Page 3 of 22 FirstFirst 1234513 ... LastLast
Results 21 to 30 of 220

Thread: Post your GenePlaza Results!

  1. #21
    Banned
    Posts
    5,173
    Sex

    Quote Originally Posted by Magnetic View Post
    the central asian component is flawed . there is no way a Kurd can score that much "Altaian, Kalmyk, Tuvinian and Yakut in Russia"

    dafuq

    also so much Arab ? hmm seems odd to me

    waiting for my results (but I am not a fan of such sites . they dont seem that accurate )


    The E Asian heavy Central Asian score may seem high to you, but that is because you are used to seeing results from ADMIXTURE based calculators that use a RELATIVE E Asian component, and not an ABSOLUTE one (a little more on this later in the post).

    Their website says that they use an algorithm developed by JOE PICKERELL. Well, J Pickerell is very much respected in the circles of genetics research, including by the folks at Harvard. I remember a paper he published with DAVID REICH, "Toward a new history and geography of human genes informed by ancient DNA". All you have to do is google his name.

    J PICKERELL has been an Investigator at New York Genome Center, an 186 employee facility for the past 3 years.
    He also was a Post Doctoral Fellow for 2yrs at Harvard Medical School, and got his PhD in Human Genetics from the Medical School at The University of Chicago in 2011.

    The GPS Origins results also showed a score of about 18% for Siberian & Tuvanian for the northern Iraqi Kurd, and my own test using rarer alleles showed a similar score, when I used non-E Asian admixed ancients for the other components:


    COMPONENT Kurd-Iraq-N MAF<12% - 26KSNPs Armenian - MAF<12% - 26KSNPs Assyrian - MAF<12% - 26KSNPs
    FARMER 59.15% 65.92% 77.25%
    Anatolia-N 16.76% 18.05% 19.21%
    Europe-EN 17.10% 20.28% 23.04%
    Levant N/BA 7.92% 10.14% 9.11%
    Iran N/Chl 17.37% 17.44% 25.89%
    STEPPE 22.82% 19.79% 17.02%
    Andronovo-Srubnaya 11.73% 11.19% 6.90%
    Yamnaya-Poltavka 11.09% 8.61% 10.12%
    WHG 1.50% 3.61% 2.42%
    SE Asian 7.95% 1.96% 0.00%
    NE Asian 5.19% 2.25% 0.00%
    SSA 2.63% 4.28% 0.00%
    Papuan 0.76% 2.18% 3.30%



    In other words, if the calculator has a Caucasian component and a Kurd scores 50% Caucasian, their E Asian score would be diminished because those alleles are part of the Caucasian variation.

    ADMIXTURE calculator, such as those on Gedmatch that uses 1000+ test samples in the run, compared to much fewer population sources. These artificially minimize variation within populations, and the component allele frequencies are skewed by the test subjects. I have addressed this many times both here and on my website. This problem has also been addressed in papers such as


    http://bmcbioinformatics.biomedcentr...859-014-0418-7
    For comparison, we also ran ADMIXTURE (in supervised mode using the HapMap reference panel of individuals) on the same dataset (see Figure 1(b)). The European and African admixture estimates for each individual were highly consistent between the two methods. For some individuals, the European component of ancestry using our method was split between the TSI and CEU populations. This could reflect one important difference between the two methods in how they use data from reference individuals. Our method finds a maximum likelihood estimate of the admixture coefficients for each individual using the fixed set of allele frequencies. In contrast, ADMIXTURE, in the supervised mode, utilizes data for all individuals (both the reference populations and the individual(s) being analyzed) to estimate the allele frequencies for each cluster or population and maximize the likelihood function summed across all individuals. Therefore, the allele frequencies are determined not only by the genotypes of the reference individuals but also by the individual(s) that are analyzed for admixture. To confirm this, we estimated allele frequencies by running ADMIXTURE twice: (1) using 800 reference individuals simulated using allele frequencies for 8 HapMap populations (100 individuals per population, see previous section) and (2) 800 reference individuals and 1 additional individual with 100% CEU ancestry simulated using the HapMap allele frequencies. Subsequently, we used our method to estimate admixture coefficients for the simulated CEU individual using the two sets of allele frequencies separately. We found that using the first set of allele frequencies, the admixture coefficients for both CEU and TSI were non-zero. In contrast, using the second set of allele frequencies, only the CEU admixture coefficient was non-zero. This was similar to the results observed in the analysis of the Mozabite data and provided an empirical validation of our hypothesis regarding the difference in the admixture coefficients estimated by the two methods.


