Loading ...
Sorry, an error occurred while loading the content.

Re: [I-M223] Re: may I ask my 'simple' question again?

Expand Messages
  • Cliff. Johnston
    Aaron, Well stated. Thank you. Cliff. ________________________________ From: sallfertorr To: I-M223@yahoogroups.com Sent: Sun, November
    Message 1 of 15 , Nov 6, 2012
      Well stated.
      Thank you.

      From: sallfertorr <sallfertorr@...>
      To: I-M223@yahoogroups.com
      Sent: Sun, November 4, 2012 7:55:44 PM
      Subject: [I-M223] Re: may I ask my 'simple' question again?



      Sorry but I do not know your name, so cannot address you directly.

      As you know, