Announcement

Collapse
No announcement yet.

Comparing DNA test companies for unknown ancestry

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

  • Comparing DNA test companies for unknown ancestry

    I have finally tested with all three of the biggest DNA testing companies: AncestryDNA, 23andMe, and FamilyTreeDNA. I also uploaded to Gedmatch.com. Here is my experience in case it helps anyone else who is adopted or otherwise knows nothing about their own genetic ancestry.

    My mother was adopted, and my father was an NPE. I had hints about who the people might be but testing confirmed.

    My pick as best overall autosomal testing kit: AncestryDNA. Why? Two reasons. A much larger database of people who have sent in their DNA. My paternal uncle who does not even know who I am showed up after a couple of years. If I had no idea about my paternity or maternity, that would have basically led me to my father. Plus, the number of cousins who match 50+ cM is at least double that of the next highest kit, 23andMe.

    The second reason is how well it is integrated into the genealogy records and family tree auto-matching. It is so easy to figure out how I may be matched with cousins even if they only provide a partial family tree. Ancestry.com will analyze both of your trees and try to help you find the common ancestor.

    If I only had enough money to do one kit, it would be AncestryDNA, but I have gotten something out of all the companies. The y-DNA test from FamilyTreeDNA pointed me towards the paternal surname based on the number of times that surname showed up in y-DNA matches. That surname was the surname of my paternal uncle that showed up on AncestryDNA. I have not been as impressed with 23andMe, although I am glad they said I don't have a higher risk of colon cancer. For some reason I don't have very many close matches with 23andMe, even though it is supposed to be second behind AncestryDNA for participants.

    So this has been my experience. Please share your comparisons. When I first started this 3 1/2 years ago, I was looking for a thread like this.



  • #2
    I think 23andMe, which is where I first tested, isn't so much supportive of genealogical endeavors. Is more about health-related stuff there. My experience with the response rate when trying to contact matches at 23andMe is that it's utterly dismal, but I still enjoy the web-site. Among other things, it was also the first one (of several) that indicated some Italian for me - a completely unexpected discovery. This has led me to dig deeper to try and uncover that part my ancestral heritage. My mother was adopted, so this is the NPE I'm speaking of, i.e., my unknown maternal grandfather. I've posted on this subject in a another string, so I won't regurgitate that here.

    Each DNA provider brings something to the table. Actually, it was both MyHeritage and FTDNA that gave me some matches (and trees) that allowed me to look backwards and, in theory, "find" a couple who potentially are MRCAs (3x great grandparents). I have many more matches at Ancestry, but the largest ones are only in the mid-50's (cMs). So, 4th cousins or later (we all have bazillions of those). Check that, I have one Ancestry match that's 115 cM's across 8 segments - probably a 3rd cousin. However, as "luck" would have it, she, too, has NPE's in her ancestral lines! Is unfortunate that the part of her tree that she does have is that with which I do NOT match. Sigh... A pretty big bummer there, as this match could've saved me at least one generation of searching - not an insignificant thing.

    I'm very close to exhausting all my "official" avenues to find out more about my mother's adoption. After that, it will be "all DNA", so I do hope there's a closer match for me out there somewhere that someday comes to the fore. It doesn't need to be a first cousin, but maybe a second cousin (or even a 3rd cousin) would at least give me more momentum (hope) moving forward.
    Last edited by mjclayton31; 25th December 2018, 12:31 PM.

    Comment


    • #3
      I have 35 matches that are 50 cM or higher on AncestryDNA. 7 matches on 23andMe are 50 cM or higher. Amazingly I have 263 50 cM or higher on FamilyTreeDNA. Maybe there are different standards for calculating total shared cM?

      Comment


      • #4
        I have, for example, 59 matches at 50 cMs or more at Ancestry. That's not the issue. The issue is that only a few are over 50 cMs as to the NPE that I'm trying to solve - who my mother's father is/was. And yes, there are different algorithms, etc., that are used by each DNA company in calculating shared segments and cMs. They're generally similar in result, but most certainly not exact, hence, the discrepancy from company to company. This also (likely) speaks to the dearth of Italians that have done the test, i.e., relative to the more common English/Irish /Scottish cousin test-takers.
        Last edited by mjclayton31; 26th December 2018, 09:23 PM.

        Comment


        • #5
          Originally posted by ec1970 View Post
          I have 35 matches that are 50 cM or higher on AncestryDNA. 7 matches on 23andMe are 50 cM or higher. Amazingly I have 263 50 cM or higher on FamilyTreeDNA. Maybe there are different standards for calculating total shared cM?
          FTDNA tests a broader section of Autosomal DNA, further, they pretty much only look at autosomal, while others "peek" into mito and Y DNA at times. (The reason FTDNA doesn't should be pretty evident) Only NatGeno and MyHeritage sample more, and not by much. Nowhere near the difference between FTDNA and Ancestry atDNA tests.

          https://isogg.org/wiki/Autosomal_DNA...​​​​

          Comment


          • #6
            I believe FTDNA still does the testing for NatGeno.