    AncestryDNA also recognizes this problem:
    Here is an excerpt from page 24 of their white paper at:

    https://www.ancestry.com/dna/resourc...ty-White-Paper
    Another assumption that follows from the likelihood equation above (Equation 5) is that all samples are independent, or unrelated. In order to meet this assumption, we preprocess the genetic samples to place any samples from related individuals into separate runs of ADMIXTURE. In a particular run, we also remove any reference samples to whom a customer appears to be related.

    It should also be noted that the approach we use is not entirely “supervised,” although we use a supervised version of the algorithm. While the reference populations are set as the “source” populations, genotypes of the tested samples can also influence the allele frequency estimates in the source clusters; i.e., the approach is not fully supervised.

    This is because the model not only estimates Q, but also P, as a function of both the reference samples and the customer samples (a total of N samples). While ideally the P values should remain stable regardless of the customer samples, the customer samples could slightly change the P estimates from their “true” values.

    Customer samples are run in batches of varying sizes; due to the details of the algorithm described above, in theory a customer’s results could vary by batch.
    Extensive tests have shown that the effect of batch on customer estimates is minimal. This is because the batch size is very small compared to the size of the reference panel. Also, removing related samples from the same batch, as described above, ensures minimal effects on customer ethnicity estimates.
    Have you looked at some "East Asian" looking Kurds, W Asians, and S Asians and sometimes wondered how come they score only say 2% E Asian on some of the genetics tests, and thought that did not make sense? Well, the reason is because these tests underestimate the E Asian admixture. There is in fact some correlation between genotype and phenotype.

    I will try to get some details on their algorithm and methodology, but they may not be so open about it.
    Last edited by Kurd; 05-21-2017 at 10:13 AM.

  2. The Following 9 Users Say Thank You to Kurd For This Useful Post:

     Amerijoe (05-21-2017),  Helgenes50 (08-02-2017),  jjensen6 (09-02-2017),  jortita (05-21-2017),  kangz (05-21-2017),  kingjohn (05-21-2017),  Reza (05-21-2017),  vettor (05-21-2017),  wombatofthenorth (08-02-2017)

  3. #22
    Gold Member Class
    Posts
    2,765
    Sex
    Location
    Calne,England
    Ethnicity
    British and Irish
    Nationality
    Great Britain
    Y-DNA
    E-Y45878
    mtDNA
    H67

    United Kingdom Scotland England Ireland
    It says my results are being calculated but it seems to be taking longer than the 2-10 minutes stated.
    Please support Mental health research and world community grid

    Hidden Content
    Hidden Content
    Hidden Content
    Hidden Content

  4. #23
    Gold Member Class
    Posts
    2,765
    Sex
    Location
    Calne,England
    Ethnicity
    British and Irish
    Nationality
    Great Britain
    Y-DNA
    E-Y45878
    mtDNA
    H67

    United Kingdom Scotland England Ireland
    Dna land 23andMe upload and Geneplaza upload


    DNA land 23andme 4th.PNG


    gene plaza ancestry.PNG

    The Finnish seems to have been swapped for North Slavic. I am not aware of having either.
    Attached Images Attached Images
    Last edited by firemonkey; 05-21-2017 at 01:32 PM.
    Please support Mental health research and world community grid

    Hidden Content
    Hidden Content
    Hidden Content
    Hidden Content

  5. The Following 3 Users Say Thank You to firemonkey For This Useful Post:

     Asimakidis (08-02-2017),  sgdavies@hotmail.com (05-21-2017),  wombatofthenorth (08-02-2017)

  6. #24
    Registered Users
    Posts
    1,534
    Sex
    Ethnicity
    Assamese
    Nationality
    Indian
    Y-DNA
    R1a1
    mtDNA
    M13'46'61

    India China Tajikistan Mongolia Russian Federation Papua New Guinea
    Quote Originally Posted by Kurd View Post
    The E Asian heavy Central Asian score may seem high to you, but that is because you are used to seeing results from ADMIXTURE based calculators that use a RELATIVE E Asian component, and not an ABSOLUTE one (a little more on this later in the post).

    Their website says that they use an algorithm developed by JOE PICKERELL. Well, J Pickerell is very much respected in the circles of genetics research, including by the folks at Harvard. I remember a paper he published with DAVID REICH, "Toward a new history and geography of human genes informed by ancient DNA". All you have to do is google his name.

    J PICKERELL has been an Investigator at New York Genome Center, an 186 employee facility for the past 3 years.
    He also was a Post Doctoral Fellow for 2yrs at Harvard Medical School, and got his PhD in Human Genetics from the Medical School at The University of Chicago in 2011.

    The GPS Origins results also showed a score of about 18% for Siberian & Tuvanian for the northern Iraqi Kurd, and my own test using rarer alleles showed a similar score, when I used non-E Asian admixed ancients for the other components:


    COMPONENT Kurd-Iraq-N MAF<12% - 26KSNPs Armenian - MAF<12% - 26KSNPs Assyrian - MAF<12% - 26KSNPs
    FARMER 59.15% 65.92% 77.25%
    Anatolia-N 16.76% 18.05% 19.21%
    Europe-EN 17.10% 20.28% 23.04%
    Levant N/BA 7.92% 10.14% 9.11%
    Iran N/Chl 17.37% 17.44% 25.89%
    STEPPE 22.82% 19.79% 17.02%
    Andronovo-Srubnaya 11.73% 11.19% 6.90%
    Yamnaya-Poltavka 11.09% 8.61% 10.12%
    WHG 1.50% 3.61% 2.42%
    SE Asian 7.95% 1.96% 0.00%
    NE Asian 5.19% 2.25% 0.00%
    SSA 2.63% 4.28% 0.00%
    Papuan 0.76% 2.18% 3.30%



    In other words, if the calculator has a Caucasian component and a Kurd scores 50% Caucasian, their E Asian score would be diminished because those alleles are part of the Caucasian variation.

    ADMIXTURE calculator, such as those on Gedmatch that uses 1000+ test samples in the run, compared to much fewer population sources. These artificially minimize variation within populations, and the component allele frequencies are skewed by the test subjects. I have addressed this many times both here and on my website. This problem has also been addressed in papers such as


    http://bmcbioinformatics.biomedcentr...859-014-0418-7




    AncestryDNA also recognizes this problem:
    Here is an excerpt from page 24 of their white paper at:

    https://www.ancestry.com/dna/resourc...ty-White-Paper


    Have you looked at some "East Asian" looking Kurds, W Asians, and S Asians and sometimes wondered how come they score only say 2% E Asian on some of the genetics tests, and thought that did not make sense? Well, the reason is because these tests underestimate the E Asian admixture. There is in fact some correlation between genotype and phenotype.

    I will try to get some details on their algorithm and methodology, but they may not be so open about it.
    Kurd, as I do not have 23andme data and cannot upload. Given what you have said, would you say its worth it to order their kit, thank you

  7. #25
    Junior Member
    Posts
    9
    Sex
    Ethnicity
    Albanian
    Y-DNA
    J2b2
    mtDNA
    J1c

    Kosovo Albania Germany

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

     CelticGerman (05-21-2017),  khanabadoshi (05-22-2017),  Kurd (05-21-2017),  vettor (05-21-2017)

  9. #26
    Registered Users
    Posts
    615
    Sex
    Location
    Germany
    Ethnicity
    European
    Nationality
    German
    Y-DNA
    +R1b/U152/Z36/CTS188
    mtDNA
    H

    Germany Denmark Czech Republic Switzerland Austria European Union
    My GenePlaza result in comparison with DNA Land Attachment 16210
    Y-DNA: R1b/U152/Z36/CTS4333/CTS7958, from Thuringia 1634, probably Alsace 1552, -- mt-DNA: H
    EUROGENES K13: North German 3.7, North Dutch 4.85, Danish 5.13, Swedish 5.44, Norwegian 5.7 -- West Eurasia K8: WHG 46.64%, ENF 39.09%, ANE 14.23%
    23andMe: speculative 26% French&German, 11.3% Scandinavian, 9.1% Eastern European, 8.8% British&Irish, 1.6% Southern European, 38% Broadly North Euro
    Known ancestry: 92.6% German (66.8% North German), 4.7% Danish, 1.8% Czech, 0.8% Austrian, 0.1% Swiss

  10. The Following 4 Users Say Thank You to CelticGerman For This Useful Post:

     Amerijoe (05-21-2017),  kangz (05-21-2017),  khanabadoshi (05-22-2017),  vettor (05-21-2017)

  11. #27
    Banned
    Posts
    5,173
    Sex

    Quote Originally Posted by jortita View Post
    Kurd, as I do not have 23andme data and cannot upload. Given what you have said, would you say its worth it to order their kit, thank you
    Sorry, that is a personal decision you have to make. I am not familiar with the details of the test. I have sent a message, but doubt that they will divulge details.

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

     jortita (05-22-2017)

  13. #28
    Registered Users
    Posts
    2,234
    Sex
    Location
    Canada
    Ethnicity
    Mixed Euro/Near East
    Nationality
    Canadian
    Y-DNA
    R1a-YP4516/YP4807*
    mtDNA
    H11a2a3

    Canada Franco-Manitoban European Union Ottoman Empire Russia Imperial United States Grand Union
    Quote Originally Posted by lilac9 View Post
    Eesh not very good! How is it that some of you got a separate geneplaza result from the DNA.land app result?
    Maybe Joe Pickrell's algorithm uses Monte Carlo modelling, in which case it could be possible to get different results even with separate uploads of exactly the same genome to DNA.Land.
     

    Hidden Content
    Hidden Content
    Hidden Content

    Other ancestral Y lines:

    E1b-M81 Ukraine (Ashkenazi)
    E1b-V13 England
    I1-M253 Ireland
    I2-M423 Ukraine
    R1a-L176.1 Scotland
    R1b-L584 Syria/Turkey (Sephardi)
    R1b-L20 Ireland
    R1b-L21 (1)England; (2)Wales?>Connecticut
    R1b-L48 England
    R1b-P312 Scotland
    R1b-FGC32576 Ireland

    Other ancestral mtDNA lines:

    H1b2a Ukraine (Ashkenazi)
    H6a1a3 Ukraine
    K1a9 Belarus (Ashkenazi)
    K1c2 Ireland
    V7a Ukraine

  14. The Following 3 Users Say Thank You to AJL For This Useful Post:

     kingjohn (08-28-2017),  Titane (05-31-2017),  wombatofthenorth (08-02-2017)

  15. #29
    Banned
    Posts
    4,088
    Sex
    Ethnicity
    karius the black hair
    Nationality
    east med

    intresting it looks like gene plaza has med category

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

     Rukha (05-22-2017)

  17. #30
    Registered Users
    Posts
    164
    Sex

    Quote Originally Posted by AJL View Post
    Maybe Joe Pickrell's algorithm uses Monte Carlo modelling, in which case it could be possible to get different results even with separate uploads of exactly the same genome to DNA.Land.
    I had that exact same problem with DNA.land. So I lost confidence in their results. I uploaded Ancestry.com's raw data twice and I got vastly different results.

  18. The Following User Says Thank You to lilac9 For This Useful Post:

     wombatofthenorth (08-02-2017)

Page 3 of 22 FirstFirst 1234513 ... LastLast

Similar Threads

  1. Post your Eurogenes K36 results
    By Passa in forum Autosomal (auDNA)
    Replies: 187
    Last Post: 09-06-2018, 11:04 PM
  2. Post Eurasia 10 CHG Results
    By Kurd in forum Autosomal (auDNA)
    Replies: 230
    Last Post: 02-25-2018, 10:59 PM
  3. Post ANE K6 Results
    By Kurd in forum Autosomal (auDNA)
    Replies: 364
    Last Post: 08-08-2017, 12:37 AM
  4. Replies: 9
    Last Post: 09-08-2016, 04:00 PM
  5. Irishmen please post ANE K8 results
    By Krefter in forum Autosomal (auDNA)
    Replies: 2
    Last Post: 05-21-2015, 11:44 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
  •