Announcement

Collapse
No announcement yet.

Advice needed on DNATribes

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

  • Advice needed on DNATribes

    Hi,
    Just wondering what peoples' views on DNATribes is. I'm a bit confused about it...on Kershner's blog the test results seem nearly always nothing like what people know about their paper trail. Can anyone explain to me how it works and what all the "matches" mean? They all seem to be like 10 this and 0.2 that?
    Thanks.

  • #2
    Originally posted by burto
    Hi,
    Just wondering what peoples' views on DNATribes is. I'm a bit confused about it...on Kershner's blog the test results seem nearly always nothing like what people know about their paper trail. Can anyone explain to me how it works and what all the "matches" mean? They all seem to be like 10 this and 0.2 that?
    Thanks.
    The information on the individual's results pages says it all, but to paraphrase, you have match probabilities for populations and then a likelihood of origins on particular continents.

    The probabilities of matches to populations are actually compared to a match to a "generic" population, something which I've never seen before. But the idea is to compare the population matches against each other to see which is most likely. Same with the continents.

    Did you really get population match probabilities of 10 or so? My weakest was 24.5, and strongest was 2,215.1.

    Jim
    Hot day in Cardiff

    Comment


    • #3
      [QUOTE=Jim Honeychuck].

      Did you really get population match probabilities of 10 or so? My weakest was 24.5, and strongest was 2,215.1.

      Sorry I was just picking random numbers there

      So where are they getting that 2,215.1 from then? I've heard they use CODIS and you can put your results on a database online to compare with other ethnic groups, but how do they interpret your results and does this database work too?
      Thanks.

      Comment


      • #4
        [QUOTE=burto]
        Originally posted by Jim Honeychuck
        .

        So where are they getting that 2,215.1 from then? I've heard they use CODIS and you can put your results on a database online to compare with other ethnic groups, but how do they interpret your results and does this database work too?
        Thanks.
        That match was from India. Bit of a mystery, but the company says it's valid. Some Indians and Pakistanis have West Asian ancestry, that has to be it. My ancestors seem to have been in southern Russia for a long time.

        I think you are referring to the downloadable Omnipop calculator. It didn't tell me anything new, but it seems to be informative for other people.

        Jim

        Comment


        • #5
          Do you have a paper trail/haplogroups to go with your results, or does this type of test go back thousands of years? I want to try and find an explanation for this 11% East Asian from the DNAPrint test.
          Thanks.

          Comment


          • #6
            Originally posted by burto
            Do you have a paper trail/haplogroups to go with your results, or does this type of test go back thousands of years? I want to try and find an explanation for this 11% East Asian from the DNAPrint test.
            Thanks.
            This test goes back to infinity, I guess.

            Apart from my Irish side, which DNA Tribes can't detect yet, and eight matches to Indian populations, my DNA Tribes results tracked well with my Y-DNA results, a few surname clues, and known population movements in Eastern Europe.

            I have a rather weak match with Japanese, which I'm ignoring.

            Can't advise on DNAPrint. I thought it would be too crude a measure, and I never took that test.

            Jim

            Comment


            • #7
              Originally posted by burto
              Do you have a paper trail/haplogroups to go with your results, or does this type of test go back thousands of years? I want to try and find an explanation for this 11% East Asian from the DNAPrint test.
              Thanks.
              DNATribes offers good guidance on how to interpret results on their Blog on their site. That would be a good place to read-up on their test offering.

              The 13 highly variable CODIS markers that DNAT uses were selected for forensic purposes - a way to determine paternity and identify criminals, victims, human remains etc. - they were never meant to be the basis for a biogeographical test. They are very good as "DNA fingerprints" and because they have become a world forensic standard there are many sets of results available from all over the world. (Check-out DNAT's Populations to see who is currently represented in their combined database. There are regular new additions).

              So, if you get your CODIS markers sequenced, DNATribes can, through the magic of statistics, match your marker profile to the profiles in their database to produce "significant" matches indicating biogeographical ancestry.

              Here are the caveats -

              Each marker has two alleles. Nominally, one allele comes from your mother and one from your father, although that is not an ironclad rule. But it is something of an ironclad rule that one cannot determine with certainty from which parent any particular allele derives.

              Another limitation to the test is that it does not proceed allele-by-allele, but by overall profile. So you might be matched to populations that share a "parallel" recent history of ethnic admixture but to whom you are not directly related. E.G. I get a cluster of matches to Central and South American mestizo populations even though my mother was Native American of (evidently) Upper Midwest North American origin.

              And finally testees get "odd" matches and weak scores because DNATribes' database does not yet contain samples from populations to whom they are most particularly related.

              It follows from all this that DNAT test results are "fuzzy." And because genealogical researchers desire conclusive proof, many may be dismayed by a test that produces as many questions or possibilites as answers.

              However, in some instances, some clusters of results have a geographical distribution close to that of a particular y-haplogroup. And I suppose when mt-haplogroup distributions are better mapped some clusters will seem to correspond to their geographic distributions as well.

              I believe that DNATribes database will grow and they will improve upon their statistical methods. I believe that I will learn more from my results in the future. And I believe that when I get into my "brick wall" I will find support and confirmation from results already in-hand - that the puzzle pieces are indicated by my results and I will, in time, know where they fit.

              Tom

              Comment


              • #8
                I have not yet taken an autosomal test because I am waiting for a more valid, reliable and specific measure. I am certainly hesitant about DnaTribes. Their forum seems to suggest that the test results can be discrepant with other results or information. This especially seems the case in geographical areas marked by significant haplogroup diversity. (The forum suggests that the results are closer for respondents of indigenous populations marked by less diversity). There are a number of statistical problems with the DnaTribes approach: small number of loci tested, reliance on a small amount of the available statistical information ( concern with modal patterns only) and limited sampling of geographical groups that are well represented in other tests. Anyone of these problems can throw the results off. The upcoming revisions for AncestrybyDna look like a better choice.
                Last edited by josh w.; 20 July 2006, 12:13 PM.

                Comment


                • #9
                  Hmm interesting stuff. I'll pass all this onto my Mum then let her decide. By the way when will the new Ancestry upgrade come out? Does anyone know how it will be better than 2.5?
                  Many thanks.

                  Comment


                  • #10
                    Originally posted by josh w.
                    I have not yet taken an autosomal test because I am waiting for a more valid, reliable and specific measure. I am certainly hesitant about DnaTribes. Their forum seems to suggest that the test results can be discrepant with other results or information. This especially seems the case in geographical areas marked by significant haplogroup diversity. (The forum suggests that the results are closer for respondents of indigenous populations marked by less diversity). There are a number of statistical problems with the DnaTribes approach: small number of loci tested, reliance on a small amount of the available statistical information ( concern with modal patterns only) and limited sampling of geographical groups that are well represented in other tests. Anyone of these problems can throw the results off. The upcoming revisions for AncestrybyDna look like a better choice.
                    Ancestryby DNA only devides Europe into Northern European, Mediteranean, Middle Eastern and South Asian. I know pretty sure I'm about 100% Northern European without testing. What I'd like to know is what parts of Northern Europe my ancestors came from. DNA Tribes has reference populations from Finland, Norway, Estonia, Iceland, Germany, Poland, etc. I'm just waiting until they also have populations from Sweden and maybe Denmark before I'll order a test.
                    Last edited by Eki; 20 July 2006, 02:23 PM.

                    Comment


                    • #11
                      I would like to read what DNAPrint has planned for their next iteration of ABDNA 2.5 as I did not attend their recent e-conference.

                      As to CODIS markers, you need not rely, exclusively, upon DNATribes' analysis as you can research the markers on your own.

                      The Europeans have a slightly different forensic standard but you can find Euro population results and search most of your CODIS markers through ENFSI at - www.str-base.org.

                      And you can download Brian Burritt's Omnipop spreadsheet to search CODIS markers against his database of populations.

                      And/or try the Royal Canadian Mounted Police db at - www.csfs.ca/pplus/profiler.htm

                      Anyone interested might also visit the site for the German lab that does the labwork for DNATribes. They offer guidance for research, links to even more databases, the Omnipop download, and an additional STR panel that will let you make full use of all the field results online. Go to - www.dna-fingerprint.com


                      Tom

                      Comment


                      • #12
                        Originally posted by Eki
                        Ancestryby DNA only devides Europe into Northern European, Mediteranean, Middle Eastern and South Asian. I know pretty sure I'm about 100% Northern European without testing. What I'd like to know is what parts of Northern Europe my ancestors came from. DNA Tribes has reference populations from Finland, Norway, Estonia, Iceland, Germany, Poland, etc. I'm just waiting until they also have populations from Sweden and maybe Denmark before I'll order a test.
                        The ENFSI db at www.str-base.org has samples for Austria, Belgium, Switzerland, Czech Republic, Denmark, Estonia, Spain, Finland, France/Lille, France/Toulouse, Croatia, Ireland, North Ireland, Italy, Netherlands, Norway, Poland, Portugal, Sweden, Slovenia, England/Wales, Scotland/Dundee, and Scotland/Glasgow.

                        Again, ENFSI has a slightly different forensic standard, but you can get all the markers they use through DNA-Fingerprint. DNATribes once offered their analytical service as an option to those with DNA-F results in-hand. Or if you got results through DNATribes, you can ask them to release your sample to DNA-F for further tests and save the cost of a redundant DNA extraction.

                        I agree that some of DNATribes outputs seemed skewed by the manner in which they weight homozygotic results and the limitations of their population databases. But there are workarounds to that because CODIS is an open standard. On the other hand, some have been shocked by DNAPrint's pronouncements and, unfortunately, those "readings" issue from a black box.

                        Standard or stick?

                        Tom

                        Comment


                        • #13
                          So do you have to pay DNA fingerprint extra in order to check your markers against the ENFSI database, or do the markers just come with the DNATribes results?

                          Comment


                          • #14
                            DNATribes test is based on 13 CODIS markers (the American standard). DNA-Fingerprint does the CODIS labwork for DNATribes - the 13 CODIS markers are covered by DNA-F's Autosomal Panel 1.

                            If you got test results through DNATribes, your DNA sample resides in Germany with DNA-Fingerprint. If you request DNATribes to waive their confidentiality agreement, DNA-Fingerprint would be able to conduct further tests (the tests DNA-F offers) on your already extracted DNA sample in their possession at their lab without the unnecessary expense of an additional and redundant DNA extraction. Absent such a request, your confidentiality agreement with DNATribes stands.

                            DNATribes only offers the analysis they offer - their test. DNA-Fingerprint only does lab work, they do not offer analysis. I suppose that if you wanted your markers analyzed against databases other than DNATribes', you would have to engage the services of another party who was willing and qualified to conduct such an analysis. This applies to either DNA-F's Autosomal Panel 1 or Autosomal Panel 2.

                            When DNATribes launched their service they had an introductory offer to persons who already had CODIS/Autosomal Panel 1 results through DNA-F, to analyze those CODIS results according to DNATribes' proprietary statistical methods. I do not know if that offer still stands.

                            Tom

                            Comment


                            • #15
                              I suggested AncestrybyDna because they rely on hundreds of markers rather than just the limited number of CODIS markers. This should make their results alot more reliable. The advertised future version of AncestrybyDna should be country or region specific (whenever it hits the market).

                              Comment

                              Working...
                              X