to wait and see.

    ReplyDelete
  26. Good methodology seeks to avoid confirmation bias, despite individual subjectivity, and cultural and ideological leanings. So the results of this study should be examined with regard to the methodology and then the results. It is interesting that few question 'Caucasian' ethnocentric biases in Western Europe or the US, where "Caucasian" is a label that has lost its connection with the peoples of the Caucasus and Transcaucasus, and the Western anthropologist responsible for the creation of the label.

    ReplyDelete

Read the rules before posting.

Comments by people with the nick "Unknown" are no longer allowed.

See also...


New rules for comments

Banned commentators list