Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 28

Thread: FTDNA vs Yfull "private" SNP confusion?

  1. #11
    Gold Class Member
    Posts
    347
    Sex
    Location
    Ohio
    Ethnicity
    German
    Nationality
    Galactic Empire
    Y-DNA (P)
    R-Z324, Z5055, L188+
    mtDNA (M)
    H5

    Quote Originally Posted by lgmayka View Post
    Trust is in the eye of the beholder. YFull prefers to apply strict academically defensible criteria, because one of their goals is to provide a complete haplotree of the human race that academia can trust.
    Considering the omissions it can hardly be considered anywhere near complete since it is based upon contributions from individuals who PAY for their service. Remember that the academically rigorous criteria is technology limited. As we get some of the results from the much longer read technologies this year it will become more obvious how limited the current criteria area.

  2. #12
    Gold Class Member
    Posts
    2,718
    Sex

    Quote Originally Posted by Cofgene View Post
    Considering the omissions it can hardly be considered anywhere near complete since it is based upon contributions from individuals who PAY for their service.
    YFull also includes academic samples if the available full-Y data has sufficient coverage. But this presumably has to be done "on the side," since no one is paying for it. You are correct that money does not magically fall from the sky (or from government treasuries) to pay for this kind of work.

    My use of the term complete was primarily meant to acknowledge YFull's coverage of the entire Y-tree, from A00 to T. Where else on earth can one find a complete and correct analysis of five N-P189 examples? Academia barely recognizes the existence of the clade at all (and grossly misreported its ethnicity).
    Quote Originally Posted by Cofgene View Post
    Remember that the academically rigorous criteria is technology limited. As we get some of the results from the much longer read technologies this year it will become more obvious how limited the current criteria area.
    The original poster asked about getting Yseq to offer individual SNP tests for some SNPs. Yseq's criteria are even stricter than YFull's.
    Last edited by lgmayka; 02-07-2017 at 01:09 AM.

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

     Amerijoe (02-07-2017),  JMcB (02-07-2017),  vettor (02-07-2017)

  4. #13
    Registered Users
    Posts
    4,410
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by A.Morup View Post
    ....
    Now - who do I have to believe - FTDNA or Yfull - and do I have 22271xxx as a private SNP or not?
    Anders
    Anders, the FTDNA automated calling system is a rough tool. To get an expert opinion out of FTDNA you have to get a project admin to email an inquiry Michael Sager. Generally the request is along the lines of adding a branch to the haplotree or updating your haplogroup label because you think it is wrong. In the process you can have him check for SNPs that are questionable and see if he there are any other folks in the database that are derived. Sager's job is to keep the haplotree and haplogroups correctly updated (the haplogroup labels are based on the haplotree). If you have to couch your request / inquiry in terms of that.

    There are some variants in the DYZ19 region that labs will sometimes accept. Expert opinions vary and the caveat is always "so far". The statement is everything looks good "so far" which means something might be acceptable until its not.

    There is a subtle terminology thing here that you may not have noticed. The Next Gen Sequencing VCF files and output involve the calls of "variants". The mutations you are looking are called variants on FTDNA's screens. They are not calling them valid SNPs, just variants... which they are. They are variants from the reference model. Whether or not they are stable and good for documentation on the tree is a different story.

    I might as well hit a pet peeve since we are on it - terminology. Newbies probably want to tune out here.

    The industry has gotten us into their "speak" which is SNPs are whatever they say they are. This is because the original concept blew up. Originally only mutations that were deemed "Unique Event Polymorphisms" (UEPs) were considered for the tree. What blew up is that with 59 Million or so base pairs on the Y chromosome and about 3.5 Billion (with a "B") males in the world its obvious that we are going to have polymorphisms that are NOT unique life-cycle of mankind events.

    Hence "Single Event Polymorphisms" (SNPs) became the replacement term. This technically means a "single" change event. Today, we call things such as insertion-deletion events (indels) SNPs. They aren't but SNP is spoke of as a UEP.

    I guess I'm trying point out we have to take all of this stuff with a grain of salt and no test provider, interpreter, institution is clean, clean.
    Last edited by TigerMW; 02-07-2017 at 02:45 PM.

  5. The Following 2 Users Say Thank You to TigerMW For This Useful Post:

     dp (02-08-2017),  JMcB (02-07-2017)

  6. #14
    Registered Users
    Posts
    714
    Sex
    Location
    Texas
    Ethnicity
    English, Irish, German
    Nationality
    US
    Y-DNA (P)
    L21>L226>FGC5639

    England Germany Netherlands France Ireland Switzerland
    There are various methods and sources to analyze NGS data:

    1) Haplogroup admins who review BAM files are usually the highest quality that you can get. They analyze YSNPs in unstable areas and find many consistent results that make good branches. They look many more NGS tests and see marginal calls on some tests that can be compensated by strong calls by other tests. These reveal even more valid branches. There are not many admins that do this kind of analysis, but if you belong to one of these haplogroup projects, you send them a link to your BAM file for maximum analysis. Dennis Wright spoils me as he does the heavy lifting for this analysis. Jame Irvine does the same for L555.

    2) YFULL and FGC offer similar analysis services. FGC reports more gray area calls that could be valid branches but this requires work on your part. YFULL is more conservative in their calls - good and bad. I now prefer YFULL over FGC since they have a viable database for R1a and I haplogroups.

    3) BigTree only uses VCF files but their coverage of R1b is so superior to YFULL that you always use Mike's procedure to upload your VCF files to Ale'x BigTree database which is no charge and covers around 90 % of all R1b NGS tests.

    4) FTDNA's haplotree includes all of the Big Y tests and is a significant resource for summaries of haplotrees. But it lacks the vast majority of private YSNPs since they do not post these results. Their haplotree constantly has numerous errors but the FTDNA only errors will be addressed by Michael Sager. FTDNA YSNP database is very useful as well but has a lot of issues - few private YSNPs, lots of bloat from older non-relevant tests, many admins default to private (project only), so their YSNP database is missing a substantial number of YSNP tests, many projects are private - stuck in the dark ages or privacy phobic, also remember to look at terminal YSNPs in the YSTR reports as five percent of YSNP testing is only found there.

    5) ISOGG's haplotree is quickly becoming less important as their criteria is too labor intensive to prove branches and they exclude on the exciting genealogical YSNP which probably 25 % of the branches and growing.

    6) YSEQ's tool known as the ISOGG Ybrowser is useful for drilling down for locations, approximate areas where YSNPs will reside, synonymous labels, etc.

    7) FGC has a haplotree for their tests and those that they analyze - this is pretty limited but worth a look at occasionally. I have not looked at it in a while and I tested at YElite 1.0 so it is a minor player compared to YFULL.

    8) YSEQ YSNP files where around five or ten percent of branches are revealed. NGS databases do not address this issue and ID translation remains an issue.

    9) FTDNA SNP results in YSNP reports are just starting to reveal many branches when these SNP packs include private YSNPs and branch equivalents. NGS databases do not address this issue. Also FTDNA are listing all of these private YSNPs included in SNP packs as branches. L226 went from 25 valid branches to 75 branches overnight - 50 branches are not real branches but are private YSNPs. So for L226, we went from a modest 10 % error rate to a 75 % error rate overnight. So FTDNA's haplotree will be becoming less reliable as these SNP packs roll out.

    These are nine good sources for YSNP information. We really need a consolidation of these databases and standardization of their output. Also, we need these organizations to provide CSV export functionality from these sources as well, so that we can review this information offline and build our database to properly provide analysis for the genetic genealogy community.
    Last edited by RobertCasey; 02-07-2017 at 03:41 PM.

  7. The Following 4 Users Say Thank You to RobertCasey For This Useful Post:

     A.Morup (02-07-2017),  dp (02-08-2017),  JMcB (02-07-2017),  lgmayka (02-07-2017)

  8. #15
    Registered Users
    Posts
    113
    Sex
    Location
    Copenhagen, Denmark
    Ethnicity
    96% Danish
    Nationality
    Danish
    Y-DNA (P)
    R-BY3604
    mtDNA (M)
    J1c4

    Denmark Germany
    Thanks to all contributors!

    My question has been answered - I now know why my 4th SNP was not to be found at Yfull.
    I am happy with my 3 Yfull "stable" SNPs and will order single SNP tests for those three SNPs at Yseq and start testing my parallel male lines, when I get the chance.

    I am already at Alex' Ytree - I really love his presentation of the P312 haplotree!
    Do you know that Ytree is now including U106?

    On another note, my close match (4 at Y111) and I have just convinced our common Y111 match (at genetic distance 10) to take a Big Y, which will hopefully divide our 25 common SNPs into two smaller groups.
    I am also in contact with an other potential BY3604 Dane, who has only done Y12 but has SNP tested to Z35 and he is now testing for BY3604 at Yseq. If he is BY3604+, we will test for Z257 and if he is Z257- maybe then for some of our common private SNPs (he is not interested in doing BigY). My local haplotree is expanding!

    Thanks to all!
    Anders

  9. The Following 2 Users Say Thank You to A.Morup For This Useful Post:

     dp (02-08-2017),  JMcB (02-07-2017)

  10. #16
    Registered Users
    Posts
    6,147
    Sex
    Location
    Australia
    Ethnicity
    Italian Alpine
    Nationality
    Australian and Italian
    Y-DNA (P)
    T1a2b- SK1480
    mtDNA (M)
    H95a
    Y-DNA (M)
    R1b-S8172
    mtDNA (P)
    T2b17

    Australia Italy Veneto Friuli Italy Trentino Alto Adige Italy Ladinia Austria Tirol
    Quote Originally Posted by A.Morup View Post
    Thanks to Jdean and Igmayka!

    I also prefer "strict academically defensible criteria"

    Would testing 22271xxx at Yseq.net on myself confirm this SNP - or should I just ignore it and stick to my 3 "good" private SNPs?

    I know that my close match and I have a common ancestor around the year 1600 - we have not found him yet, but we are getting closer. I have contacted decendants from other male lines in my known male line and asked them to test (at my expense) at Yseq for these 3 or 4 private SNPs, so I can find out at what generation they occured and thereby making my private SNPs really private! Any thoughts?
    I have already got the first positive answer!

    Thanks
    Anders
    As discussed recently with Yfull team, ..........you can loose some or all of your known "private SNPs" as they discover and analyse others with the same "private SNPs" .


    My Path = ( K-M9+, LT-P326+, T-M184+, L490+, M70+, PF5664+, L131+, L446+, CTS933+, CTS3767+, CTS8862+, Z19945+, BY143483+ )


    Grandfather via paternal grandmother = I1-CTS6397 yDna
    Great grandmother paternal side = T1a1e mtDna
    Son's mtDna = K1a4p

  11. #17
    Registered Users
    Posts
    113
    Sex
    Location
    Copenhagen, Denmark
    Ethnicity
    96% Danish
    Nationality
    Danish
    Y-DNA (P)
    R-BY3604
    mtDNA (M)
    J1c4

    Denmark Germany
    Of course - private means simply "until now only found in you".
    If I tested my brother, I would most probably end up having no "private" SNPs, but then all my SNPs would be on the public Yfull tree.

    Quote from Yfull FAQ:
    A new SNP is placed on the YFull YTree at the time YFull identifies a specific mutation that is phylogenetically consistent and shared by two or more people, even if they are closely related.
    An added SNP may previously have been defined by YFull as private or as novel and private.


    I think that makes good sense!
    Anders

  12. #18
    Registered Users
    Posts
    4,410
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by RobertCasey View Post
    ....
    5) ISOGG's haplotree is quickly becoming less important as their criteria is too labor intensive to prove branches and they exclude on the exciting genealogical YSNP which probably 25 % of the branches and growing.
    ...
    I think they/we need a redo. This is probably being worked on but everyone else seems to have conquered the infrastructure problems already.

  13. #19
    Registered Users
    Posts
    113
    Sex
    Location
    Copenhagen, Denmark
    Ethnicity
    96% Danish
    Nationality
    Danish
    Y-DNA (P)
    R-BY3604
    mtDNA (M)
    J1c4

    Denmark Germany
    Sorry guys, but I will need your advice again!

    The thing is that Yfull just finalized the results of my very close match and my 28 "novel" SNPs, were reduced to 3 as expected, but not the ones I expected!
    Lets take it from the start:

    FTDNA originally gave me:
    16604xxx
    22271xxx (turned out to be in the DYZ19 zone)
    22974xxx
    7511xxx

    On Yfull, I could only find (among my 28 "novel" SNPs):
    16604xxx - "Best quality" (0 probality of error)
    22974xxx - "Acceptable quality" (0.072 probality of error)
    7511xxx - "Acceptable quality" (0.013 probality of error)

    Now after finalizing my match, Yfull gives me:
    16604xxx
    3882xxx - "Acceptable quality" (0.0 probality of error - homologous, only 4 reads)
    7511xxx

    So where did 22974xxx go?
    Maybe FTDNA has a higher setting for minimum "reads" and my match DID have 22974xxx, but with less reads than FTDNA accepts?
    And that would also explain why 3882xxx shows up now.

    It is a bit disturbing - meanwhile, I have contacted and got surprisingly positive response from two distant male line relatives - one branching of my line in the year 1790 and the other in the year 1833. One kit is already at Yseq and the other is on the way. But which SNPs should I test them for? The original idea was to see which of my private SNPs were created in what generation, but now I have 5 SNPs in play - of which only 3 can be trusted, but which?

    EDIT: Should I maybe test myself for 3882xxx and 22974xxx at Yseq - or is that a waste of money?

    And what exactly does homologous mean?

    Thanks
    Anders!
    Last edited by A.Morup; 02-20-2017 at 09:02 PM.

  14. #20
    Registered Users
    Posts
    4,410
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by A.Morup View Post
    ....
    FTDNA originally gave me:
    Who at FTDNA? Are you talking about what you see in your Big Y results tab or some employee at FTDNA? and if so, the help desk or someone on the Y team?

    FTDNA has at least two levels of calling results out of Big Y. The first is the automated calling system that is fairly strict but also fairly dumb and does not look at phylogenetic consistency.

    Their second level is a direct BAM file interpretation done by their Y team, specifically Michael Sager. He is their haplotree maintenance guy and he can force a manual haplogroup update as well as add SNPs to their haplotree. What he does is similar to YFull in that it is a BAM file interpretation and he looks at similar kinds of things.

    If you think you have identified a new branch, particularly if YFull thinks they have, I would submit it to Michael Sager with the kit #s proving the branch positioning and the SNP position-anc-der details. Probably, Michael will agree with YFull. Sometimes he finds additional SNPs or additional branching or problems with recurrency because he has access to all FTDNA's BAM files. Sager does not provide you with interpretation and analysis reports, he just reviews the data, updates the haplotree and reports back to you. Generally, he responds within two business days.

    If you haven't discovered new branching, the important thing you can do is check your Y matches and try to recruit someone else in your cluster to do a Big Y also.... which is how you will discover new branching. Genetic genealogy is a team sport.

    Quote Originally Posted by A.Morup View Post
    And what exactly does homologous mean?
    It just means that an SNP you are looking at sits in a region of the Y chromosome that is similar to some region on a different chromosome, making it more difficult to sequence.

    See this paper. "X and Y Chromosomes: Homologous Regions"
    http://onlinelibrary.wiley.com/doi/1...05793/abstract
    Last edited by TigerMW; 02-21-2017 at 04:04 PM.

  15. The Following 2 Users Say Thank You to TigerMW For This Useful Post:

     A.Morup (02-21-2017),  JMcB (02-21-2017)

Page 2 of 3 FirstFirst 123 LastLast

Similar Threads

  1. A Mutation Tree for YFull's NGS "500 STRs"
    By Dave-V in forum R1b-L21
    Replies: 27
    Last Post: 05-16-2017, 04:39 PM
  2. Replies: 15
    Last Post: 01-30-2017, 09:51 PM
  3. Replies: 12
    Last Post: 04-20-2016, 03:09 PM
  4. Replies: 1
    Last Post: 06-07-2015, 03:05 AM
  5. "Missing Thread" Confusion
    By Administrator in forum Forum Support
    Replies: 0
    Last Post: 05-16-2015, 04:40 PM

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
  •