Announcement

Collapse
No announcement yet.

Relative Importance of Matches

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Relative Importance of Matches

    In the DNA test results, there seem to be three major criteria for matching others tested. Surname, allele matches and haplogroup. In addition to the surname and allele matches, Recent Ethnic Origins, Most Recent Common Ancestor, and Haplogroups are reported. Within the surname, I understand the allele matches, but do matches outside of the surname group have any signifigance? I notice that those tested are asked to enter "unknown" if they are not sure of the country of origin. There are also "unknowns" shown in the REO and Haplogroup reports. Does the large numbers of unknowns in the REO result just reflect the input of those tested? Do the unknowns shown in the haplogroup report mean that although a haplogroup may have been predicted, it isn't "known" until another $65 is spent on tests? Also, do 12/12 matches have any signigance if those tested are of different surnames and haplogroups?

  • #2
    Within the surname, I understand the allele matches, but do matches outside of the surname group have any signifigance?
    It depends on how many markers you are talking about. At the 12 marker level, matches with other surnames don't mean much at all, unless there is a known connection between the two surnames. And even then I would not draw any conclusions until further testing was done. Even within the same surname 12/12 matches can go bad when the test is upgraded to 25 markers.

    At the 25 marker level, an exact or very close match might be worth further investigation, particularly if there is a known connection between the surnames. Were your ancestors and those of the other match in the same area at the same point in time? If so, then maybe there is a possible connection. If you or your match seem to be an "orphan" within your surname projects maybe you're connected.

    At the 37 marker level, an exact or very close match would certainly require further investigation.


    Does the large numbers of unknowns in the REO result just reflect the input of those tested?
    Yes - Testees are instructed to answer "Unknown Origin" when their ancestor's origin is uncertain or not known.

    Do the unknowns shown in the haplogroup report mean that although a haplogroup may have been predicted, it isn't "known" until another $65 is spent on tests?
    No - the "unknown" refers to country of origin

    Also, do 12/12 matches have any signigance if those tested are of different surnames and haplogroups?
    See my answer to the first question above concerning 12/12 matches with different surnames.

    As for different haplogroups - if two people have different haplogroups they can NOT be related. Haplogroups pertain to ancient origins (think 10,000 or 10's of 1000's of years) and can only be truly determined by SNP(Single Nucleotide Polymorphism) testing. SNP is a completely different type of test than the STR testing used to determine your haplotype (marker values).

    If you have not taken the SNP test to determine your haplogroup, FTDNA compares your haplotype to the haplotypes of those who have taken the SNP test to predict your haplogroup. If FTDNA cannot predict your haplogroup based on your haplotype they leave your predicted haplogroup blank.

    FTDNA further states "for any predicted results we see no reason for ordering a SNP test to confirm the Haplogroup. If a – is in the HAPLO field then we feel that the comparative results are not clear and unambiguous and if the kit holder wants to know their SNP with 100% confidence they may consider ordering a SNP confirmation test."

    The haplogroup page on your FTDNA personal site compares your known or predicted haplogroup with "the world-wide database of Dr. Hammer and customers who have had their SNP tested by us. The comparative Haplogroups shown were confirmed by SNP tests."

    Although in theory it is possible for two people to have the same haplotype but have different haplogroups, no one has yet to show a single case where this has happened. Until someone can provide an example of this I see no need to be concerned about haplogroups and SNP testing. Although it may be great interest to those into anthropolgy it really does not provide any real information for those interested in convention genealogy (within the time frame of surnames).

    John
    Last edited by jablair; 4 May 2005, 11:22 AM.

    Comment


    • #3
      John, for those of us not tested at ftDNA, could you comment on the reliability of haplogroup estimates provided by https://home.comcast.net/~whitathey/haptest.htm ? Mine is based on 30 markers.

      Comment


      • #4
        Originally posted by [email protected]
        John, for those of us not tested at ftDNA, could you comment on the reliability of haplogroup estimates provided by https://home.comcast.net/~whitathey/haptest.htm ? Mine is based on 30 markers.
        I could not reach the page you referred to, but it really doesn't matter. As I hinted at in my earlier post, I have no use of haplogroups, don't pay any attention to them, and would have no idea about the reliability of the haplogroup estimates provided.

        Hopefully someone who is interested in haplogroups will be able to answer your question. Good luck,

        John

        Comment


        • #5
          Originally posted by [email protected]
          John, for those of us not tested at ftDNA, could you comment on the reliability of haplogroup estimates provided by https://home.comcast.net/~whitathey/haptest.htm ? Mine is based on 30 markers.
          What I hear is that it is pretty reliable, although not infalible.

          If your haplogroup is one of the rare ones, one approach you can take is to notice how the percentage figures change as you go entering each marker. For example, you should compare the results you get on 12 markers versus 25, etc.

          If you you feel there is much ambiguity the definitive answer might be an SNP test.

          Victor

          Comment

          Working...
          X