Announcement

Collapse
No announcement yet.

About the latest “My Origins Version 3”

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
    KATM
    mtDNA: K1a3 / YDNA: R-FGC46379

  • KATM
    replied
    Updates to reference populations are based on further research of results from more people representing populations tested. Ideally each project, or DNA testing company, would provide descriptions for any updates to specific populations, so we are at their mercy for that information. Usually a scientific paper used for the projects (such as the projects in the chart I posted), or a white paper in the case of the DNA testing companies, will get into such details.

    While apparently FTDNA has so far not issued a white paper for myOrigins 3.0 (latest version), you can find the white papers for myOrigins from 2014, and an updated version from 2017, at the Internet Archive (aka "Wayback Machine"). The 2014 white paper contains the original myOrigins Cluster Descriptions used at that time, which were later revised. Other examples of white papers for admixture/bioethnicity are:

    Leave a comment:

  • KATM
    mtDNA: K1a3 / YDNA: R-FGC46379

  • KATM
    replied
    John McCoy, without doing any more research into your questions (and thus having no other answers), I recalled I had this admittedly dated image which I downloaded in 2016, probably from some presentation. We can see that this represents the overlap of the reference populations used at or before that time for 23andMe, Ancestry.com, and FTDNA. I suspect that for some purposes, any or all of these companies may include some degree of input based on results from customers (as 23andMe and FTDNA are shown using in the chart). I don't believe that FTDNA has released a white paper for myOrigins 3.0, and off the top of my head I don't recall if earlier white papers discussed where the reference populations used for earlier myOrigins versions were obtained.

    Ref_pops_3_companies_overlap.png
    • The 1000 Genomes Project apparently continues to be updated.
    • I'm not sure if the Human Genome Diversity Project gets updated.
    • Apparently the International HapMap Project website was taken down due to security issues, before it was going to be decommissioned due to the 1000 Genomes Project becoming the research standard.
    • The Estonian Biocentre was established in 1986 with "the goal is to promote research and technological development (RTD) of gene and cell technologies in Estonia."
    • The SMGF was the Sorenson Molecular Genealogy Foundation, and its history is described in the ISOGG Wiki. Ancestry acquired the SMGF data in 2012, and in 2015 removed the SMGF database, to much chagrin in the DNA community.

    Leave a comment:

  • John McCoy
    FTDNA Customer

  • John McCoy
    replied
    The question that occurs to me is where the underlying data come from and how the "reference groups" used in the various admixture algorithms have been developed and confirmed.

    Are these samples submitted by individuals to one or more autosomal DNA testing companies? Or samples collected by scientists specifically for genetic anthropology research? Or collected from residents of the specified regions for some other purpose? As the admixture results offered by different vendors become more and more specific, I think it is important to know where the samples have come from and how the data have been manipulated.

    Leave a comment:

  • khazaria
    FTDNA Customer

  • khazaria
    replied
    Thank you for telling us about you family seeing this phenomenon of false estimates in the Jewish categories, Dr.SYSTEM. It is well-established that Mizrahi Jewish and nearby non-Jewish populations didn't mingle during the autosomal timeframe and were basically genetically isolated from each other since the 7th century at the latest.

    Unfortunately, some Middle Eastern people from non-Jewish families trust the fleeting glimpses of possible Jewish ancestry from the small percentages in the Mizrahi Jewish and/or Sephardic Jewish and/or Yemenite Jewish and/or Ashkenazic Jewish categories that show up in MyOrigins or MyHeritage or 23andMe and then when they find distant identical-by-chance matches who are Jewish who supposedly match them on small segments, they think they've proven it exists in their heritage. In populations that test infrequently, the proportion of false matches is higher than for someone who comes from a well-tested population. False matches are those that fail to pass rigorous phasing and triangulation tests inside GEDmatch. The segments often are only 6, 7, or 8 cM in length and the total percentages of DNA shared are usually under 0.2%.

    Leave a comment:

  • KATM
    mtDNA: K1a3 / YDNA: R-FGC46379

  • KATM
    replied
    FTDNA has not yet put out information about the reference populations used for myOrigins 3.0, such as in a white paper. The reference populations are used to determine to which populations clusters each customer is closest. FTDNA, as far as we've been told, does not use information provided by individual testers to come up with the population clusters and percentages for that tester. On the FTDNA Learning Center page "Does the myOrigins tool use non-autosomal DNA results?" it says:
    Only your autosomal DNA results from the Family Finder microarray chip are used.
    While this statement is addressing other forms of DNA (mtDNA, X chromosome, or Y-DNA), I would also take it as some evidence that ONLY your autosomal DNA results are used to compare to reference populations. Your forum profile shows you are from Gedrosia, so your result of 97% for "West Central Asia: Indus Valley" sounds pretty accurate, unless I'm wrong about the location. I would disregard the trace amounts. If you look at your myOrigins 3.0 results, and hover over the small "i" in a black circle shown with your trace amounts, it shows this message:
    A trace percentage indicates a very small amount of shared DNA in common with the corresponding population. Smaller DNA segments are more likely to be misattributed.
    So, consider that these trace amounts in your estimate may be misattributed. There are similarities between different populations which can be misidentified (misattributed) as one or the other. Among others, some people with Mediterranean and Middle Eastern ancestry sometimes show a misattributed percentage of Jewish in their results. Also consider that myOrigins may reflect ancestry beyond what we can find out about our own in the last few hundred years: the page "How many generations in the past do myOrigins results reflect?" in the Learning Center says:
    myOrigins results are your personal genetic ancestry that reflects the last 100 to 2,000 years (about four to 80 generations). They may also reflect one population that mixed with another in ancient times and became fixed in one of your populations.
    We all need to keep in mind that ALL ethnicity estimates, from ALL the companies, are still their most accurate only at the continental level; once beyond that, interpretations may be off. And ethnicity estimates ARE NOT the most accurate part of DNA testing for genealogy. Rather, the closeness of your matches and where they match you on the chromosomes ARE the most accurate. The ancestry of your matches will reflect your known ancestry better than the ethnicity estimates.
    KATM
    mtDNA: K1a3 / YDNA: R-FGC46379
    Last edited by KATM; 3 October 2020, 08:25 PM.

    Leave a comment:

  • Dr.SYSTEM
    Y-DNA: G-SK1127

  • Dr.SYSTEM
    started a topic About the latest “My Origins Version 3”

    About the latest “My Origins Version 3”

    Greetings,

    Lately I have checked the updated results for my V3 Origins and they are as follow:

    West Central Asia: Indus Valley 97%
    Middle East Jewish: Mizrahi Jewish <2%
    Arabian Peninsula <1%
    Anatolia, Armenia, & Mesopotamia <1%
    ==========================================
    My V2 old results were as follow:

    Central/South Asian 55%
    Asia Minor 45%
    =========================================
    I wonder on what genetic bases these percentage have been calculated?!

    It’s seemed to me they gave me the highest percentage because I have given them my latest ancestor’s location!!

    Is the new "My Origins Version 3" cluster zones represent a realistic genetic map distribution??

    Finally, Why I’m getting a trace of a “Jewish Diaspora” even it’s very low?!! (With all my respect to those people cultural heritage)

    Can anyone explain in a scientific logical sense about these calculations!

    Regards
Working...
X