Page 1 of 11 123 ... LastLast
Results 1 to 10 of 102

Thread: FTDNA my origins 2.0 failure......

  1. #1
    Gold Class Member
    Posts
    5,235
    Sex
    Location
    Groningen
    Ethnicity
    SGC+TRB Creool
    Nationality
    NL
    Y-DNA (P)
    E-V22

    Netherlands

    FTDNA my origins 2.0 failure......

    FTDNA recently released a new version of my origins. For much people this meant significant changes in the image of their origins. Certainly for many people form the Low Lands and Northwestern Germany. Confusion everywhere. How is this possible?

    What may well be the case for the Low Lands/ NW Germany/Denmark is that FTDNA his models ignores the "West Germanic" or "North Sea Germanic" autosomal DNA. They identify "British Isles" or "Scandinavia". But they don’t identify the “West-Germanic” factor.

    In that respect is 23 andme closer to the fire. They have a category "Broadly Northwest European" this is autosomal DNA that is not attributable to a specific country in northwestern Europe, but to DNA that occurs in several places along the North Sea coast. And this DNA peaks in the (Northern) Low Lands up to Denmark. But since FTDNA doen’t recognize this category, we see a very distorted picture ....

    In this respect has Maciamo (Eupedia) a more spot on analysis!!!
    "Broadly Northwest European admixture
    This admixture peaks in the northern Dutch provinces of Frisia and Groningen (40%), as well as in East Anglia (35-40%), Denmark (34%), the central Netherlands (32%), Germany (31%) and the northern French département du Nord (31%) and the Cotentin peninsula in Normandy (33%). Its distribution correlates mostly with West Germanic ancestry, but could also include some broader Celto-Germanic elements in Germany, the Benelux and France. It appears to be linked to the Proto-Celto-Germanic Y-haplogroup R1b-U106, which almost reaches its maximum frequency in Frisia, East Anglia and Denmark."


  2. The Following 15 Users Say Thank You to Finn For This Useful Post:

     AnnieD (04-10-2017),  asquecco (04-09-2017),  castle3 (04-09-2017),  Helgenes50 (04-11-2017),  J1 DYS388=13 (04-09-2017),  JohnHowellsTyrfro (04-09-2017),  Mike_G (04-10-2017),  Oleg (Rus) (04-09-2017),  Saetro (04-10-2017),  Shadogowah (04-10-2017),  Simon_W (04-09-2017),  sktibo (04-09-2017),  Solothurn (04-10-2017),  Theconqueror (04-11-2017),  wombatofthenorth (04-11-2017)

  3. #2
    Silver Class Member
    Posts
    4,650
    Sex
    Nationality
    Canadian
    Y-DNA (P)
    R-Z198 (DF27)
    mtDNA (M)
    T2B-T152C

    England Scotland Poland Germany Canada Quebec Netherlands
    Hopefully it's not too far off topic but if I count my broadly northwest European score (23andme of course) as Germanic, it comes pretty close to my paper trail, which is about 18-19% German and 1-2% Dutch. It assigns me 28.5% broadly northwest European, which I think is pretty good for a DNA test, especially considering it might include Germanic elements from places like England. I also note my French and German (8.1%) is within one percent of my paper trail for my French ancestry, so I ask: do you think 23andme's French and German is more representative of French?

    The more I dig into DNA testing the more impressed I am with 23andme. I think you've made a very good analysis of the updated ftdna. It definitely seems to include a Germanic element now in the British category, whereas it didn't before. This would explain why it's more accurate for some and less accurate for others. I observed an increase in British testers British percentages while Irish testers appear to have a little less. I interpret this as evidence of your theory.
    Last edited by sktibo; 04-09-2017 at 11:48 AM.

  4. The Following 5 Users Say Thank You to sktibo For This Useful Post:

     AnnieD (04-10-2017),  castle3 (04-09-2017),  Finn (04-09-2017),  Mike_G (04-10-2017),  Ron from PA (04-10-2017)

  5. #3
    Gold Class Member
    Posts
    5,235
    Sex
    Location
    Groningen
    Ethnicity
    SGC+TRB Creool
    Nationality
    NL
    Y-DNA (P)
    E-V22

    Netherlands
    Quote Originally Posted by sktibo View Post
    Hopefully it's not too far off topic but if I count my broadly northwest European score (23andme of course) as Germanic, it comes pretty close to my paper trail, which is about 18-19% German and 1-2% Dutch. It assigns me 28.5% broadly northwest European, which I think is pretty good for a DNA test, especially considering it might include Germanic elements from places like England. I also note my French and German (8.1%) is within one percent of my paper trail for my French ancestry, so I ask: do you think 23andme's French and German is more representative of French?

