Page 1 of 8 123 ... LastLast
Results 1 to 10 of 76

Thread: DF27: Let's not be rash

  1. #1
    Registered Users
    Posts
    149
    Sex
    Location
    Britain
    Ethnicity
    British
    Nationality
    British
    Y-DNA (P)
    R1b DF27>FGC17112
    mtDNA (M)
    V

    DF27: Let's not be rash

    It's been a wonderful year for DF27 with over 100 of the very early subclades now discovered as well as thousands of downstream snips revealed. That said there is still some way to go however. The Big Tree suggests that there are are at least 14 more subclades yet to be discovered at just the second level under DF27. With more testing this could well turn out to be 20 subclades or more. At just three or four levels under DF27 there are many snips still to be discovered.

    With the commercial success of the FTDNA R1b panel, an early release DF27 panel must seem tempting. There are still too many gaps in the early branching however. There are also large blocks of snips that still need splitting to determine what really is the lead SNP in the block. This is already evident in the R1b panel where DF27 lead snips that looked really old only a couple of months ago are now known to be 5, 6 or 7 levels below DF27 itself whilst much older snips are missing simply because their seniority was unknown just a couple of months ago.

    I hope we can wait at least another 3 to 6 months or more before rushing to design a DF27 panel at FTDNA.

    Earl.

  2. The Following 13 Users Say Thank You to Earl Davis For This Useful Post:

     Agamemnon (09-03-2015),  gogogenes (08-31-2015),  gotten (09-01-2015),  Gray Fox (08-30-2015),  Krefter (09-03-2015),  Kvenlander (09-01-2015),  Lugus (09-03-2015),  R.Rocca (09-01-2015),  razyn (08-29-2015),  REWM (08-29-2015),  Rick (09-01-2015),  Táltos (09-01-2015),  TigerMW (09-01-2015)

  3. #2
    Registered Users
    Posts
    4,393
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by Earl Davis View Post
    It's been a wonderful year for DF27 with over 100 of the very early subclades now discovered as well as thousands of downstream snips revealed. That said there is still some way to go however. The Big Tree suggests that there are are at least 14 more subclades yet to be discovered at just the second level under DF27. With more testing this could well turn out to be 20 subclades or more. At just three or four levels under DF27 there are many snips still to be discovered.

    With the commercial success of the FTDNA R1b panel, an early release DF27 panel must seem tempting. There are still too many gaps in the early branching however. There are also large blocks of snips that still need splitting to determine what really is the lead SNP in the block. This is already evident in the R1b panel where DF27 lead snips that looked really old only a couple of months ago are now known to be 5, 6 or 7 levels below DF27 itself whilst much older snips are missing simply because their seniority was unknown just a couple of months ago.

    I hope we can wait at least another 3 to 6 months or more before rushing to design a DF27 panel at FTDNA.

    Earl.
    This is the constant problem of fixed SNP panels and packs of all sorts. The next Big Y test result could bring some valuable new information. Hence fixed SNP packages are doomed to be obsolete on the day they are released or soon thereafter.

    Earl, I've argued your point on this in knock down drag out discussions with the Thomas Krahn on the R1b-L21-project yahoo group. It essentially comes down to that you should get a Next Generation Sequencing test (Big Y or FGC Elite) if you can and are inclined to at all and are interested in genetic genealogy. That's always a good investment. I also say upgrading to 111 STRs is always a good investment.

    ... but in the discussions I've had with Mr. Krahn, I have to concede that if Big Y is not in the picture then fixed SNP packages are a good value and have a time value. It is worth something to know something now rather than later.

    On another thread, another project administrator emphasizes this time value in relation to SNP pack availability.
    Quote Originally Posted by lgmayka View Post
    ... My primary concern is that with every passing year, more testees die and more account custodians become unreachable or simply disinterested. Delay has its own cost--in knowledge lost.
    http://www.anthrogenica.com/showthre...l=1#post100059

    Oftentimes (maybe all of the time), one's view depends on one's position. I've found that generally Big Y testers that have results in from multiple members of the cluster are oftentimes anxious to have all of their SNPs out their in the market on these fixed SNP tests. Those that have NGS results waiting and want just a little more time.

    If interested in NGS testing, the good news is there is no need to wait, The below posting is from the U106 group. This is NOT the rule but he says he as seen a turn around of only 16 days. There is hope for those who want Big Y results included.
    ---------- Forwarded message ----------
    From: Charles Moore .... [R1b1c_U106-S21] ....
    The coupon code swede2015 will apparently reduce the price of Big Y from $575 to $488.75. We recently had a Big Y result received only 16 days after it was ordered during this slow time of year. Who knows, but something to consider.
    Surprisingly, SNPs can be added to packs and panels at the last minute.
    Last edited by TigerMW; 09-02-2015 at 11:18 PM.

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

     Agamemnon (09-03-2015),  gotten (09-01-2015),  lgmayka (09-01-2015),  R.Rocca (09-01-2015),  red (11-08-2015),  Rick (09-01-2015),  Táltos (09-01-2015)

  5. #3
    Registered Users
    Posts
    68
    Sex
    Ethnicity
    Finnish
    Y-DNA (P)
    R1b - FGC17112
    mtDNA (M)
    W1b1

    Finland
    Do you know, generally, how many people need to share an SNP before it is taken into offering of FTDNA?

  6. The Following User Says Thank You to Kvenlander For This Useful Post:

     dp (09-03-2015)

  7. #4
    Registered Users
    Posts
    4,393
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    Quote Originally Posted by Kvenlander View Post
    Do you know, generally, how many people need to share an SNP before it is taken into offering of FTDNA?
    At least two. That's the minimum I've seen. Of course, I would think it helps if they are of different surnames.

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

     dp (09-03-2015),  Kvenlander (09-03-2015),  red (11-08-2015)

  9. #5
    Registered Users
    Posts
    4,393
    Sex
    Y-DNA (P)
    R1b
    mtDNA (M)
    H

    What does Razyn think about packs and panels? He's Mr. DF27.

  10. #6
    Gold Class Member
    Posts
    2,112
    Sex
    Location
    Virginia, USA
    Y-DNA (P)
    DF27, FGC15743
    mtDNA (M)
    T2f3

    Quote Originally Posted by Mikewww View Post
    What does Razyn think about packs and panels? He's Mr. DF27.
    Whoa. Mr. DF27, heavy responsibility.

    Well, I think it's coming along, at YSEQ -- but I'm not supposed to say that at FTDNA, which is where I have the actual reins of the project. Last weekend Thomas sent me the current version of his DF27 panel and it's really obvious, visually or graphically, how much better-defined (as a tree) the Z195+ side is -- compared with the ZZ12 side. That still looks more like a comb structure, or "haircomb" as the Brits seem to say. This is pretty apparent on the Big Tree, too; it's just more obvious in the vertical format Thomas uses. Maybe I can do a couple of illustrative screen shots or something. I'm trying to reduce these to a size you can't actually read, just to illustrate the difference between how well we have sorted the Z195+ branch in 4 1/2 years of testing it (and that's the one, part of which exploded in Iberia in the past 2-3 millennia), versus the rest of DF27. Those of us who believe palindromic SNPs are important call it the ZZ12+ branch, but anyway, it's the rest of DF27.

    So, here's the part that looks like a tree -- Z195+ with its numerous subclades:
    Screen Shot 2015-09-02 at 7.42.44 PM.png

    And here's the rest of DF27 -- the ZZ12+ part that still looks like a comb. All of those little tags in red at the right mean "new," not always in discovery sequence, but in having been added to the YSEQ panel for DF27:
    Screen Shot 2015-09-02 at 7.44.04 PM.png

    That's as of 8/28/15 -- and as perceived by a guy who does not intend to test for SNPs on palindromes, or in repeat regions, mirroring the X chromosome, near the centromere, etc. (so it is never going to agree completely with Alex Williamson's Big Tree). If we were to include those additional known SNPs -- and/or wait about a month, while new discoveries (and matches with older ones) keep trickling in -- it would look substantially different. Mainly, it would look like a wider comb. That being the case, my instincts are to wait a bit -- and get it a little closer to right, before trying to sell it. YMMV. I have no veto at YSEQ -- and as far as I know, no voice whatever with the lab side of FTDNA.

    YFull shares most of the preconditions I attributed to YSEQ, plus one against indels (that I think Thomas will test, anyhow). Some of these distinctions exclude the basic documentation of SNPs that form, or once formed, major branching points on the phylogenetic tree of DF27. NextGen testing, coupled with creative analysis of many results, can get around these lab-specific or technology-limited problems. The desired quick-and-cheap pre-loaded panel, pack, and chip tests cannot get past those obstacles; and unless either attitudes or technological capabilities at the relevant labs change quite a bit, will not.

    So, Mr. DF27 does not really carry much of a torch for them. I kind of like targeted NextGen testing; analysis by the people who are good at that; sorting the project by SNP trees to see what STR patterns might point toward what SNPs -- and then testing a handful of one's most likely SNPs for a reasonable price, maybe more like $20 than $40 each.

    Another thing I would like, were it ever to happen, would be intelligent revision of the woefully inadequate criteria that generate the Novel Variants list, after a BigY test. If that feature worked as it should, perhaps we wouldn't need to jump a hopscotch path through a maze wearing a blindfold, to learn what useful results our $575 test has found.

  11. The Following 9 Users Say Thank You to razyn For This Useful Post:

     Agamemnon (09-03-2015),  ehjelt (09-12-2015),  gogogenes (09-03-2015),  gotten (09-03-2015),  Gray Fox (09-03-2015),  ISC80 (09-03-2015),  Kvenlander (09-03-2015),  R.Rocca (09-03-2015),  red (10-09-2015)

  12. #7
    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 razyn View Post
    Whoa. Mr. DF27, heavy responsibility.

    SNIP
    I'm trying to reduce these to a size you can't actually read, just to illustrate the difference between how well we have sorted the Z195+ branch in 4 1/2 years of testing it (and that's the one, part of which exploded in Iberia in the past 2-3 millennia), versus the rest of DF27.
    SNIP27.
    You should be able to see haplogroup "bottlenecks" corresponding to the Iron Age cold epoch around DF27... There are explosions of new haplogroups after things warmed up appearing in the U106 region.

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

     red (10-09-2015)

  14. #8
    Gold Class Member
    Posts
    2,718
    Sex

    Quote Originally Posted by razyn View Post
    Last weekend Thomas sent me the current version of his DF27 panel and it's really obvious, visually or graphically, how much better-defined (as a tree) the Z195+ side is -- compared with the ZZ12 side.
    ...
    If we were to include those additional known SNPs -- and/or wait about a month, while new discoveries (and matches with older ones) keep trickling in -- it would look substantially different. Mainly, it would look like a wider comb.
    That is not a difference in "definition," as if greater testing would reduce the difference. What you are describing is a major difference in fanout, which roughly but accurately reflects the actual speed of expansion. A tree node which has two or three subclades indicates steady expansion, whereas a tree node with ten subclades (or singletons) suggests a demographic and/or geographic explosion.

    This kind of difference is not unique to DF27. Take a look at the tree for I-CTS10228 (sometimes called "Dinaric"). We see:
    - One singleton, CTS10228*, that barely survived
    - The Z17855 clade, which did only a little better
    - The Y4460 clade, which shows substantial growth
    - The S17250 clade, which in its Y3548 subclade shows a veritable population explosion of singletons as well as several subclades.

    If we were to test the "Dinaric" population more heavily, this difference would not disappear. The subclades would become bushier, and the singletons would become small subclades, but the basic difference in fanout would remain, and in fact probably increase.

    To be fair, there is indeed a way to reduce the fanout (i.e., make the tree bushier): To increase the number of base pairs tested. If everyone were to order FullGenomes' higher-coverage product rather than the Big Y, we would find additional SNPs to consolidate some of the fanout. But we have to live with the current practicalities of pricing, sample availability, etc.
    Last edited by lgmayka; 09-03-2015 at 03:03 AM.

  15. The Following User Says Thank You to lgmayka For This Useful Post:

     Gray Fox (09-03-2015)

  16. #9
    Registered Users
    Posts
    68
    Sex
    Ethnicity
    Finnish
    Y-DNA (P)
    R1b - FGC17112
    mtDNA (M)
    W1b1

    Finland
    Quote Originally Posted by Mikewww View Post
    At least two. That's the minimum I've seen. Of course, I would think it helps if they are of different surnames.
    Thanks Mike. Then there is hope for FGC17112, as apart from Earl and another Davis, and myself (Kainulainen/Kajnul), there is also Rice?

  17. #10
    Gold Class Member
    Posts
    2,112
    Sex
    Location
    Virginia, USA
    Y-DNA (P)
    DF27, FGC15743
    mtDNA (M)
    T2f3

    Quote Originally Posted by lgmayka View Post
    That is not a difference in "definition," as if greater testing would reduce the difference. What you are describing is a major difference in fanout, which roughly but accurately reflects the actual speed of expansion. A tree node which has two or three subclades indicates steady expansion, whereas a tree node with ten subclades (or singletons) suggests a demographic and/or geographic explosion.
    That may be true theoretically; but in practice [noting that the practice is virtually limited to (a) FTDNA testers at the BigY level and (b) the heavily skewed sample of the 1000 Genomes project] a tree node that has two or three (or zero) subclades indicates a very small tested sample, like two to four guys. Whereas the Z195 side has a very large sample, many hundreds of guys (they don't have to have NGS testing if they tested the individual SNP in question, mainly with Sanger sequencing since 2011). Additionally, much of the apparent Z195+ "explosion" happened in the population toward which 1000 Genomes is skewed.

    I think the ZZ12 side of the DF27 project -- or at least several specific, recently-discovered branches of it -- might be as bushy as the Z195 side. We lack the sample, so far, with which the rest of its bushiness can be discovered and differentiated. I don't for a moment believe that the fifty or so subsets of DF27+, ZZ12+ [my second illustration, plus the ones Alex includes but YSEQ doesn't test, plus over a dozen who have NextGen results or 1000 Genomes sequencing with no match yet found, plus some unknown number we have yet to discover at all] are "basal," or almost basal (DF27**) -- meaning they have had one or two mutations in 4900 years.

    I think it's a sampling difference.

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

     gogogenes (09-03-2015),  Kvenlander (09-03-2015),  red (10-09-2015)

Page 1 of 8 123 ... LastLast

Similar Threads

  1. Replies: 45
    Last Post: 06-06-2021, 07:55 PM
  2. Replies: 82
    Last Post: 10-06-2020, 03:06 PM
  3. Replies: 5
    Last Post: 07-27-2019, 01:53 AM
  4. Replies: 56
    Last Post: 12-20-2015, 05:54 PM
  5. Replies: 5
    Last Post: 09-28-2013, 05:49 PM

Posting Permissions

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