Page 7 of 7 FirstFirst ... 567
Results 61 to 67 of 67

Thread: The Insoluble Dilemma of Commercial Haplogroup Predictors used by FTDNA/Genographic

  1. #61
    Banned
    Posts
    231
    Sex
    Location
    Canada
    Ethnicity
    Punjabi Sidhu Jatt
    Nationality
    Canadian
    Y-DNA (P)
    F
    mtDNA (M)
    R30B1

    Canada Pakistan Finland
    Quote Originally Posted by Amerijoe View Post
    I my opinion and with on going personal interaction, testing at The Genographic Project 2.0+ is no better than spinning a roulette wheel.

    Geno refuses to acknowledge any errors in it's results and FTDNA keep saying it's not our fault and the customer is left holding the bag. I blame both companies, Geno, for it's broke down chip and FTDNA for knowingly producing erratic results from said chip.

    To separate itself from this debacle, one must first openingly admit the problem exists. FTDNA should sub-contact the work out, thereby separating it's association with Geno and avoid further tainting from this matter.
    Quote Originally Posted by Afshar View Post
    This is a paper that has studied microarray chips http://www.pnas.org/content/99/20/12975.full .
    The point is that its mainly used for economical reasons otherwise it wouldnt be able to provide these tests to the public at an affordable price. Its impossible to provide these test at an equal price with either sanger sequencing or next gen sequencing (with the same accuracy).
    Quote Originally Posted by wombatofthenorth View Post
    You also fail to mention that BigY costs hundreds of dollars more and tests only a single thing and uses next gen testing. Why do you keep comparing it to all-purpose chips?

    You seem to think Geno can not do a single thing correct ever and that FTDNA has zero problems of their own ever (then how come they had so many bugs in FamilyFinder they took months to get any matches to appear at all? granted they did take me of me well in the end).

    Yes, some annoying things with Geno 2.0 NG test such as poor sampling for V mtDNA and not really all that much results for most compared to the old test and they did over sell the upgrade potential for haplogroups. And yeah for some they have even made errors (although even more often the haplo errors were on FTDNA transfer but not also on original site, although that has happened too).

    But the way you hate on them 100% seems a bit over the top. Did they kick your dog?
    Quote Originally Posted by miiser View Post
    Every chip of this size is going to have some SNPs that don't work properly, giving both false positives and false negatives. There needs to be a verification process to establish which reads are legitimate and which are garbage. And then you just ignore (don't report to the customer) any SNPs that don't give reliable results. My impression, based on this thread and other comments I've seen elsewhere, is that this new chip was rushed to market without going through any such verification process. They are essentially doing open beta testing to figure out which SNPs need to be filtered out and thrown away.
    Quote Originally Posted by Saetro View Post
    This is very much at the heart of what is going on.
    Ideally, the customer should be properly informed about the outcome before purchasing a test.
    Unfortunately emotion is brought into the mix by both individuals wanting to do the test as well as testing labs touting for business.
    And this is a complex area.
    I regularly spend an hour on the phone with one or other member of my family history society just going through what they hope to get and what test will best do it for them. In the end, it is often a matter of hoping that some distant cousin has also tested, or enough people with the same haplogroup.
    People like Armando have been very patient with helping me, so my helping others is only giving forwards.
    Thank you to them all, and to all the bloggers in this area.
    And to this forum that enables us to help each other.

    While so many people make their choice on the basis of 1)whichever test lab shouts loudest, and 2)the lowest price, in what is a complex field; it is inevitable that some individuals will continue to be led to expectations that are unrealistic.
    I strongly advise potential testers to inform themselves in whatever aspect they want to research, to read what the tester says in FAQs and if necessary, to ask the tester what they can do for them.
    Quote Originally Posted by vettor View Post
    I got more SNP both positive and negative from Geno 2 than any other company ...............I got more help from Geno 2 than anyone else............I got the first truly accurate ydna and mtdna than anyone else ...................and the worst is 23andme ......
    In 23andme v3 I had T marker ( which I am ) since november 2015 they are using v4 format and have placed me under F marker , because my v3 SNP's are no longer in their v4 system. Utterly useless company
    Quote Originally Posted by ArmandoR1b View Post
    That was the first version of Geno 2.0 and not Geno 2.0 NextGen. There is a higher percentage of people that get an incorrect haplogroup or subclade with Geno 2.0 NextGen than did with the first version of Geno 2.0 or with 23andme. I haven't seen you post that you have had a BigY test so you could see that with a YFull or FGC analysis it would give you more positive and negatives than any other company. The FTDNA FMS tests more of the mtDNA than any other company except a much more expensive FGC test. With a BigY test you would probably create a new subclade at https://www.yfull.com/tree/T-CTS8489/


    23andme tests a far too low of a number of Y-DNA SNPs. There are less than 1,000 Y-DNA SNPs without no-calls. 884 in one of the files that I looked at. The first version of Geno 2.0 had about 12,000. Geno 2.0 NextGen probably has the same number or even less if all of the extra false positives and false negatives are removed. BigY has almost 36,000 SNPs in the VCF file the actual number of usable SNPs will be far less but a lot more than Geno 2.0.
    Hi Folks,

    As you know that the main purpose of this thread was to discuss the pitfalls/weaknesses in the overall testing platform that the commercial DNA testing companies use to assign haplogroups to their clients and share with each other our experiences. In context with my post no 1, I am going to eloborate with examples the weaknesses of the system. The called SNPS (results of your DNA Analysis) are compared to the DNA reference tree mainly ISOGG 2013 and based on the where these SNPS fit on the tree – a Haplogroup is assigned. Two main type of strategies are used in assigning haplogroups :

    • Vertical Alignment of SNPS to the reference tree

    • Horizontal alignment

    Combination of both

    For the Sake of discussion lets say that we have three clients who tested their DNA with a commerical DNA testing company. Their DNA was analyzed by using the chip technology. For the simplicity purpose, let us consider an imaginary haplotree with only 6 Haplgroups i.e. Haplogroup A, Haplogroup B, Haplogroup C, Haplogroup E, Haplogroup D and Haplogroup G

    Person 1 is tested positive for the following SNPS : M90, M92, M94, M95 and he fits well in to the green path of the Haplotree Diagram and hence he is assigned Haplogroup A2B1. This is perfect example of Vertical alignment of SNPS on the tree since all the SNPS follow a vertical pattern as drawn in green in the sketch.

    Person 2
    is tested positive for the following SNPS: P266, P230, P90 and P60 and he fits well in to the Purple path of the Haplogroup tree and therefore he is assigned haplgroup G. This is is perfect example of horizontal alignment of SNPs since the person was not tested positive for any other snps then the ones in purple.

    Person 3 is tested positive for SNPs that fall erratic on the reference tree. He is positive for the SNPS L21, L22, L23, L25, L30, M200 and M201 (SNPS in red). As you would notice that there is no clear pattern of SNPs that can put him on the tree with confidence. Now the most cautious/conservative approach would be to assign him B ( where he has the most red SNPS) but the commercial algorithms used by various companies fall short of recognizing this dilemma and would wrongfully assign him a :

    C : based on SNP L25
    D2A : based on SNP M201 or
    E1b basied on SNP L30



    Most clients would not question the results and would not know the great amount of uncertainty that is involved in assigning haplogroups. You are on your own if your case is similar to person 3 and would not get much support from the vendors. It will be all guess work and less science!
    Last edited by jatt2016; 09-02-2016 at 07:59 PM.

  2. #62
    Banned
    Posts
    231
    Sex
    Location
    Canada
    Ethnicity
    Punjabi Sidhu Jatt
    Nationality
    Canadian
    Y-DNA (P)
    F
    mtDNA (M)
    R30B1

    Canada Pakistan Finland
    The Below is real life example of how my raw data ( from FTDNA/NatGeo) fits in various haplogroups.


    I am a F if you consider the following tested positive SNPS below :


    I am an H if you consider one tested positive SNP under H below:



    I am a J if you consider the following tested positive SNP below:



    and an R if you consider the following tested positive SNPs below:



    As, you would notice that these SNPs are erractic and no clear inference can be derived from these results.

    In situations likes these most vendors take advantage of the client's ignorance or lack of knowledge in this field and erroneously assign a haplogroups without explaining the uncertainty factor to the client.

    FTDNA and Genographic after trial and error, from J to R and then to H ( All of these can be questioned caz these are erratically spread on the tree). If I am placed on H, then it means that Geno/FTDNA has ignore the following positive SNPS that would have otherwise placed me in J or R

    23andme
    was very cautious in assigning me a haplogroup and placed me in F. I discuss with them my case that how I was previously assigned a J then an R and finally an H by other vendors and 23andme explanation was that I do not have enough positive SNPS under H, J and R and that they assigned me F where I had the most positive SNPs. To be assigned an H, R or J I should show a pattern of SNPs and not just some isolated SNPs. 23andMe will assign you the haplogroup that you are the most similar to in regards to the reference tree.

    Personally, I think that 23andme's explanation is more logical and scientific whereas, FTDNA and Genographic were just shooting arrows in the sky - guess game and taking advantage of client's ignorance and lack of understanding of the haplo assignment process.




    The below is a list of tested positive SNPS that were ignored by FTDNA/Genographic when they assigned me an H


    Code:
    BY1357
    BY136
    BY1551
    BY1585
    BY20
    BY2510
    BY65
    BY763
    BY876
    CTS10031
    CTS10147
    CTS10188
    CTS10229
    CTS103
    CTS10300
    CTS10428
    CTS10433
    CTS10448
    CTS10552
    CTS10572
    CTS10615
    CTS10648
    CTS1066
    CTS10723
    CTS10745
    CTS10761
    CTS10821
    CTS10847
    CTS11012
    CTS11041
    CTS11071
    CTS11088
    CTS11126
    CTS11148
    CTS11190
    CTS11261
    CTS11354
    CTS114
    CTS1141
    CTS1148
    CTS11503
    CTS11530
    CTS1164
    CTS11651
    CTS11731
    CTS11759
    CTS11816
    CTS11907
    CTS11949
    CTS1216
    CTS12578
    CTS12856
    CTS12933
    CTS12934
    CTS12948
    CTS12950
    CTS1340
    CTS1353
    CTS1413
    CTS1758
    CTS1806
    CTS202
    CTS2067
    CTS210
    CTS2275
    CTS2284
    CTS2289
    CTS2670
    CTS2800
    CTS305
    CTS316
    CTS3202
    CTS3268
    CTS3366
    CTS3403
    CTS3466
    CTS347
    CTS3519
    CTS352
    CTS3543
    CTS3654
    CTS3700
    CTS3802
    CTS3868
    CTS4053
    CTS4086
    CTS4094
    CTS4137
    CTS4178
    CTS4209
    CTS4235
    CTS4326
    CTS4443
    CTS4463
    CTS4608
    CTS4643
    CTS4715
    CTS4853
    CTS5052
    CTS5152
    CTS5156
    CTS5304
    CTS5334
    CTS5359
    CTS5370
    CTS585
    CTS5860
    CTS5933
    CTS5939
    CTS597
    CTS5998
    CTS6062
    CTS6063
    CTS6117
    CTS6285
    CTS6352
    CTS636
    CTS6403
    CTS6438
    CTS6447
    CTS6449
    CTS6468
    CTS6481
    CTS6506
    CTS656
    CTS6704
    CTS6848
    CTS687
    CTS6918
    CTS6967
    CTS7022
    CTS7147
    CTS7186
    CTS7227
    CTS7266
    CTS7275
    CTS7299
    CTS730
    CTS7317
    CTS7335
    CTS7451
    CTS7611
    CTS7626
    CTS7767
    CTS7810
    CTS7929
    CTS8126
    CTS8203
    CTS8216
    CTS8300
    CTS8440
    CTS8506
    CTS8521
    CTS860
    CTS8645
    CTS8723
    CTS8815
    CTS9056
    CTS9154
    CTS9325
    CTS9387
    CTS9471
    CTS9518
    CTS9525
    CTS9539
    CTS9558
    CTS97
    CTS9894
    CTS9900
    CTS9925
    CTS9975
    DF109
    F1013
    F1024
    F1055
    F1127
    F1207
    F1291
    F1302
    F1329
    F1412
    F1442
    F1474
    F1540
    F1567
    F1703
    F1704
    F1712
    F1714
    F1753
    F1767
    F1796
    F1848
    F1874
    F188
    F1899
    F1924
    F1946
    F1956
    F2018
    F2048
    F2142
    F2155
    F2165
    F2224
    F2236
    F2304
    F2370
    F2402
    F2516
    F2558
    F2587
    F2654
    F2688
    F2724
    F2742
    F2753
    F2767
    F2837
    F287
    F2873
    F2947
    F2961
    F2985
    F2990
    F3048
    F3056
    F3111
    F313
    F3136
    F3195
    F3335
    F3368
    F3402
    F3422
    F3444
    F3491
    F3556
    F3582
    F3595
    F3625
    F3632
    F3692
    F3697
    F371
    F3735
    F3806
    F3869
    F3892
    F3907
    F3949
    F3956
    F3994
    F4003
    F4042
    F4111
    F4176
    F4188
    F4244
    F4257
    F4282
    F4338
    F489
    F492
    F526
    F557
    F565
    F646
    F716
    F719
    F736
    F741
    F836
    F837
    F862
    F896
    F995
    FGC11678
    FGC11897
    FGC12120
    FGC12627
    FGC13364
    FGC16362
    FGC16683
    FGC1721
    FGC20866
    FGC20874
    FGC28639
    FGC29572
    FGC3718
    FGC6948
    K257
    K386
    L1186
    L253
    L294
    L325
    L387
    L442
    L484
    L492
    L498
    L503
    L556
    L782
    L817
    L840
    L99 (Geno 2.0 results are erratic for this SNP)
    M116
    M129
    M171
    M238
    M251
    M262
    M275
    M288
    M367
    M9574
    M9588
    N4
    P102
    P117
    P118
    P268
    P269
    P289
    P52
    P59
    P84
    P91
    PAGES00048
    PAGES00049
    PAGES00081
    PAGES00101
    PAGES00105
    PF1015
    PF1026
    PF1031
    PF1085
    PF1097
    PF1141
    PF1147
    PF1152
    PF1164
    PF1169
    PF121
    PF1232
    PF1252
    PF1269
    PF1270
    PF1279
    PF1283
    PF133
    PF1368
    PF147
    PF1562
    PF1577
    PF1587
    PF22
    PF228
    PF2334
    PF2481
    PF2495
    PF2579
    PF2593
    PF2611
    PF2624
    PF2635
    PF2643
    PF2745
    PF276
    PF288
    PF293
    PF3051
    PF3086
    PF3107 (Geno 2.0 results are erratic for this SNP)
    PF3220
    PF328
    PF3298
    PF3320
    PF3823
    PF3890
    PF3964
    PF3986
    PF4105
    PF4246
    PF4533
    PF4573
    PF4576
    PF4589
    PF4592
    PF4647
    PF4720
    PF4957
    PF5014
    PF5122
    PF535
    PF5381
    PF5735
    PF5787
    PF5792
    PF601 (mis-called by FTDNA?)
    PF6063
    PF625
    PF6275
    PF6286
    PF6300
    PF6314
    PF6369
    PF644
    PF6673
    PF668
    PF671
    PF679
    PF6864
    PF698
    PF700
    PF7067
    PF7328
    PF7341
    PF7374
    PF7527
    PF7557
    PF794
    PF806
    PF829
    PF888
    PF907
    PF93
    PF946
    PF966
    PF968
    PF971
    PF997
    PR805
    S7123
    SK1247
    SK1410
    V186
    V205
    V21
    V216
    V227
    V79
    V90
    V94
    Y1038
    Y1049
    Y1083
    Y1113
    Y1122
    Y32
    Y4864
    Y5272
    Y5305
    Y5759
    Y6181
    Y763
    YP321
    YSC0000081
    YSC0000150
    YSC0000216
    YSC0000292
    YSC0001071
    Z11180
    Z1149
    Z14303
    Z1456
    Z1476
    Z1483
    Z1504
    Z1518
    Z1589
    Z1593
    Z1616
    Z17153
    Z18140
    Z1835
    Z1975
    Z315
    Z3723
    Z39
    Z477
    Z625
    Z767
    Z966
    ZS1727
    RS12116413
    RS16980396
    RS16980499
    RS17842387
    RS2032652
    RS2075640
    RS35249440
    RS35407486
    RS7067237
    RS7892924
    RS7893044
    RS9306845
    RS9306848
    RS9785670
    RS9785743
    RS9785853
    RS9785905
    RS9786082
    RS9786181
    RS9786247
    RS9786290
    RS9786325
    RS9786774
    BY1142
    Last edited by jatt2016; 09-03-2016 at 04:39 AM.

  3. The Following User Says Thank You to jatt2016 For This Useful Post:

     vettor (09-08-2016)

  4. #63
    Banned
    Posts
    231
    Sex
    Location
    Canada
    Ethnicity
    Punjabi Sidhu Jatt
    Nationality
    Canadian
    Y-DNA (P)
    F
    mtDNA (M)
    R30B1

    Canada Pakistan Finland
    My Conclusion :

    The commerical testing falls short of accuracy when it comes to assigning haplgroups to non european populations mainly due to lack of adequate DNA samples from these groups. The Y DNA tree is incomplete and does not capture all ethnic groups. Many of the clients are assigned wrong/inaccurate haplgroups without them being aware of it.

    A complete and accurate Y-Haplo tree that fits all populations/ethnic groups looks undoable with the technology and resources available at this point in time.
    Last edited by jatt2016; 09-03-2016 at 01:46 AM.

  5. The Following User Says Thank You to jatt2016 For This Useful Post:

     Amerijoe (09-08-2016)

  6. #64
    Registered Users
    Posts
    4,324
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    The 2001 paper referenced criticizing a MicroARRAY technology does not apply and should not be cited on this topic unless some how there is an historical reference. It was received May 2001 so much of the work for it may have been done in 2000.

    I don't know what or how the National Geographic Society's (they go by NGS too) Geno 2.0 NG is done but I absolutely agree it is full of errors.

    I can say that SNP Packs are running on a MassARRAY technology (not "Micro"), and this MassARRAY technology is much more modern than 2000 and pre-2000 stuff. FTDNA also has a vetting process on SNPs for the Packs. Regardless, sometimes project admins want marginal SNPs in packs to see if they will work.
    Last edited by TigerMW; 09-08-2016 at 05:50 PM.

  7. The Following 3 Users Say Thank You to TigerMW For This Useful Post:

     Amerijoe (09-08-2016),  ArmandoR1b (09-08-2016),  jatt2016 (09-08-2016)

  8. #65
    Registered Users
    Posts
    4,324
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by jatt2016 View Post
    My Conclusion :

    The commerical testing falls short of accuracy when it comes to assigning haplgroups to non european populations mainly due to lack of adequate DNA samples from these groups. The Y DNA tree is incomplete and does not capture all ethnic groups. .... .
    The solution is get more testers from those regions and ethnic groups. Someone has to pay. A leader in this is your favorite - the National Geographic Society and their National Genographic Project. I agree there are problems but you should probably work with them to blacklist SNPs as is appropriate and try to redirect them to populations you are interested in, or help them with donations for the National Genographic Project. My guess is you can write off donations on your taxes to the project.
    Last edited by TigerMW; 09-08-2016 at 05:50 PM.

  9. The Following 4 Users Say Thank You to TigerMW For This Useful Post:

     Amerijoe (09-08-2016),  ArmandoR1b (09-08-2016),  jatt2016 (09-08-2016),  wombatofthenorth (09-15-2016)

  10. #66
    Registered Users
    Posts
    961
    Sex
    Location
    Sarasota, Fl. USA
    Ethnicity
    Scot
    Nationality
    American Naturalized
    Y-DNA (P)
    YP6291 R1a1a1b2a2a*
    mtDNA (M)
    J1c23

    United States of America Scotland
    Quote Originally Posted by Mikewww View Post
    The solution is get more testers from those regions and ethnic groups. Someone has to pay. A leader in this is your favorite - the National Geographic Society and their National Genographic Project. I agree there are problems but you should probably work with them to blacklist SNPs as is appropriate and try to redirect them to populations you are interested in, or help them with donations for the National Genographic Project. My guess is you can write off donations on your taxes to the project.
    I also am having an ongoing dialogue about an erroneous mtdna result. I'm quite confident it will be corrected even as frustrating a process on responsibility as it has become. Pride supersedes reason.

    I totally agree with supporting science. So many grants have been established to push results in the direction of those with the most to gain. The little guy, me and you for an example, are used as tools to accomplish this mean. Sorry for the digress, but it saddens me to witness such a rise in the publication inferior science studies. Now, that money you were going to donate to the study of climate change on the sex habits of the dung nettle, donate it to a just cause or NG or find someone through your group administrated who needs a helping hand. Being the helping hand gave me the best feeling.

  11. The Following 2 Users Say Thank You to Amerijoe For This Useful Post:

     jatt2016 (09-08-2016),  TigerMW (09-09-2016)

  12. #67
    Gold Class Member
    Posts
    516
    Sex
    Location
    England
    Ethnicity
    Northern English
    Nationality
    British
    Y-DNA (P)
    Brother I2-M26 I2a1a
    mtDNA (M)
    H4a1a1a

    Thank you Jatt for raising this issue, it has highlighted numerous aspects of concern that the naive customer (like myself) is totally unaware of.

  13. The Following User Says Thank You to Judith For This Useful Post:

     jatt2016 (10-05-2016)

Page 7 of 7 FirstFirst ... 567

Similar Threads

  1. Replies: 50
    Last Post: 02-26-2021, 09:46 PM
  2. Replies: 1
    Last Post: 03-06-2018, 11:13 PM
  3. Replies: 1
    Last Post: 01-06-2018, 05:37 PM
  4. A moral dilemma.
    By Grossvater in forum Genealogy
    Replies: 14
    Last Post: 08-08-2015, 07:09 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •