Page 13 of 13 FirstFirst ... 3111213
Results 121 to 126 of 126

Thread: A clustering of the Europeans from Global 25

  1. #121
    Registered Users
    Posts
    624
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by Molfish View Post
    Would you mind doing mine and my parents whenever you have time?

    ,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
    Molfish,0.0113,0.0136,0.018,0.0157,0.0127,0.008,0. 0021,0.0033,0.0017,0.0011,-0.0073,0.0057,-0.0106,-0.0096,0.0145,-0.0009,-0.0155,0.0035,0.0011,-0.0039,0.004,0.0044,-0.0008,0.0099,-0.003

    ,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
    MolfishFather,0.0114,0.0137,0.0176,0.018,0.0129,0. 0079,0.0016,0.004,-0.0002,0.0032,-0.002,0.0006,-0.0088,-0.0043,0.0193,-0.004,-0.0173,0.0028,0.0003,-0.0051,0.0035,0.002,-0.0003,0.008,-0.003

    ,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
    MolfishMother,0.0117,0.0133,0.0163,0.0144,0.0123,0 .0077,0.0015,0.0033,0.0023,-0.0015,-0.0043,0.0058,-0.0084,-0.0119,0.0106,0.0076,0.0013,0.0015,0.0044,0.0006,0 .0047,0.0065,-0.0017,0.0077,-0.0002

    Thanks.
    Molfish.jpg

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

     JMcB (12-03-2018),  Molfish (12-03-2018)

  3. #122
    Registered Users
    Posts
    266
    Sex
    Location
    Australia
    Ethnicity
    British Islander
    Y-DNA
    R1b-M222 > DF109
    mtDNA
    K1a10

    Australia England Ireland
    Quote Originally Posted by Huijbregts View Post
    Consistent with other maps, my father overlapping with I presume Icelandics, my mother with Dutch, and me deeply British/Irish. Goes to show you can't read too much into personal variation within a small cluster.
    Last edited by Molfish; 12-03-2018 at 01:15 AM.
    LivingDNA: 97.6% British Isles, 2.4% Scandinavia. 45.1% Central England related.
    AncestryDNA: 53% England/Wales/NW Euro, 47% Ireland/Scotland

  4. The Following 2 Users Say Thank You to Molfish For This Useful Post:

     Finn (12-03-2018),  JMcB (12-03-2018)

  5. #123
    Registered Users
    Posts
    429
    Sex
    Omitted
    Location
    Cardiff
    Ethnicity
    NW European
    Nationality
    Welsh
    Y-DNA
    R1b-U106 (Z306/S497)
    mtDNA
    H13a1a

    England Ireland Wales European Union
    Quote Originally Posted by Huijbregts View Post
    cluster 1
    Cluster 1 contains 64 of the 68 French samples, and 10/10 of the French_East samples, but not the French_South and Basque samples.
    Also 26/27 of the Belgian samples are in #1.
    59/87 German are in cluster1.
    Cluster 1 also has 17/22 of the Swedish and 1/1 Nordic IA.
    This cluster is the non-British NW cluster.
    I spent a long time building a duplicate of a calculator from the ground up using unscaled data. I reached the end of the process (lots of LDA in Past3) and created a new data file for use with nMonte3. I then ran some tests and the output was unusual - not what would be expected compared to using scaled data. So ultimately I shelved the calculator.

    Is it possible that this effect of using unscaled data has occured here, with an unusual clustering of French and Swedes? I always keep in mind "genes mirror geography" to guide my output but maybe I'm wrong.
    My ancestry: 53% S Eng, 2% N Eng, 39% Ire, 6% S Wal
    LivingDNA: 53% S Eng, 7% E Anglia, 15% SW Scot, 9% S Wal, 8% Brit Isles, 4.5% Basque/Sardinian/Tuscan, 3.4% Chechen/Indian/Kurdish

  6. The Following User Says Thank You to Capitalis For This Useful Post:

     Ruderico (12-10-2018)

  7. #124
    Registered Users
    Posts
    64
    Sex
    Nationality
    US

    France England Germany Palatinate Switzerland
    Could you plot mine? Thanks! ,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
    DavidC_scaled,0.132035,0.137096,0.049403,0.031654, 0.037853,0.013945,0.005405,0.003461,0.008385,0.012 028,-0.001786,0.007943,-0.011596,-0.007707,0.009093,0.004906,-0.001173,-0.004307,0.005154,0.006753,0.004243,-0.002968,0.002835,0.006989,0.007784

    ,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
    DavidC,0.0116,0.0135,0.0131,0.0098,0.0123,0.005,0. 0023,0.0015,0.0041,0.0066,-0.0011,0.0053,-0.0078,-0.0056,0.0067,0.0037,-0.0009,-0.0034,0.0041,0.0054,0.0034,-0.0024,0.0023,0.0058,0.0065

  8. The Following User Says Thank You to dchicn For This Useful Post:

     Ohio (12-04-2018)

  9. #125
    Registered Users
    Posts
    624
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by Capitalis View Post
    I spent a long time building a duplicate of a calculator from the ground up using unscaled data. I reached the end of the process (lots of LDA in Past3) and created a new data file for use with nMonte3. I then ran some tests and the output was unusual - not what would be expected compared to using scaled data. So ultimately I shelved the calculator.

    Is it possible that this effect of using unscaled data has occured here, with an unusual clustering of French and Swedes? I always keep in mind "genes mirror geography" to guide my output but maybe I'm wrong.
    The unscaled data are the PCA scores from Global 25, so I do not understand what you mean with 'unscaled effect'.
    But maybe there is something like a 'Past3 effect'
    Arithmetically the results of Past3 are correct. But they may be useless for drawing conclusions.
    For instance, when you offer Past3 a sample of two Chinese, it will return the mathematically correct result. But you cannot generalize that result to the average length of the Chinese population. Past3 doesn't warn you that the result is not generalizable.
    Davidski has cleaned the Global 25 from outliers and family relationship. But there are several methodological problems left: in the first place overfitting ("lots of LDA"). Also the dataset is very unbalanced. And lastly, many populations are overlapping.

    For this thread I did not use the population labels in Global 25 because there are so many of them that overfitting is unavoidable; also many of the (sub)populations are known to be heavily overlapping.
    I just used the genetic coordinates for a clustering into 9 clusters. I am aware that most of these clusters are overlapping.
    Also it is evident that there is a severe unbalance between the populations. I presume that if I could have used more Norwegian and Swedish samples, the Swedish would have split from the French, and the Norwegians from the British.

    Btw, if you always keep in mind the "genes mirror geography" rule, why do you want to build a calculator?
    Last edited by Huijbregts; 12-05-2018 at 01:10 PM.

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

     anglesqueville (12-05-2018),  Ruderico (12-10-2018)

  11. #126
    Registered Users
    Posts
    624
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by dchicn View Post
    Could you plot mine? Thanks! ,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
    DavidC_scaled,0.132035,0.137096,0.049403,0.031654, 0.037853,0.013945,0.005405,0.003461,0.008385,0.012 028,-0.001786,0.007943,-0.011596,-0.007707,0.009093,0.004906,-0.001173,-0.004307,0.005154,0.006753,0.004243,-0.002968,0.002835,0.006989,0.007784

    ,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
    DavidC,0.0116,0.0135,0.0131,0.0098,0.0123,0.005,0. 0023,0.0015,0.0041,0.0066,-0.0011,0.0053,-0.0078,-0.0056,0.0067,0.0037,-0.0009,-0.0034,0.0041,0.0054,0.0034,-0.0024,0.0023,0.0058,0.0065
    dchicn.jpg

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

     dchicn (12-05-2018)

Page 13 of 13 FirstFirst ... 3111213

Similar Threads

  1. Global 25 from the Web.
    By Mixed in forum Autosomal (auDNA)
    Replies: 1
    Last Post: 08-26-2018, 05:22 PM
  2. Question about global 25??
    By Samuel7312 in forum General
    Replies: 0
    Last Post: 08-13-2018, 05:42 PM
  3. Global 25 con't
    By sktibo in forum Autosomal (auDNA)
    Replies: 308
    Last Post: 08-08-2018, 08:18 AM
  4. Replies: 2
    Last Post: 05-30-2018, 01:24 AM
  5. Replies: 3
    Last Post: 02-18-2017, 09:05 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
  •