Sunday, December 11, 2011

Participant results for 'K12a' calculator

The participant results can be found in the "Individual Results" tab of the K12a spreadsheet.
You can read more about the K12a calculator at my other blog; if you are not a Project participant, you can also find a DIY version of it there, which can be used in conjunction with DIYDodecad 2.1.


  1. Hi Dienekes', just two questions:

    ¿don't you think the Caucasus and Gedrosian components are overdimensioned in Southern Europe? Specially while looking at the Iberian results, it's hard to believe they have more of this admixtures together than the vast majority of Northern Europeans.

    This is fairly different from what we saw in the v3 results. Second question then is if the Dodecad v4 will fix some of this things if finally based on the same model (or similar).


  2. They are what they are, I have no reason to think that they are over- or under-represented in any particular region.

    If they are in any way different than the v3 results, the K12a results are more trustworthy as a truer representation of Eurasian variation.

  3. What is this Gedrosia thing? Does it mean probable Gypsy admixture?

  4. It is a component that reaches its maximum in the populations of ancient Gedrosia, Balochs, Brahuis, and Makranis. I see no particular reason to associate it specifically with Gypsies, although anyone with Gypsy ancestry will probably have elevated levels of this component, since it occurs at a much higher level throughout South Asia compared to Europe.

  5. I'm allegedly 6.81% Gedrosian. All hte other calculators said I was roughtly 6% SW Asian.

    Now, first if that's really what it means, I bet it's approximate, based on genetic findings for some sample that happened to come from there, and it applies far more broadly to the general region.

    Second, there is NOTHING here that tells if that's what it means, or what it does mean. I don't know if your breakdown is simply a list of populations that genetically resemble you, or that genetically resemble some other population like that of Britain, or that of Northwest Europe.

    So, finding out that you're 6.81% Gedrosia could mean 6.81% of your ancestors allegedly came from Gedrosia, or it could mean that the people of Gedrosia 6.81% resemble the people of Britain where you actually came from. Like,???? I wouldn't really worry about it without clearer information.

    It isn't even clear if you're supposed to subtract values from each other to get actual percentages.

    I mean, I don't know what Dieneke thinks we're supposed to interpret this stuff with, because there is no guide to interpreting it anywhere that I can find, and there aren't clear instruction son most of it either. I mean, what does K=12 mean? Something you're supposed to plug into one of the programs somehow?

  6. Second, there is NOTHING here that tells if that's what it means, or what it does mean.

    Here is a primer on what sharing an ADMIXTURE component means

    A variety of technical articles can be found at the bottom of the blog.

  7. Dienekes,

    There's a rumor making the rounds on DNA-Forums that you need to change the first line of K21a.par as follows:
    From: 1d-7
    To: 1d-11

    Care to comment?

  8. That is a parameter that can be changed. You can run it in "progress" mode to see whether the percentages are still changing noticeably and whether the extra precision is worth the trouble.

    For the purposes of comparing against the reference populations, I recommend (at least for participants) to use the spreadsheet values, for a couple of reasons: (1) these were calculated with the exact same termination criteria by ADMIXTURE, and (2) the population averages and the individual results are both calculated with the exact same genotype files I have received, whereas testing companies occasionally fiddle with their genotype files.

  9. Thanks, There is a lot to digest here.

  10. @Acid if you look at "fst" you will see that Gedrosia and Caucasus are very similar (probably the same ancestor) I believe that "Gedrosia" represents a Caucasian Gene flow into South-Central Asia and is comparable with "ANI" of the "South Asian" component just more West Asian as ANI was. I would generally not connect "components" too "literally". The names tell you were this component reaches its highest but this doesnt mean that they actually "developed" there. Take Ukrainians_Y as example, on dodecad v3 they had more of the "West European" component than "East European".

  11. I understand it Kurti, however, I still think there's something strange there. Perhaps this Northern European cluster is masking some of the Asian affinities, and that's why Northern Europeans now look (generally) less "Asian" than before. Actually, Mediterranean seems to be better defined as representative of Europe, but it's just personal apreciation.

    Also, the example of this Ukranians and the v3 run isn't probably the best, since there was some kind of controversy around it. Anyways, I got the message.

  12. I am trying to understand my grandmother's results. In her mixed mode she came out to be: [1,] "74.4% N_Italian_D + 25.6% Bantu_N.E." "11.1848" I can understand the North Italian. The N.E. Bantu I am not clear about. I looked at the graphs and it showed the NE Bantu with some 30% East African and some 5% SW Asian. I am trying to picture this geographically to try to understand better my grandmother's origins. I am almost certain she has ancient Sephardic Jewish roots. If her NE Bantu places her near Ethiopia and what not, then this would make more sense to me.

  13. >> I am trying to understand my grandmother's results. In her mixed mode she came out to be: [1,] "74.4% N_Italian_D + 25.6% Bantu_N.E." "11.1848" I can understand the North Italian.

    If your grandmother isn't a quarter-black, you probably entered the numbers wrong. Make sure you enter all the 12 numbers in order (including zeros).

  14. I doubled checked all the numbers. Out of the 1132 joint cousins that we do have on 23andme, only several of them are primarily south or west African. When I analyse the cMg's we have in common with them using the by segment texts, the ethnic profile that shows is that of a Eurasian/Caucasian/Mediterranean mixed with Native American. Only one 8cMg segment shows as Mulatto. All her distant cousins tend to be Ashkenazi, American, European or Caribbean. This is going all the way back to the very beginnings of the Transatlantic slave trade. The Caribbeans she and I have are overwhelmingly Northern or Southern European. European genes dominate Puerto Rico. I believe she is primarily of white Canariote stock (I used your k12 analysis on another verified Canarian's genome, and the Canarian's minority percentage showed up as Bantu as well. The Canariote is of white Cuban stock that has not blended with African stock in Cuba.). I can trace her back genealogically to the 1790's. Her family was linked to other's of verifiable Canariote Stock. So my conclusion is that her African may be a product of the Transaharan slave trade. Her Ancestry is Canarian/Sephardic/Spanish/Berber. I believe her ancestors were one of those that fled the blood cleansing Spanish Inquisitorial progroms. These are her numbers: DodecadOracle(c(27.70, 4.09, 3.43, 21.74, 0.84, 16.94, 8.98, 2.79, 3.45, 4.95, 0.71, 4.39), k=30 . I am her grandson, yet my mixed mode shows up as 79.5% Dutch and 25.5% Egyptian. I have no recent Middle Eastern heritage. My paternal lineage is of British descent so the Dutch makes sense. But as I have said before, years of research have lead me to the conclusion that our ancestors did enter Iberia over a thousand years ago as Sephardim. Having her ancestors enter history via Ethiopia into North Africa and unto Iberia before emigrating to the Islands is the most plausible trajectory I can make for her unique blend of European, Mediterranean and African.

  15. Charles is in denial about his recent Subsaharanness. Chucky, your grandmother is Puerto Rican. Her k12a West African is 17%. That is not atypical. Deal.

  16. No Brandon. I never said that is not Subsaharan. North East Bantu extend to the edges of what is currently Ethiopia. As Dienekes explained about L3, L3 may have originated in Arabia. I am mitochondrial L3b1a. This may have originated in Central Africa. This is not a Black vs White issue. It is about understanding the many ways that the African genome entered into the European scene. What many forget is that Islam played a very important role via the Transaharan slave trade. Women slaves were favored two to one over men in Islam. They served as servants and wives. This is a given fact. Just about all the L's in my extended family are Mitochondrial L. Most of those L's can be genetically classified as Southern European (As my grandmother is.). North Americans have had an essentially Black/White view of history. Things were not so clear cut in the Iberian empires. I would not do Iberia or Latin America justice by imposing such a overly simplistic view point.

  17. Chuck, you have the most convoluted theories here. SEVENTEEN percent is way too large to be a smidgen of hidden black in the Iberian gene pool. You keep saying trans-Saharan, as if some small traces of maternally transmitted "African genome" came into Spain from the Muslim era, and somehow that got "diluted" *up* to a whopping near-quarter of grannie's inheritance.

    It's obvious that her black ancestry is recent and part of the trans-ATLANTIC slave trade and that it was Sephardic Jewry who brought that fully-African set of slave ancestors of yours directly to the Carribean.

  18. Not necessarily Brandon. The fact is that the Berber average can be anywhere from 0-40%. It is also a fact that my ancestors were isolated on Islands for the last 500 years. This I believe have isolated much of their genome from the mayor changes that occurred in the Spanish metropolis after 1492. It is also a fact that only 4 out of 17 mitochondrial L's are predominantly African. The other's are predominantly Southern European. I also have a Canarian cousin that has up to 11% SSA. He verifies that his family has not had any recent exchange with Africans. This is entirely possible. People on the islands did cluster into tight nit groupings. And again, I do not have enough predominantly African cousins in my extended network of cousins to account for all that extra SSA. Even today in remote areas of Southern Portugal and Spain you can encounter outliers with a large amount of SSA. So, as convoluted as you may think it is, 5 years of research have led me to this conclusion. The fact is that the blood cleansing progroms in Spain were very effective in white washing Iberia's genetic and cultural history.

  19. I meant to say that 4 out of 14 mitochondrial L's that show up in my extended network of cousins are predominantly West or South African. Products of the transatlantic slave trade I presume. I do assume a possibility of 1-5% SSA entering my grandmother's genome via drift in the form of intermarriage with mulattoes and other free people of color (I have been able to determine at least two mulatto insertions into my grandmother's genome using Dienekes by segment analysis. I also have one cousin who declares her ancestor as being a free woman of color.) What I do not accept, based on the evidence I have gathered, is that all or the majority of her SSA is due to the transatlantic slave trade. The numbers do not add up! 75% of the puertorican population descend of Canariote stock. These Canarians are already distinguished from the rest of the tribes in Iberia as having a large amount of both NW African and E African in their genome. This is a fact. My grandmother's top tribes are Portuguese and Tuscan (Genovese). This is the same for other cousins of Canarian descent I have studied as well as an original Canarian that can trace his lineage back 500 years. My grandmother's SSA does not affect her overall genomic picture, just like Europeans are not affected by their 12% middle eastern. That reminds me that my Grandmother is almost 30% middle eastern and she is a carrier of a rare condition called Familiar Mediterranean Fever. This is carried mostly by those of sephardic descent. Over two thirds of her surnames are those of Sephardic descent. Bla bla bla. I can go on and on. The percentages do not matter as much as the fact that the component is there. The true meaning of science is to be able to distinguish one thing from another. I am distinguishing between what she may have obtained from the transaharan slave trade from what may have been added unto her via the transatlantic slave trade. I say, from what I have gathered, that up to two thirds of her SSA is transaharan and one third is due to transatlantic SSA drift. I believe this is an educated guess and I will stand by it until there comes such a time when we can accurately place a time line on each component of a genome.

  20. The rumor about changing the parameter file to 1D-10 was purely based on my request to a project participant to see if it made any difference in their results as I have heard a lot about the DIY results corrupted with noise. A few of them ran it and in their case they found that the noise is not a big issue.

  21. I am clearly out of my element here so please bare with me. I have a slight understanding of what an admixture is, although not nearly as much as I thought before reading these comments, but I am trying to learn. I typed in irish as a substring and my results showed 43% North European, 39.66% Atlantic Med. I'm not sure how that specifically answers the Irish substring question but it does fall in line with all other admixture projects I've compared my DNA to. I also have repeatedly had Caucasus show up in decent numbers but this time it was lower (3.34%) and Gedrosia was11.36%. This is the first time I've heard of Gedrosia but it is related to Caucasus, yes? I'm not asking that anyone do my homework for me, but I would certainly appreciate anyone who could shed light on what my results mean for me in a simpler, "laymans" way.

  22. I suspect that I have a remote Jewish ethnicity. Here are my results from GEDMatch:

    Dodecad K12B Admixture Proportions

    Gedrosia 10.34%
    Siberian -
    Northwest_African -
    Southeast_Asian -
    Atlantic_Med 40.86%
    North_European 40.63%
    South_Asian 0.71%
    East_African 0.28%
    Southwest_Asian 0.50%
    East_Asian -
    Caucasus 6.08%
    Sub_Saharan 0.60%

    Gedrosia K11 Test

    Single Population Sharing:

    # Population (source) Distance
    1 French 5.99
    2 Bulgarian 12.75
    3 Norwegian 16.91
    4 French_South 20.4
    5 Albanian 20.86
    6 Greek 24.13
    7 Sicilian 29.72
    8 Ashkenazi_Jew 31.14
    9 Russian 32.12
    10 Finnish 35.16
    11 Estonian 37.22
    12 Lithuanian 38.44
    13 Tajik_Pomiri 39.18
    14 Turkish 39.58
    15 Turkmen 41.55
    16 Turkmen_Afghan 41.93
    17 Uzbek 43.3
    18 Tajik_Afghan 44.17
    19 Uzbek_Afghan 44.72
    20 Cypriot 45.83

    Gedrosia K15 Test

    Single Population Sharing:

    # Population (source) Distance
    1 French 6.6
    2 English 6.99
    3 Norwegian 13.88
    4 Hungarian 14.64
    5 Croatian 14.72
    6 Bulgarian 14.73
    7 Czech 16.55
    8 Spanish 17.58
    9 Bergamo 21.75
    10 Albanian 22.22
    11 French_South 22.81
    12 Greek 25.38
    13 Ukrainian 26.46
    14 Sicilian 31.05
    15 Belarusian 31.54
    16 Mordovian 31.65
    17 Russian 32.27
    18 Finnish 33.13
    19 Ashkenazi_Jew 33.56
    20 Estonian 37.5