Announcement

Collapse
No announcement yet.

Match on GEDmatch but not FTDNA - newbie needing help

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

  • Match on GEDmatch but not FTDNA - newbie needing help

    Hello all

    I received my Ancestry Autosomal DNA last week and uploaded it to GEDmatch. I've been in contact with my closest DNA match relative, with whom I share the following (we have no x matches),

    Chr 3 Start Location 97,523,082; End Location 110,720,621; cM 8.2
    Chr 17 Start Location 7,870,502; End Location 11,119,018; cM 11.1
    Chr 19 Start Location 211,912; End location 3,103,982; cM 9.0
    Chr 19 Start Location 3,597,642; End location 8,920,808; cM 15.7
    Chr 22 Start Location 31,838,996; End location 35,701,055; cM 8.1


    I have also uploaded my Ancestry DNA to FTDNA, which is supposed to give me full access to the FTDNA Family Finder database (I also upgraded to the Chromosome Browser).

    Although I have 129 people matching me on FTDNA, none of these are with my closest match relative on GEDmatch. (She is also on FTDNA, by the way).

    If someone could explain this, I would be grateful; e.g. is my match (as shown above) from GEDCOM insufficient to qualify as a match on FTDNA?

    Thanks

    Liz

  • #2
    For one, if you have only transferred Ancestry Raw Data, you will not get "full matching" so to speak.

    Anyone FTDNA normally would deem the 4th to Remote Range or Further are not included with transfers due to the compatibility issues between the testing chip each company uses.

    Each test their own set of SNPs, only the ones (same SNPS) tested by both companies can be compared.
    Between ancestry and FTDNA only approximately 427,000 SNPs are shared between the two companies chip versions.(FTDNA tests about 716,000, Ancestry around 671,000)

    Due to FTDNA's confidence level about distant relatives they have chosen not to show them in transfers.

    Gedmatch on the other hand is a free site set up with no extra restrictions (scientific) other then by default one needs a single segment of 7cM/500SNPs to show as a match.

    Testing companies algorithms weigh each segment based on their own standards to determine if they feel it is Identical by descent (positive/true), Identical by state (negetive/false), or can be discounted due to segments being part of a pile up region

    https://isogg.org/wiki/Identical_by_descent

    "Chr 3 Start Location 97,523,082; End Location 110,720,621; cM 8.2
    Chr 17 Start Location 7,870,502; End Location 11,119,018; cM 11.1
    Chr 19 Start Location 211,912; End location 3,103,982; cM 9.0
    Chr 19 Start Location 3,597,642; End location 8,920,808; cM 15.7
    Chr 22 Start Location 31,838,996; End location 35,701,055; cM 8.1


    How many SNPs does Gedmatch list for each of these above segments?

    Comment


    • #3
      Originally posted by osullivanbere View Post
      If someone could explain this, I would be grateful; e.g. is my match (as shown above) from GEDCOM insufficient to qualify as a match on FTDNA?

      Thanks

      Liz
      I asked FTDNA about a similar case. Here is their reply.

      "The new Ancestry file does not have a large amount of overlap with the Family Finder chip. That's what delayed us accepting the new files. Bennett did not want to impute as much data as would be necessary to match the full chip, so Elliott came up with a compromise. We would display all but speculative matches, which means those in the 4th-remote and 5th-remote cousin categories. The closer matches are not affected much by limited overlap with fewer SNPs, but the ones that are speculative anyway would be. GEDMatch, however, simply matches everything, imputing data throughout. If Kellie only matches your wife with 14.4 cM, that would be considered a speculative match and will not show up on the Family Finder match list."

      Comment


      • #4
        if your GEDmatch cousin..... hasn't loaded their data onto FTDNA you won't find them on FTDNA.

        Comment


        • #5
          And besides, GEDmatch, as far as I am aware, does NOT "impute" anything. They just have a different matching algorithm, which, among other things, doesn't impose the same overall threshold that FTDNA uses to declare a match.

          Comment


          • #6
            Hi - thanks all for the replies. But I'm still struggling

            I've set out some data to better explain my puzzle: In the list below, the first 4 people are from FTDNA, showing my top two and bottom two matches. Person 5 is from GEDmatch, and is my top GEDmatch match (she's the person missing from my FTDNA list of matches. To confirm, she is on FTDNA).

