Totalt antall sidevisninger

onsdag 18. september 2013

Saami ancestry and the MDLP Oracle-x Population Fitting

Update 08.10.2014: The MDLP calculators have been updated and now do not give the results as shown below. This post is then considered outdated and should not be used as guide vs the MDLP calculators.

This is a short test of MDLP Oracle-x Population Fitting at Gedmatch ability to catch Saami ancestry. Only MDLP calculators have been tested as these are the only one with Saami population reference.

As the test shows finding Saami ancestry using these calculators may give very different and even erroneous results (no Saami ancestry or minor Saami ancestry when there is actually major Saami ancestry) but one stand clearly out as the preferred choice. The "test subject" (with consent) is a North-Saami individual participating in the project with mostly Saami ancestry.


Absent = Saami ancestry not detected.
Minor = Saami ancestry detected but as minority ancestry.
Top but minor = Saami ancestry detected as top population but with less than 50%.
Top majority = Saami ancestry detected as top population with more than 50%.



As the result shows using MDLP K=5 Oracle X "Pct. Calc. Option 1" appears to be the preferred choice to detect Saami ancestry.

EDIT 20/9-13

Please note that all the other functionalities of the different versions of the MDLP calculators like Oracle and Oracle-4 was neither able to find that the Saami individual had majority Saami ancestry. Some where able to infer Saami minority ancestry and some didnt detect anything at all. The exception again is the MDLP K=5 that managed in Oracle to get a Saami population as number two and 3 of 4 in Oracle-4. Thie MDLP 27 calculator (not in Gedmatch) managed using Gaussian method 1 population mode to infer the correct population but failed in the 2,3 and 4 approximations.


27 kommentarer:

  1. These are the top 4 populations from my results (the rest being <1%):
    1. FIN 76.22%
    2. SAM 9.51%
    3. SWD 9.22%
    4. BASQ 4.99%
    Is 'SAM' Saami? How should I read this? By the way, I'm a western Finn (C. Ostrobothnia), which explains the Swedish, right?

    SvarSlett
  2. Yes SAM is Saami. Your western Finn origin may explain the Swedish minority ancestry. Why you have a Basque score I must admit I dont know, maybe its below the margin of error.

    How to read this would be first to check your geneology (any Saamis? Ancestors from Saami areas?) and second to what extent this also appear in other Finns from different parts of Finland. Maybe its background relatedness seen in many Finns.

    SvarSlett
  3. Very nice blog! I have just recently discovered it. My results from the MDLP K=8 Oracle-X Population Fitting/Pct. Calc. Option 1 are as follows. As you can see, there is a small SAM percentage (1.18%). Do you think that this is just noise? And I am wondering how you got those comparison results shown above (Absent/Minor, etc.) I am relatively new to Gedmatch.com, but have mainly only seen my results in pie-chart form.

    SvarSlett
  4. Liz: This short tests shows that the MDLP K=5 Oracle X Pct. Calc. Option 1 appears to describe this Saami individual ancestry best. K=8 appears to be able to detect it but only as minority ancestry. It is possible that some small Saami percentages may show up for random individuals for these MDLP calculators but it is quite possibly not real if it doesn't show up consistently.

    SvarSlett
    Svar
    1. Thank you, Anders, for the information. I show low-level (1.18% and below; usually about 1.10%) for MDLP's K8 through K12 Oracle-X option calculators.

      Slett
    2. Liz: difficult to say if it is real or not. I will probably guess not as it is so small but you never know. If you have partly ancesty from known Saami populated areas or even better known geneologies connecting to Saami individuals it would be even better.

      Slett
    3. Thanks, Anders. Some of the Gedmatch.com admixture calculators (various ones) returned results for Finnish, Swedish and Norwegian ancestry ranging from 8%-ish to maybe twice that percentage, and some quite a bit lower (AncestryDNA also gave me "Finnish/Volga-Ural" results - about 12%). Anyway, as you suggest, without known genealogies, it's hard to say!

      Slett
    4. This is interesting to me because my ancestry is mostly British, except that I have two great-grandparents who were from areas of Sweden where there are a lot of Saami. K=5 Oracle X Pct. Calc. Option 1 gives me no Saami at all (is there a legend somewhere that explains what all the abbreviations are?)

      0 Unable to determine 0.26%
      1 CEU 86.34%
      2 EUKR 5.49%
      3 NOG 4.40%
      4 GRK 1.64%
      5 BLG 1.34%
      6 BSHK 0.38%
      7 KSV 0.12%
      8 LZG 0.02%
      9 GYP 0.01%
      10 TRK 0.00%

      Whereas most of the other MDLP utilities give me results with Saami ranging from 1-3% (this is K=8 Oracle X Pct. Calc. Option 2):

      0 Unable to determine 0.01%
      1 GER 62.69%
      2 BASQ 8.17%
      3 SRD 7.11%
      4 GBRARG 7.10%
      5 LTV 4.16%
      6 SAM 3.26%
      7 BLG 2.96%
      8 UD 2.36%
      9 TTR 2.17%
      10 CPR 0.00%

      Do you think this, along with what appear to be central Asiatic clusters, would be consistent with a Swedish grandparent having some Saami blood or is it just noise?

      Slett
    5. Angie: That I dont know but I would guess that Saami ancestry should be seen directly as Saami population is in the reference so no need for Central Asian proxies. You may want to contact the authors of these calculators about this. The contact details should be available when you run the calculators.

      Slett
  5. My K=4
    1 KRL 89.58%
    2 BASQ 8.92%
    3 HKS 1.43%
    4 SRD 0.02%
    5 FRN 0.02%
    6 GBRCORN 0.01%
    7 IBR 0.01%
    8 GBRORK 0.01%
    What is HKS population?
    My son:
    1 KRL 56.56%
    2 SWD 17.27%
    3 GBRARG 8.32%
    4 FIN 7.32%
    5 SAM 7.26%
    6 BASQ 3.23%
    7 NRW 0.03%
    8 CEU 0.01%
    What is GBRARG population?

    SvarSlett
  6. HKS = Hakas. GBRARG = Great Britain, Argyll.

    SvarSlett
  7. Besides the 9.51% that I got with K=5 Opt2, my other results for Saami were: K=12 Opt1 4.14% Opt2 6.33%, World-22 Opt2 2.04%. The rest returned no Saami. I have matched with some Saami individuals but I thought that it was probably due to my ancestors moving to Saami areas in Sweden and Finland.

    In any case, the MDLP World test gave me weird results. It gave me 85-95% Russian and the top two and three were Koryak and Hanti. Speaking of Russian, I probably have a Russian ancestor but it isn't confirmed yet.

    Sometimes, I wish that people weren't so worried about privacy... :P The amount of declines on 23andMe really sucks, too.

    By the way, are Paleo-Scandinavian and North-European-Mesolithic both Saami components?

    SvarSlett
  8. I dont know. I do not have the MDLP K=5 oracle files. The answer would be in there.

    SvarSlett
  9. MDLP K=5 Pct. Calc. Option 1
    1 Karelian 34.66%
    2 FIN 28.48%
    3 SWD 21.61%
    4 SAM 8.99%
    5 BASQ 4.52%
    6 GBRARG 0.93%
    7 Veps 0.81%


    MDLP World test Pct. Calc. Option 1 also had Saami and something else:
    1 Finnish 47.62%
    2 Swedish_V 32.34%
    3 Ire8 [the Gotlander] 14.09%
    4 Saami 4.76%
    5 Yukagir 1.19%

    SvarSlett
  10. I read this article with interest and ran my 23andMe results through all of the MDLP with Oracle-x calculators and interestingly K=5 was the only one showing no Saami. The others show some Saami mostly under option 2, with K=6 and K=7 the strongest.

    K=5, opt 1 - 0.00, opt 2 - 0.00
    World 22, opt 1 - 2.98, opt 2 - 2.09
    K=6, opt 1 - 0.00, opt 2 - 5.38
    K=7, opt 1 - 0.04, opt 2 - 5.59
    K=8, opt 1 - 0.00, opt 2 - 2.84
    K=9, opt 1 - 0.00, opt 2 - 2.92
    K=10, opt 1 - 0.00, opt 2 - 2.89
    K=11, opt 1 - 0.00, opt 2 - 2.89
    K=12, opt 1 - 0.99, opt 2 - 1.08

    My known background; paternal Grandparents were both born in Southern Sweden and my maternal gg grandfather was born in Norway, the rest of my mothers family is a mix of German, Colonial British and French.

    SvarSlett
    Svar
    1. As K=5 was the only calculator that was able to disover that this Saami individual was of majority Saami ancestry I am assuming it is the most correct calculator of all these. Anything as low as these scores you show I then would assume is far below the margin of error and as you score 0 at K=5 it seems you have no Saami ancestry.

      Slett
  11. Thanks Anders. The reason I was searching for Saami ancestry was That I thought the Yakut affinity that I show might have been East Asian that admixed with my Scandinavian ancestors. I've always shown East Asian on the same spot on Chromosome 2 now 23andMe shows it as 0.02% Yakut on their new Ancestry Composition. I'll have to keep searching to solve that.

    SvarSlett
    Svar
    1. It could theoretically be that that you have a Saami ancestor far back and the little this ancestor contributed to you happend to be a segment that fits the Yakut even the far majority of any Saami ancestors is European. However there is no guarantee for this is the case. It may have came other ways too but its difficult without a proper investigation of the segment in question and even then there might not be a sure answer.

      Slett
  12. Sorry, I meant to say 0.2% Yakut.

    SvarSlett
  13. All four of my mother's Grandparents are Finnish and only two of my Grandparents are Finnish. Yet my results are nothing like hers.

    My mom
    1 FIN 64.68%
    2 SAM 21.35%
    3 SWD 9.21%
    4 BASQ 4.70%
    5 GBRARG 0.04%
    6 GBRWAL 0.02%

    Me:
    1 SWD 69.63%
    2 CHU 22.98%
    3 EUKR 5.30%
    4 BLG 1.21%
    5 TTR 0.48%
    6 MCD 0.35%
    7 NOG 0.03%
    8 LZG 0.02%

    SvarSlett
    Svar
    1. It could be that because your "mixed" it uses the closest proxies. I suggest you use Oracle-X with 4 pops to maybe get a answer closer to the truth.

      Slett
  14. Hello Anders:

    I ran my 23andme results through the MDLP K=5 Oracle-x Pct. Calc. Option 1, looking specifically for Saami ancestry, as my maternal grandparents were from Sodankyla, Finland and raised reindeer. My results were:

    1 KRL 77.07%
    2 BASQ 15.13%
    3 SAM 6.40%
    4 GYP 1.11%
    5 NRW 0.03%

    Everything else was 0.01 or less; "unable to determine" was 0.23%.

    The thing that has me confused is the Basque ... all 16 of my great-great grandparents were born in Finland, and 5 of my 8 great-grandparents were born in Finland (the other 5 were born in Michigan after their parents emigrated). Any idea where the Basque might be coming from or why it is showing? Thank you!

    SvarSlett
    Svar
    1. I have not tried to experiment what score people with Saami minority ancestry would get in these tests. But my guess from this experiemnt is that getting some higher score here would be better but as you can see this test also claim you to be 15% Basque so it also make your Saami score questionable.

      Slett
    2. n this PCA dimension plotted on a gradient map of Europe appear to show some connection between Finns and Basque, maybe the genetic traces of an ancient agricultural expansion that have moved from South-East Europe tot he west and north to the Finns.

      https://sites.google.com/site/fennobga/CCEuropeD5map191213.png

      Slett
  15. Hello, maybe this can give an answer to your question. Most Basque are probably the same stock as the farmers that came from Balkans, and Middle East, it seems that the Basque language show such affiliation.
    But this work is about Sámi and Berbers and about the Ice Age refugee in Franco-Kantabria and Iberia/North Africa.

    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1199377/

    SvarSlett
    Svar
    1. In this PCA dimension plotted on a gradient map of Europe appear to show some connection between Finns and Basque, maybe the genetic traces of an ancient agricultural expansion that have moved from South-East Europe tot he west and north to the Finns.

      https://sites.google.com/site/fennobga/CCEuropeD5map191213.png

      Slett