    The more I dig into DNA testing the more impressed I am with 23andme. I think you've made a very good analysis of the updated ftdna. It definitely seems to include a Germanic element now in the British category, whereas it didn't before. This would explain why it's more accurate for some and less accurate for others. I observed an increase in British testers British percentages while Irish testers appear to have a little less. I interpret this as evidence of your theory.
    I guess so Sktibo! It's an educated guess that my known ancestors all lived in the Northern Dutch provinces or bordering Germany within max 100 miles.
    These are my new origins results, 0% West European (explained by Simon W. as a signature of La Tene, I guess he is right).


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

     AnnieD (04-10-2017),  sktibo (04-09-2017)

  7. #4
    Registered Users
    Posts
    1,919
    Sex
    Location
    South East Wales UK
    Ethnicity
    Welsh
    Nationality
    British
    Y-DNA (P)
    U106 Z326 R-BY27310
    mtDNA (M)
    J1c1b2a

    United Kingdom Wales
    Quote Originally Posted by sktibo View Post
    Hopefully it's not too far off topic but if I count my broadly northwest European score (23andme of course) as Germanic, it comes pretty close to my paper trail, which is about 18-19% German and 1-2% Dutch. It assigns me 28.5% broadly northwest European, which I think is pretty good for a DNA test, especially considering it might include Germanic elements from places like England. I also note my French and German (8.1%) is within one percent of my paper trail for my French ancestry, so I ask: do you think 23andme's French and German is more representative of French?

    The more I dig into DNA testing the more impressed I am with 23andme. I think you've made a very good analysis of the updated ftdna. It definitely seems to include a Germanic element now in the British category, whereas it didn't before. This would explain why it's more accurate for some and less accurate for others. I observed an increase in British testers British percentages while Irish testers appear to have a little less. I interpret this as evidence of your theory.
    That could account for the unexpected accuracy of my new result - big difference. They have me virtually spot-on as all - UK ancestry, I had a big chunk of Scandinavian the first time. John

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

     Finn (04-09-2017),  sktibo (04-09-2017)

  9. #5
    Registered Users
    Posts
    264
    Sex
    Location
    Oregon, U.S.A.
    Ethnicity
    Irish
    Nationality
    American/Canadian
    Y-DNA (P)
    RL21-DF23>ZP149
    mtDNA (M)
    T2g1

    Northern Ireland Ireland Scotland Canada Quebec
    Just an example of the myOrigins failure. Known Ancestry 75% Isles, 25% French Canadian:

    Old my Origins:
    Isles - 71%
    Scandinavia - 8%
    West & Central Euro - 0% (yes, 0%)

    New myOrigins:
    Isles - 4% (yes! 71% to 4%!!)
    Scandinavia - 0%
    West & Central Euro - 96% (yes!! 0% to 96%!!)

    Mine is the biggest swing I've see, and completely inaccurate.

    This an MDLP K16 for comparison.

    MDLP K16.png
    Y-DNA R-DF23>ZP149>ZP171 MDKA Thomas Doherty, b. 1825, Three Trees, Donegal, Ireland.
    mtDNA T2g1 MDKA Francoise Arguin, b. 1698, Camaret-Sur-Mer, Bretagne, France

  10. The Following 11 Users Say Thank You to MacEochaidh For This Useful Post:

     Amathusian (04-14-2017),  AnnieD (04-10-2017),  Camulogène Rix (04-09-2017),  Dewsloth (04-10-2017),  Jessie (04-10-2017),  JMcB (04-09-2017),  kikkk (04-09-2017),  Mike_G (04-10-2017),  sktibo (04-09-2017),  Táltos (04-09-2017),  Tolan (04-09-2017)

  11. #6
    Registered Users
    Posts
    2,449
    Sex

    There isn't just a problem of DNA changing from one region to another which FTDNA and others try to explain away as what the person inherited. There are also examples of very large discrepancies between parents and children and that can't be explained away. There is a kit at FTDNA where both parents get 37% Iberian DNA and one child gets 46% Iberian DNA. Another child gets 0% Iberian DNA but gets a very large amount of West & Central Europe instead, more than a parent that does have West & Central Europe, even though the other parent doesn't have any West & Central Europe. The differences between the parents and the child without Iberian DNA and the differences between siblings are statistically impossible to be real.

    I expected 23andme to still be better after the release of myOrigins 2.0 and it is. The results there are much closer to documented genealogy for most western Europeans and they use phasing so parents and children don't have large discrepancies.

  12. The Following 9 Users Say Thank You to ArmandoR1b For This Useful Post:

     Amathusian (04-14-2017),  AnnieD (04-10-2017),  Jessie (04-10-2017),  JMcB (04-09-2017),  MacEochaidh (04-09-2017),  Mike_G (04-10-2017),  REWM (04-09-2017),  Simon_W (04-16-2017),  sktibo (04-09-2017)

  13. #7
    Banned
    Posts
    404
    Sex
    Nationality
    USA
    Y-DNA (P)
    R1b>L21>Z253>L1066

    Ireland Germany England Netherlands Switzerland Scotland
    Anyone who truly understands how these admixture calculators work will not give much significance to them. There are several features of these analyses that need to be understood by testers.

    1. Such admixture calculators are based on the assumption that a modern person's DNA is comprised of a mixture of a finite number of well defined source populations, mixed together in various ratios, and that those source populations are genetically pure collections of ancient people with homogeneous DNA.

    This assumption is certainly not true. However, it is close enough to true that some useful information about a tree's history may be obtained by supposing it to be true.

    In reality, some of the hypothetical source populations lived in close proximity to each other and had continuous genetic mixing between them over extended time periods. So the working assumption upon which the model is based will sometimes lead to erroneous conclusions.

    2. There is a large degree of overlap in the genetic makeup of neighboring populations, especially in Western Europe. So, for example, a person from England may have a genetic makeup that looks similar to the genetic makeup of an average person from Germany, and a person from Germany may have a genetic makeup that looks similar to the genetic makeup of an average person from England. But this does not mean that the person from England actually is from Germany, or visa versa.

    But the admixture models assume the average, or modal, genetic makeup from a particular locale to be the "true" genetic definition of that locale. In other words, they may label an English person as German, when in reality they are just an English person who is not identical to the average English person.

    3. Admixture models are not a fully defined set of equations that can give a single, unique, correct solution. These models mix together hypothetical source populations in various ratios in order to derive the modern person's genetic makeup. But there are many possible recipes that can give the same outcome. By analogy, if one wants to make gray paint using primary colors, they can mix blue with orange, but they could also mix green with red. Both recipes give the same outcome, more or less. And it is not possible, simply from looking at the final grayish color, to know for certain which source colors were mixed together in order to obtain the resulting color.

    In other words, it is not possible for the admixture models to identify one unique, correct genetic recipe, and to exclude all other possible recipes that would give the same outcome.

    So the approach most often used by the testing companies is to randomly try a variety of recipes, and pick one of the recipes that creates the right shade of gray, so to speak. But this approach does not guarantee that the chosen recipe reflects the historical reality of how our modern person derived their genetic makeup. This is why parents and their children, or siblings, sometimes receive dramatically different results from their ethnic admixture analyses. Indeed, I have even read of cases in which the same person submitted their own DNA two different times, and received a significantly different outcome from the same DNA sample.

    Unfortunately, the testing companies present the test results with much more certainty that what the tests are actually capable of producing.

    So, to summarize, these ethnic admixture tests may contain some useful information. But they should all be taken with an enormous grain of salt.
    Last edited by miiser; 04-09-2017 at 06:50 PM.

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

     A Norfolk L-M20 (04-09-2017),  AnnieD (04-10-2017),  avalon (04-16-2017),  Babatunde (04-11-2017),  Camulogène Rix (04-09-2017),  Dewsloth (04-10-2017),  Jessie (04-10-2017),  JMcB (04-09-2017),  Judith (04-11-2017),  Mike_G (04-10-2017),  MitchellSince1893 (04-09-2017),  randwulf (04-09-2017),  Saetro (04-10-2017),  wombatofthenorth (04-11-2017)

  15. #8
    Silver Class Member
    Posts
    4,650
    Sex
    Nationality
    Canadian
    Y-DNA (P)
    R-Z198 (DF27)
    mtDNA (M)
    T2B-T152C

    England Scotland Poland Germany Canada Quebec Netherlands
    Quote Originally Posted by MacEochaidh View Post
    Just an example of the myOrigins failure. Known Ancestry 75% Isles, 25% French Canadian:

    Old my Origins:
    Isles - 71%
    Scandinavia - 8%
    West & Central Euro - 0% (yes, 0%)

    New myOrigins:
    Isles - 4% (yes! 71% to 4%!!)
    Scandinavia - 0%
    West & Central Euro - 96% (yes!! 0% to 96%!!)

    Mine is the biggest swing I've see, and completely inaccurate.

    This an MDLP K16 for comparison.

    MDLP K16.png
    What is the break down on the 75% Isles ancestry? You're mostly Irish aren't you?

  16. #9
    Registered Users
    Posts
    264
    Sex
    Location
    Oregon, U.S.A.
    Ethnicity
    Irish
    Nationality
    American/Canadian
    Y-DNA (P)
    RL21-DF23>ZP149
    mtDNA (M)
    T2g1

    Northern Ireland Ireland Scotland Canada Quebec
    Quote Originally Posted by sktibo View Post
    What is the break down on the 75% Isles ancestry? You're mostly Irish aren't you?
    Yes, mostly Irish from Ulster; Donegal, Antrim, Down, Derry. I also have a 2x great grandfather from South Uist in the Outer Hebrides of Scotland and several lines that trace from Ulster to Scotland. I have little doubt that there is some English ancestry as well, as I have ancestors in Ireland with the names Hemmingway and Saville.

    My Dad was born in the north of Ireland and adopted. My maternal grandfather is from Belfast. Every other ethnicity estimate shows my Isles scores swamping my 25% French Canadian ancestry. This includes older FTDNA estimates. 96% West Central Europe over 4% Isles is ridiculous
    Y-DNA R-DF23>ZP149>ZP171 MDKA Thomas Doherty, b. 1825, Three Trees, Donegal, Ireland.
    mtDNA T2g1 MDKA Francoise Arguin, b. 1698, Camaret-Sur-Mer, Bretagne, France

  17. The Following 4 Users Say Thank You to MacEochaidh For This Useful Post:

     AnnieD (04-10-2017),  Jessie (04-10-2017),  JMcB (04-09-2017),  sktibo (04-09-2017)

  18. #10
    Silver Class Member
    Posts
    4,650
    Sex
    Nationality
    Canadian
    Y-DNA (P)
    R-Z198 (DF27)
    mtDNA (M)
    T2B-T152C

    England Scotland Poland Germany Canada Quebec Netherlands
    Quote Originally Posted by MacEochaidh View Post
    Yes, mostly Irish from Ulster; Donegal, Antrim, Down, Derry. I also have a 2x great grandfather from South Uist in the Outer Hebrides of Scotland and several lines that trace from Ulster to Scotland. I have little doubt that there is some English ancestry as well, as I have ancestors in Ireland with the names Hemmingway and Saville.

    My Dad was born in the north of Ireland and adopted. My maternal grandfather is from Belfast. Every other ethnicity estimate shows my Isles scores swamping my 25% French Canadian ancestry. This includes older FTDNA estimates. 96% West Central Europe over 4% Isles is ridiculous
    Well it looks like West and Central Europe has a lot of overlap with Britain, and might be less Germanic than the British Isles category now. Still inaccurate, yes, but it might be an indicator of that shared Brythonic-French DNA connection. As I currently understand it, if we go back in time far enough Western Europe and Britain were more or less Bell Beaker stock, so there's that similarity. 96% is still just stupid though.

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

     AnnieD (04-10-2017),  MacEochaidh (04-09-2017)

Page 1 of 11 123 ... LastLast

Similar Threads

  1. Accuracy of FTDNA My Origins Results
    By Oceane81 in forum FTDNA
    Replies: 58
    Last Post: 09-11-2019, 07:12 PM
  2. FTDNA My Origins results
    By galon07 in forum FTDNA
    Replies: 205
    Last Post: 09-17-2018, 03:22 AM
  3. ftdna ancient origins results, italian
    By patrizio22 in forum Autosomal (auDNA)
    Replies: 75
    Last Post: 03-23-2017, 06:12 PM
  4. My FTDNA My Origins results; confused!
    By Lizzy28 in forum FTDNA
    Replies: 19
    Last Post: 09-15-2016, 09:53 PM

Posting Permissions

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