            Although the top match GEDmatch person 5 doesn't have SNPs as high as my top FTDNA people, she still has higher SNPs than both my lowest FTDNA matches - & has considerably more longish matching length blocks. So I don't grasp how FTDNA confidence levels would allow the two lowest matches (persons 3 and 4) into my list of matches but exclude person 5.

            (I don't think the smaller blocks under 4cM mean much, but I have included these just for information.)

            1. FTDNA Top match: 2-4th Cousin est
            Total number of shared cM: 62
            Longest block: 18.64 (2158 SNPs)
            Total number of shared blocks: 24
            Length/SNPs of blocks other than longest:
            Three between 3.1 - 4 cM (SNPs = 380, 369 282)
            Nineteen between 1.22 - 3cM (SNPs range between 342 - 250)

            2. FTDNA 2nd top match: 2-4th Cousin est
            Total number of shared cM: 62
            Longest block: 18.87 (1906 SNPs)
            Total number of shared blocks: 17
            Length/SNPs of blocks other than longest:
            Four between 3.1 - 4 cM (SNPs = 468, 436, 379, 290);
            Twelve between1.43 - 3cM (SNPs range between 333 - 254)

            3. FTDNA 2nd Lowest match: 3-5th Cousin est
            Total number of shared cM: 16
            Longest block: 13.53 (1228 SNPs)
            Total number of shared blocks: 2
            Length/SNPs of blocks other than longest
            One at 2.73 (SNPs = 333 SNP)

            4.FTDNA Lowest match 3-5th Cousin est
            Total number of shared cM: 13
            Longest block: 12.74 (907 SNPS)
            Total number of shared blocks: 1
            Length/SNPs of blocks other than longest:
            Nil

            5. GEDmatch top match 2-4th Cousin est
            Total number of shared cM: 59
            Longest block: 18.60 (1227 SNPs)
            Total number of shared blocks: 41
            Length/SNPs of blocks other than longest:
            Seven between 11.1 - 4cM (at 1,406, 799, 763, 592, 428, 384, 322 SNPs)
            Three between 3.1 - 4 cM (between 386 -356 SNPs)
            Thirty between 1.5 - 3cM (between 419 - 254 SNPs)

            (Again, I don't expect that the smaller blocks mean much but I've included them just for info)

            Thanks in anticipation of someone trying again to explain this to me.
            Liz

            PS: Also, prairielad, hese are the GEDmatch SNPs that correspond to the blocks I set out in my original post query.

            Chr 3 = 1,406 SNPs; cM 8.2
            Chr 17 = 763 SNPs; cM 11.1
            Chr 19 = 592 SNPs; cM 9.0
            Chr 19 = 1,045 SNPs; cM 15.7
            Chr 19 = 799 SNPs; cM 8.1


            PPS: Jim Barrett: Interesting that your reply from FTDNA says that " The closer matches are not affected much by limited overlap with fewer SNPs"

            Comment


            • #7
              Hi Liz,
              There is another factor you have to consider.

              Have you been in contact with the Gedmatch person? If you have, please ask them if they have changed their FTDNA default settings. It is possible that they have restricted email contacts which will automatically cut off certain FTDNA matchings (based on the level they choose to restrict). In such a situation, you could not see them on your FTDNA (nor you on theirs) even though lower matches who have not touched the settings would appear.

              It is worth contacting the person, since often people are unintentionally restricting the match lists without understanding this can happen.

              Comment


              • #8
                ech124 - thanks for the suggestion. I've been in contact with my cousin via Gedmatch, and she's checked and just confirmed her settings are not restricted for email contacts.

                If there are any other thoughts, I'd be grateful.
                Might it be worth me contacting FTDNA to ask about this?
                (I was considering ordering some more DNA tests, this time with FTDNA, but I've been put off a bit by this problem.)

                Comment


                • #9
                  I don't think I can help you much.

                  Programming algorithms between all the testing companies are different. Even within Gedmatch you can find different answers between one-to-many and one-to-one lookups (see the red warning on the one-to-many page).

                  I suspect that the match in question has been categorized by FTDNA as beyond a 3rd cousin and so did not make your match list.

                  If you want to pursue it, you probably have to take the familyfinder test which requires you to submit a sample. I believe they will discount the price with the unlock fee that you have paid.

                  Finally, you could also go to chromosome browser and download the CSV file there. You can then put that into a spreadsheet and sort it by chromosome number. That will give you a rough map of how your existing matches line up.

                  You can also generate a DNA map on Gedmatch by subscribing to the advanced Tier 1 utilities.
                  Last edited by ech124; 26 July 2017, 09:50 AM.

                  Comment


                  • #10
                    Hi ech124

                    Thank you again for your suggestions I already had the FTDNA chromosome browser but have now signed up to the full tier 1 utilities for GEDmatch.

                    I am still surprised me that my top GEDmatch cousin (a mile ahead of all my other GEDmatch matches) isn't listed as a match on FTDNA. This was a bit of a shock - especially as FTDNA does list lots of low 3-5th cousin matches.

                    I knew that the algorithms differ between companies, but not to this extent

                    I was considering a familyfinder test, but at the moment, this experience has put me right off.

                    Thanks again,
                    Liz

                    Comment


                    • #11
                      I explained this to you in reply #3 on July 15.

                      The problem isn't with FTDNA, it is with the down graded test at Ancestry. This causes enough gaps in your matching that FTDNA does not feel they can call the two of you a match.

                      If you are going to be "put right off" it should be with Ancestry for using a chip with fewer markers.

                      Comment


                      • #12
                        Hi Jim, I appreciate your reply. I did read your earlier reply of 15th July - but I didn't feel FTDNA's response to you was something you should rely on: To explain, they said to you that if the person only matched your wife "with 14.4cM, that would be considered a speculative match and will not show up on Family Finder match list"

                        But in my reply of 18th July I give the example of someone who is in my Family Finder match list and who shares only 12.74 cM (at 907 SNPS) with me. This doesn't seem to fit with what you were told

                        Comment


                        • #13
                          Originally posted by osullivanbere View Post
                          Hi Jim, I appreciate your reply. I did read your earlier reply of 15th July - but I didn't feel FTDNA's response to you was something you should rely on: To explain, they said to you that if the person only matched your wife "with 14.4cM, that would be considered a speculative match and will not show up on Family Finder match list"

                          But in my reply of 18th July I give the example of someone who is in my Family Finder match list and who shares only 12.74 cM (at 907 SNPS) with me. This doesn't seem to fit with what you were told
                          I believe that FTDNA's matching algorithm declares a match with a largest segment of 7.0 cM and a total shared of 20.0 cM, when counting the smaller segments or if there is one segment of 12.0 cM or larger, without need for smaller matching segments adding up to 20.0 cM. The match you cited in your post which I've bolded meets the criterion of a 12.0 or larger segment and is regarded as a match.

                          Comment


                          • #14
                            Hi Liz,
                            Your logic of saying FTDNA is wrong is faulty.

                            FTDNA has excluded your match at a 3rd cousin level.

                            Gedmatch shows the match as someone strong enough for you to consider a match.

                            Because FTDNA does show "lower ranking" people (who have nothing to do with your match), you conclude that the match was missed by FTDNA.

                            So if FTDNA has, in this case, perhaps stronger standards then Gedmatch, why can't the Gedmatch results be considered a "false positive" and they be wrong?

                            Your argument will be, look at all the Gedmatch segments. Well, I went ahead and researched it. First of all, most of the matching is on the shorter higher chromosomes which is a red flag. Some of the segments are at the very beginning of the chromosome, another red flag.

                            Finally, you failed to mention that this Match's daughter is also on Gedmatch. When we look at you compared to the daughter, all the segments are gone except for 1 segment of 11 cM. That is another red flag. Yes, the act of inheritance can lead to missing segments, but loss of that much is a little excessive and makes me question if the original segments were really true matches.

                            We will probably never know the answer, because I don't think you are inclined to test. That is fine, but please be aware that in my opinion there are enough red flags to question your conclusions about FTDNA.

                            Comment


                            • #15
                              Originally posted by osullivanbere View Post
                              Hi Jim, I appreciate your reply. I did read your earlier reply of 15th July - but I didn't feel FTDNA's response to you was something you should rely on: To explain, they said to you that if the person only matched your wife "with 14.4cM, that would be considered a speculative match and will not show up on Family Finder match list"

                              But in my reply of 18th July I give the example of someone who is in my Family Finder match list and who shares only 12.74 cM (at 907 SNPS) with me. This doesn't seem to fit with what you were told
                              You are welcome to your opinion!

                              Comment

                              Working...
                              X