            Comment


            • #7
              Addressing the last two comments (and of course, noting specifically that the Genographic test by National Geographic does not provide matching to others):

              For any of the older Geno 2 tests (pre-Nov. 2016), but not the current version (except for international kits; see below), FTDNA has and still does the lab testing. From the National Geographic Geno 2.0 website, with applicable statement in red:
              1. How do I know what Geno version I have?
                The actual product will tell you everything you need to know. If you have a white Helix box that slides inside a black case, that is the Geno 2.0 Next Gen Helix Product Kit. This kit is processed by our partner lab, Helix. If you have a black box that opens on the right side, that is either our Geno 2.0 Next Gen Kit, or Geno 2.0 or Geno 1.0 kit. These three versions of the kit are processed by our partner lab, FTDNA.
              Regarding the SNPs tested by the current Geno 2.0 Next Generation test, they also say on the above page (my bolding):
              The new test also scans more than 250,000 (autosomal) markers from across your entire genome that were inherited from both your mother and father, revealing insights into those ancestors who are not on a strictly maternal or paternal line. Included among these markers is a set that will also reveal if you are carrying DNA that came from our hominin cousins, the Neanderthals
              According to the ISOGG Autosomal DNA Testing Comparison chart, the amount of autosomal SNPs tested on the latest Geno 2.0 Next Generation Helix product test (per ISOGG's page for the Genographic Project, from Nov. 2016 to present, US only) is described as "Full details of new test yet to be announced." That Genographic ISOGG page states that "US kits are now processed by Helix. International kits are processed by Family Tree DNA." It does not show how many autosomal SNPs the Helix-processed kits will test, although the National Geographic Geno 2.0 page linked above does (250,000 is well less than half of the autosomal SNPs that FTDNA tests).

              On the ISOGG comparison chart, the previous version of the Geno 2 kit is shown as having tested about 700,000 autosomal SNPs, whereas FTDNA is still shown as testing about 690,000. Yet ISOGG's Autosomal SNP comparison chart shows FTDNA as testing 702,540 markers, Genographic 2.0 Next Generation Helix with 144,333 markers, and MyHeritage's GSA chip testing 702,442. Note: FTDNA's lab does the testing for MyHeritage. The SNP amounts for the various editions of chips for 23andMe and Ancestry are also shown on the latter chart, as well as for Genes for Good and LivingDNA. It's apparent that the days of optimum compatibility are pretty much over for companies that accept files from others.

              I'm not sure if the charts are up to date for those companies that have changed over to the Global Screening Array (GSA) chip, which is the only kind now offered by Illumina (which has no competition in this market). Louis Kessler has done a very detailed SNP comparison of his test results from five DNA testing companies (23andMe, Ancestry, FTDNA, LivingDNA and MyHeritage), if anyone is interested in a breakdown.

              Comment


              • #8
                Originally posted by KATM View Post
                I'm not sure if the charts are up to date for those companies that have changed over to the Global Screening Array (GSA) chip, which is the only kind now offered by Illumina (which has no competition in this market). Louis Kessler has done a very detailed SNP comparison of his test results from five DNA testing companies (23andMe, Ancestry, FTDNA, LivingDNA and MyHeritage), if anyone is interested in a breakdown.
                LivingDNA has a new chip produced by a new vendor. Good news indeed!:

                Link removed (please no links to outside company websites)
                Last edited by Darren; 4th January 2019, 12:10 AM. Reason: please no links to outside company websites

                Comment


                • #9
                  Originally posted by bartarl260 View Post

                  FTDNA tests a broader section of Autosomal DNA, further, they pretty much only look at autosomal, while others "peek" into mito and Y DNA at times. (The reason FTDNA doesn't should be pretty evident) Only NatGeno and MyHeritage sample more, and not by much. Nowhere near the difference between FTDNA and Ancestry atDNA tests.

                  https://isogg.org/wiki/Autosomal_DNA...​​​​
                  Interesting. Still with all these 50+ cM relatives on FTDNA, most do not provide a family tree so the matches are nearly useless unless you want to get in touch with each person individually. I guess the automatic tree comparison that AncestryDNA does has been fairly useful for me. I have been able to create a convention in my main tree where I add a special suffix to any person that I have verified as an ancestral line to another AncestryDNA relative. That way I can easily visualize on my family tree which lines I have "proven" with DNA. There are only a few lines that have come up empty so far and therefore I think there may be a mistake in the "paper" ancestry record. Or else no relatives on those lines have yet tested. But with so many people doing DNA genealogy today, almost all ancestral lines should have a representative who has tested.

                  Thanks to everyone who has added interesting information about various testing methodologies by different companies.

                  Comment


                  • #10
                    For my matches on ancestry I've actually taking to noting "the path" a match takes to reach a known ancestor. If it's a distant enough ancestor(or one I don't know too well), I'll some times even backt-trace from there.

                    Makes it a lot easier for trying to figure out where a match's line might tie in when you see most of what you need to know with just a mouse-over or mouse-over + click rather than needing to switch to a spreadsheet or load another tab on my webbrowser.

                    Comment


                    • #11
                      Originally posted by bartarl260 View Post
                      For my matches on ancestry I've actually taking to noting "the path" a match takes to reach a known ancestor. If it's a distant enough ancestor(or one I don't know too well), I'll some times even backt-trace from there.
                      I definitely do this. Basically every ancestor along the path to the common ancestor has become a "proven" DNA ancestor. I know it is not 100% because we could both have mistaken trees. I should probably differentiate actual common DNA ancestors with just those along the path to a common ancestor.


                      Comment


                      • #12
                        I agree, although I still wish Ancestry would add a chromosome browser, I have had the most success there, including identifying my grandfather's bio father.

                        Comment

                        Working...
                        X