Announcement

Collapse
No announcement yet.

Raw Files Available now????

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

  • #31
    Originally posted by Ann Turner View Post
    It doesn't look like my raw data download has Build 37 numbers, either. I would conjecture that maybe just new kits have B37 numbers, but there are some differences in my old and new files: 823 SNPs have been dropped, and there are a few SNPs (just 31) with different base calls per David Pike's utility

    http://www.math.mun.ca/~dapike/FF23utils/diffs.php
    Here are my differences - old on left, new on right
    Differently Reported SNPs:
    File1 File2
    1 rs4551569 66033587 GG GT
    1 rs17629008 214971951 GG GT
    2 rs17774013 46920718 CT CC
    6 rs4282439 150947 CC AC
    6 rs405059 42772654 AA AG
    6 rs228859 134292887 GG GT
    7 rs765971 127207878 TT CT
    8 rs12676149 83329290 AA AC
    8 rs4313191 138310309 CT CC
    11 rs11223278 132298355 CC AC
    13 rs7995279 28742061 AG AA
    14 rs1761038 56006600 AC AA
    14 rs4902843 70278778 AC AA
    16 rs374950 1019583 GG AG
    16 rs1559350 22617350 AG AA


    707269 SNPs appear in both files

    0.261 % (1847) of these 707269 common SNPs are NoCalls (i.e. NoCall in at least one file)

    0.002 % (15) of the 707269 SNP results differ

    0 differing SNPs are both homozygous
    9 differing SNPs are homozygous in File1 but not in File2
    6 differing SNPs are homozygous in File2 but not in File1

    Comment


    • #32
      Only two of your SNPs are in my list, so at first glance it doesn't look like a universal adjustment.

      rs12676149
      rs7995279

      Comment


      • #33
        Originally posted by Ann Turner View Post
        It doesn't look like my raw data download has Build 37 numbers, either. I would conjecture that maybe just new kits have B37 numbers, but there are some differences in my old and new files: 823 SNPs have been dropped, and there are a few SNPs (just 31) with different base calls per David Pike's utility

        http://www.math.mun.ca/~dapike/FF23utils/diffs.php

        Differently Reported SNPs:
        Old on left and new on right.

        File1 File2
        1 rs1892294 189768722 CC CT
        2 rs2042481 23810303 TT CT
        3 rs13083085 76776174 TT CT
        8 rs9329186 9244609 CC CT
        9 rs7874846 80838177 CC CT
        10 rs16906634 52041921 GG AG
        10 rs1909423 54825708 AC AA
        10 rs7084430 122430603 CT CC
        11 rs1822339 24315938 AG GG
        11 rs1784454 129132958 CT CC
        12 rs2604259 16141980 AA AG
        12 rs10842848 27108744 AG AA
        13 rs7995279 28742061 GG AG
        13 rs7318768 40331276 GT TT
        13 rs1992228 68297819 GG AG
        16 rs1559350 22617350 AG AA
        17 rs34278206 16840828 AA AG
        22 rs17808489 16190053 GG AG
        22 rs12158884 25247848 AA AG


        483169 SNPs appear in both files

        0.612 % (2957) of these 483169 common SNPs are NoCalls (i.e. NoCall in at least one file)

        0.004 % (19) of the 483169 SNP results differ

        0 differing SNPs are both homozygous
        12 differing SNPs are homozygous in File1 but not in File2
        7 differing SNPs are homozygous in File2 but not in File1

        Is everyone getting similar changes? or do you have more?

        Comment


        • #34
          Build 36 on left, 37 on right

          Differently Reported SNPs:

          File1 File2
          2 rs212761 32282792 CT TT
          2 rs17034557 46198158 CC AC
          2 rs13004652 170930644 GT GG
          3 rs2979989 127074115 GT TT
          4 rs7698581 145002291 CT TT
          4 rs1429123 148573390 CT CC
          4 rs2114005 185733572 AG GG
          5 rs16899394 23340638 TT CT
          6 rs1924690 23779859 CT TT
          7 rs10249342 5819993 GT GG
          7 rs816400 56233931 CT TT
          7 rs2077180 65902880 AG AA
          7 rs2157745 92543595 GT TT
          8 rs2717537 79861584 AG AA
          8 rs12676149 83329290 AA AC
          9 rs2309966 73305164 AC AA
          9 rs7870765 92329604 AG GG
          10 rs10905673 10256379 AG GG
          10 rs12764755 110898762 AC AA
          11 rs10768905 42941293 CT TT
          12 rs10845211 10814215 CT TT
          12 rs1168757 64880383 CT TT
          13 rs7995279 28742061 AG AA
          13 rs13378229 53691832 CT TT
          13 rs9523437 91311654 AG AA
          15 rs2572217 63849281 CT CC
          15 rs749161 73808889 AC AA
          19 rs7507174 8153444 AG AA
          19 rs1625762 42019132 AC AA
          19 rs7248492 46705059 CC CT


          707269 SNPs appear in both files

          0.303 % (2142) of these 707269 common SNPs are NoCalls (i.e. NoCall in at least one file)

          0.004 % (30) of the 707269 SNP results differ

          0 differing SNPs are both homozygous
          4 differing SNPs are homozygous in File1 but not in File2
          26 differing SNPs are homozygous in File2 but not in File1

          For the X I got this but I'm not sure it finished running the file:

          Differently Reported SNPs:

          File1 File2
          X rs1000530 78010734 CT undefined

          So if I'm understanding this correctly, I didn't 'lose' any SNPs but they changed the RESULTS at 31 locations??? You KNOW that's just wrong.. There is no reason why a conversion would change my testing results UNLESS somehow they frame shifted the results in several locations. Looks to me like they botched the files..

          Also, I have no idea why anyone would want to upload build 37 to GEDmatch just to have them convert it back to build 36... ?

          Comment


          • #35
            Originally posted by Kasandra View Post
            Build 36 on left, 37 on right

            So if I'm understanding this correctly, I didn't 'lose' any SNPs but they changed the RESULTS at 31 locations??? You KNOW that's just wrong.. There is no reason why a conversion would change my testing results UNLESS somehow they frame shifted the results in several locations. Looks to me like they botched the files..

            Also, I have no idea why anyone would want to upload build 37 to GEDmatch just to have them convert it back to build 36... ?
            All of the positions are still build 36.

            In addition to the above listed changes I had hundreds of SNPs changed to no-calls.

            With no new statements from Family Tree DNA we have no idea if these files are valid or pure garbage. And where are the files with build 37 positions?

            Comment


            • #36
              Originally posted by rod View Post
              All of the positions are still build 36.

              In addition to the above listed changes I had hundreds of SNPs changed to no-calls.

              With no new statements from Family Tree DNA we have no idea if these files are valid or pure garbage. And where are the files with build 37 positions?
              So now I have to hope that they have back ups of my original build 36 files. (I'm glad I do) but what are they using to base matches on? Is that done offline somewhere on a computer with who knows what build configuration?

              I'll be honest, what it looks like to me, is that they converted my 36 file to 37, there were errors, so they converted it back (rather than go back to the original build 36 file) and threw up their hands. It really makes me wonder on what basis are they coming up with my matches? It makes no sense at all that I got all of them back plus a bunch. I wonder if someone will need to contact the attorney general of their state to get this investigated?

              Comment


              • #37
                I have downloaded my build 37 family finder raw data files and would like to ask for advice on some of the uses they might be put to.

                The data looks - and I use the word quite properly here I believe - awesome.

                Maybe there is some explanatory stuff around somewhere?

                Comment


                • #38
                  Originally posted by Kasandra View Post
                  what it looks like to me, is that they converted my 36 file to 37, there were errors, so they converted it back (rather than go back to the original build 36 file) and threw up their hands. It really makes me wonder on what basis are they coming up with my matches? It makes no sense at all that I got all of them back plus a bunch.
                  A dead giveaway is that when the change to Build 37 went live and FTDNA saw that it was not working, they were not able to do what any competent company would have done:

                  Roll back and continue with Build 36 until they had fixed their software so it could handle the change to Build 37.

                  Instead we got this long wait with large but inexplicable changes in the sets of matches, chaotic attempts of making the raw data available, announcements of "Quality Control" and dead lines that quickly zipped by.

                  So I can believe that FTDNA "threw up their hands" in their inability to handle their own data.

                  PS. However, I do think that FTDNA are honest fools. I have on purpose made rather provocative and very negative postings here on the FTDNA controlled forum and unlike other companies, they have not removed these postings. That speaks in their favour.
                  Last edited by Lklundin; 13th April 2013, 05:06 AM. Reason: PS.

                  Comment


                  • #39
                    Originally posted by Lklundin View Post
                    A dead giveaway is that when the change to Build 37 went live and FTDNA saw that it was not working, they were not able to do what any competent company would have done:

                    Roll back and continue with Build 36 until they had fixed their software so it could handle the change to Build 37.

                    Instead we got this long wait with large but inexplicable changes in the sets of matches, chaotic attempts of making the raw data available, announcements of "Quality Control" and dead lines that quickly zipped by.

                    So I can believe that FTDNA "threw up their hands" in their inability to handle their own data.

                    PS. However, I do think that FTDNA are honest fools. I have on purpose made rather provocative and very negative postings here on the FTDNA controlled forum and unlike other companies, they have not removed these postings. That speaks in their favour.
                    The bottom line is we do not know what happened. As NYMark and I have indicated there has been no explanation of what kind of error was present in the earlier algorithm for Build 37. Was it simply a reduction in matches or was there a different kind of identifiable error independent of the reduction. A reduction in matches could result from a change in priorities rather than an error.
                    Last edited by josh w.; 13th April 2013, 08:06 AM.

                    Comment


                    • #40
                      Originally posted by josh w. View Post
                      The bottom line is we do not know what happened. As NYMark and I have indicated there has been no explanation of what kind of error was present in the earlier algorithm for Build 37. Was it simply a reduction in matches or was there a different kind of identifiable error independent of the reduction. A reduction in matches could result from a change in priorities rather than an error.
                      My question is did they know in advance that there would be a reduction in matches. There can be no answer to this question without an explanation from FTDNA.
                      Last edited by josh w.; 13th April 2013, 08:31 AM.

                      Comment


                      • #41
                        for royfarnol

                        One thing you can do with the raw data is upload it to Gedmatch.com and use the various utilities there to analyze it. Unfortunately the site has been down for maintenance for a couple of a days--we probably overloaded it!

                        In the meantime, I recommend reading the informative posts at http://dna-explained.com/ if you haven't already (or re-reading them--I read everything a month or so ago and once I got the raw data I needed to re-read some and refresh my memory).

                        Comment


                        • #42
                          @ khuebner

                          Much obliged, thanks.

                          Comment

                          Working...
                          X