Page 10 of 10 FirstFirst ... 8910
Results 91 to 99 of 99

Thread: Towards a better modelling of NorthWest Europeans in Global 25

  1. #91
    Registered Users
    Posts
    785
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by anglesqueville View Post
    Huijbregts kindly made the job for my parents.

    A B C
    angle_mom 0.988 0.010 0.002
    angle_dad 0.094 0.111 0.795

    My mum is as expected. My dad is dramatically shifted toward eastern Europe. Of course, the first explanation that comes is his Finnish part. But I will add that with many analyses his German part (genealogically as minor as his Finnish part) "eats" his Dutch part. I always explained this phenomenon with problems of sampling, but this explanation is perhaps not sufficient.

    I wish to add a remark. The case of my father is very revealing of the gap that exists between individual analysis and populations inferences. Many friends (I think of Finn and of some French Britton users) have the hope that, from the observation of some individuals, they will be able to speculate about some micro-regional structures. The case of my father shows clearly that some recent genes flow, known or unknown, has the ability to shift dramatically the results in an unexpected direction. If the study is made with a reasonable number of individuals, say enough to compute standard errors and detect outliers, it's not a problem. But if the study is made on a handful of individuals it's not the same story.
    I endorse your warning against 'micro-regional structures' (aptly worded).
    And even if the study is made with a reasonable number of individuals, there remains a problem if the target is a singleton sample.
    Also there are problems when the model is unbalanced (i.e. the subgroups have very uneven frequencies).
    The simple model that I used for this project has 505 samples to estimate 20 parameters. Overall that is a nice relation, but it offers no guarantee that I have sufficient degrees of freedom for each of the parameters.
    Especially since the model is heavily unbalanced.

    Having said all that, I could not resist te temptation to refine my model for your dad.
    What I did not like about the model of your dad, was that his main component was C, which is at the ultimate border of the model ('German_East') and you have good reasons to think that he is much more Eastern.
    So I crafted a VERY experimental model with a fourth component D for 'Polish'; the idea was that if he is more Eastern than German, he would be attracted by 'Polish'. However that didn't materialize: he got stuck in the 'Middle-European' component.
    A B C D
    angle_mom 0.960 0.015 0.025 0
    angle_dad 0.026 0.045 0.929 0

    Maybe the correct migration path is through Swedish. But I fear that in a model Swedish would be eaten by component A.

    To create an impression of the unbalance of the model, I prepared the next density plot.
    density.d.jpg
    The huge peak is component A, which dwarfs everything in the surrounding area.
    At the foreground the 'French/Southern' component B. At the left a small peak of Polish and German_East (C/D).

  2. The Following 8 Users Say Thank You to Huijbregts For This Useful Post:

     anglesqueville (05-09-2019),  Jessie (05-10-2019),  JMcB (05-09-2019),  JonikW (05-09-2019),  randwulf (05-09-2019),  Ruderico (05-09-2019),  timberwolf (05-10-2019),  Trelvern (05-09-2019)

  3. #92
    Moderator
    Posts
    5,567
    Sex
    Location
    Normandy
    Ethnicity
    northwesterner
    Y-DNA
    U152>L2>Z367
    mtDNA
    H5a1

    Normandie Netherlands Friesland Finland Orkney
    Huijbregts, getting G25+nMonte models for my father has always been quite a headache. What I can say: with all combinations that include Finnish and Saami
    1) The Finnish part is always so overestimated that I sometimes wonder whether there could not be an unknown source, besides his gr-gr-mother. Example:

    [1] "distance%=0.3501"
    dad
    German,75.6
    Finnish,9.6
    Finnish_East,8.6
    English,4.2
    Saami,2

    This observation is converging with another: I have stopped counting his (and my) Finnish relatives on FamilyFinder. Every update brings us at least two or three more. Moreover, our Scandinavian (Swedish or Norwegian) relatives are all by segments that are shared very predominantly by the Finns. Btw those of these Scandinavians I have contacted are all from far northern Norway or Sweden. Too bad for my father's "Viking Myth", "his" Scandinavians are very likely mixed with Saamis, and ...

    2) ... the Saami part always resists, even if I add Karelian to the references. And if on the contrary I keep the Saami among the references and throw away the Finnish, the Saami percentage grows up to 6% and more, like on this model:

    [1] "distance%=0.3912"

    dad

    German,93.4
    Saami,6.6


    I often say that my father is a black sheep for the ancestry analyses. It would perhaps be better to call him a black reindeer.

    3) What happens if I add Scandinavian references to the German ones? First, the Norwegian references are rejected (too bad for the Vikings). With the Swedish, the Saami part falls back to roughly 2%:

    [1] "distance%=0.3525"
    dad
    German,79
    Swedish,19.2
    Saami,1.8

    and you see that the mixture of Swedish+Saami (21%) is nearly equal to the mixture Finnish+Finnish_East+Saami of the model 1) (20.2%)

    My conclusion is: my dad's shift towards eastern Europe is very very likely towards northeastern. What puzzles me: how a so minor genealogical part can produce a so dramatical deviation? I have no answer.
    En North alom, de North venom
    En North fum naiz, en North manom

    (Roman de Rou, Wace, 1160-1170)

  4. The Following 8 Users Say Thank You to anglesqueville For This Useful Post:

     fabrice E (05-09-2019),  Finn (05-12-2019),  Helgenes50 (05-09-2019),  Jessie (05-10-2019),  JMcB (05-09-2019),  Nibelung (05-09-2019),  randwulf (05-10-2019),  Ruderico (05-09-2019)

  5. #93
    Registered Users
    Posts
    785
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by anglesqueville View Post
    Huijbregts, getting G25+nMonte models for my father has always been quite a headache. What I can say: with all combinations that include Finnish and Saami
    1) The Finnish part is always so overestimated that I sometimes wonder whether there could not be an unknown source, besides his gr-gr-mother. Example:

    [1] "distance%=0.3501"
    dad
    German,75.6
    Finnish,9.6
    Finnish_East,8.6
    English,4.2
    Saami,2

    This observation is converging with another: I have stopped counting his (and my) Finnish relatives on FamilyFinder. Every update brings us at least two or three more. Moreover, our Scandinavian (Swedish or Norwegian) relatives are all by segments that are shared very predominantly by the Finns. Btw those of these Scandinavians I have contacted are all from far northern Norway or Sweden. Too bad for my father's "Viking Myth", "his" Scandinavians are very likely mixed with Saamis, and ...

    2) ... the Saami part always resists, even if I add Karelian to the references. And if on the contrary I keep the Saami among the references and throw away the Finnish, the Saami percentage grows up to 6% and more, like on this model:

    [1] "distance%=0.3912"

    dad

    German,93.4
    Saami,6.6


    I often say that my father is a black sheep for the ancestry analyses. It would perhaps be better to call him a black reindeer.

    3) What happens if I add Scandinavian references to the German ones? First, the Norwegian references are rejected (too bad for the Vikings). With the Swedish, the Saami part falls back to roughly 2%:

    [1] "distance%=0.3525"
    dad
    German,79
    Swedish,19.2
    Saami,1.8

    and you see that the mixture of Swedish+Saami (21%) is nearly equal to the mixture Finnish+Finnish_East+Saami of the model 1) (20.2%)

    My conclusion is: my dad's shift towards eastern Europe is very very likely towards northeastern. What puzzles me: how a so minor genealogical part can produce a so dramatical deviation? I have no answer.
    Did you try Swedish? For instance:
    distance%=0.6976

    Angles_dad

    Swedish,57
    German_East,41
    Nganassan,2

  6. The Following 4 Users Say Thank You to Huijbregts For This Useful Post:

     anglesqueville (05-09-2019),  Jessie (05-10-2019),  JMcB (05-09-2019),  randwulf (05-10-2019)

  7. #94
    Registered Users
    Posts
    70
    Sex

    Quote Originally Posted by anglesqueville View Post
    Huijbregts, getting G25+nMonte models for my father has always been quite a headache. What I can say: with all combinations that include Finnish and Saami
    1) The Finnish part is always so overestimated that I sometimes wonder whether there could not be an unknown source, besides his gr-gr-mother. Example:

    [1] "distance%=0.3501"
    dad
    German,75.6
    Finnish,9.6
    Finnish_East,8.6
    English,4.2
    Saami,2

    This observation is converging with another: I have stopped counting his (and my) Finnish relatives on FamilyFinder. Every update brings us at least two or three more. Moreover, our Scandinavian (Swedish or Norwegian) relatives are all by segments that are shared very predominantly by the Finns. Btw those of these Scandinavians I have contacted are all from far northern Norway or Sweden. Too bad for my father's "Viking Myth", "his" Scandinavians are very likely mixed with Saamis, and ...

    2) ... the Saami part always resists, even if I add Karelian to the references. And if on the contrary I keep the Saami among the references and throw away the Finnish, the Saami percentage grows up to 6% and more, like on this model:

    [1] "distance%=0.3912"

    dad

    German,93.4
    Saami,6.6


    I often say that my father is a black sheep for the ancestry analyses. It would perhaps be better to call him a black reindeer.

    3) What happens if I add Scandinavian references to the German ones? First, the Norwegian references are rejected (too bad for the Vikings). With the Swedish, the Saami part falls back to roughly 2%:

    [1] "distance%=0.3525"
    dad
    German,79
    Swedish,19.2
    Saami,1.8

    and you see that the mixture of Swedish+Saami (21%) is nearly equal to the mixture Finnish+Finnish_East+Saami of the model 1) (20.2%)

    My conclusion is: my dad's shift towards eastern Europe is very very likely towards northeastern. What puzzles me: how a so minor genealogical part can produce a so dramatical deviation? I have no answer.
    complete shot in the dark but maybe it's from multiple sources or the group the ancestor came from had uncommon admixture from another Uralic group separate from Finns/Estonians/Sami

  8. The Following 2 Users Say Thank You to xenus For This Useful Post:

     anglesqueville (05-09-2019),  JMcB (05-10-2019)

  9. #95
    Registered Users
    Posts
    70
    Sex

    @anglesqueville I haven't vetted the paper myself but here is a paper on poles of diversity in NE Europe. If it holds any water at all it might explain the substantial pull you get but i really have no idea about your family or the actual relevance of the paper.

    https://journals.plos.org/plosone/ar...l.pone.0058552

  10. The Following 2 Users Say Thank You to xenus For This Useful Post:

     anglesqueville (05-10-2019),  JMcB (05-10-2019)

  11. #96
    Registered Users
    Posts
    849
    Sex
    Location
    Sweden
    Ethnicity
    Germanic + Italo-Celt
    Nationality
    Swedish
    Y-DNA
    R-L2 / R1b-U152
    mtDNA
    H1a1

    Sweden Italy Italy 1861-1946 Sami Vatican
    Quote Originally Posted by Nino90 View Post
     
    ,PC1,PC2,PC3,PC4,PC5,PC6,PC7,PC8,PC9,PC10,PC11,PC1 2,PC13,PC14,PC15,PC16,PC17,PC18,PC19,PC20,PC21,PC2 2,PC23,PC24,PC25
    Nino_scaled,0.126344,0.119832,0.063356,0.036822,0. 032621,0.015618,0.004935,0.007615,0.011044,0.00400 9,-0.002111,0.009741,-0.010406,-0.016377,0.010043,0.01538,0.009909,-0.002534,0.008045,0.012756,0.009733,-0.00272,0.006655,0.006145,0.000239

    ,PC1,PC2,PC3,PC4,PC5,PC6,PC7,PC8,PC9,PC10,PC11,PC1 2,PC13,PC14,PC15,PC16,PC17,PC18,PC19,PC20,PC21,PC2 2,PC23,PC24,PC25
    Nino,0.0111,0.0118,0.0168,0.0114,0.0106,0.0056,0.0 021,0.0033,0.0054,0.0022,-0.0013,0.0065,-0.007,-0.0119,0.0074,0.0116,0.0076,-0.002,0.0064,0.0102,0.0078,-0.0022,0.0054,0.0051,0.0002


    If you got time. I would appriciate help.
    Thanks.
    Please give it a try.

  12. The Following User Says Thank You to Nino90 For This Useful Post:

     JMcB (05-10-2019)

  13. #97
    Registered Users
    Posts
    785
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by Nino90 View Post
    Please give it a try.
    I don't understand. I posted your result in #76 and you thanked me for it.

  14. #98
    Registered Users
    Posts
    849
    Sex
    Location
    Sweden
    Ethnicity
    Germanic + Italo-Celt
    Nationality
    Swedish
    Y-DNA
    R-L2 / R1b-U152
    mtDNA
    H1a1

    Sweden Italy Italy 1861-1946 Sami Vatican
    Quote Originally Posted by Huijbregts View Post
    I don't understand. I posted your result in #76 and you thanked me for it.
    My bad. I missed it.
    Thanks.

  15. #99
    Registered Users
    Posts
    1,274
    Sex
    Location
    Viseu
    Ethnicity
    Romanised Paesuri
    Nationality
    Portuguese
    Y-DNA
    E-BY36858
    mtDNA
    H20

    Asturias Portugal 1143 Portugal 1485 Portugal Order of Christ
    Huijbregts, would you mind running Scotland_LBA:I3130 and Halberstadt_LBA:I0099, please?

    Over the French section we've been toying with a few models with these two north European LBA samples, since some people's models appear to chose one over the other, despite in theory they not being too dissimilar, but plot in different areas on the PCA plots. For example the Hallstatt DA111 overwhelmingly prefers the Scottish sample, I do too, and as expected we do get relatively short distances from each other. We're wondering if there could be some ancient bronze age Atlantic connection at play.
    YDNA - E-Y31991>PF4428>BY36857. Domingos Rodrigues, b. circa 1680 Hidden Content , Viana do Castelo, Portugal
    mtDNA - H20. Maria Josefa de Almeida, b. circa 1750 Hidden Content , Porto, Portugal

    Global25 PCA West Eurasia dataset Hidden Content
    Hidden Content


    [1] "distance%=1.7726"

    Ruderico

    Celtiberian,77.6
    Roman_Imperial_proxy,13.6
    Guanche,8.8

  16. The Following User Says Thank You to Ruderico For This Useful Post:

     JonikW (05-17-2019)

Page 10 of 10 FirstFirst ... 8910

Similar Threads

  1. A clustering of the Europeans from Global 25
    By Huijbregts in forum Autosomal (auDNA)
    Replies: 148
    Last Post: 01-29-2019, 04:07 PM
  2. NorthWest Admixures
    By tipirneni in forum Hindi
    Replies: 37
    Last Post: 01-03-2019, 02:32 AM
  3. Global 25 Ancient Britain & Irish vs Present day NW Europeans
    By MitchellSince1893 in forum Autosomal (auDNA)
    Replies: 2
    Last Post: 12-29-2018, 03:19 AM
  4. Replies: 0
    Last Post: 08-06-2012, 12:15 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •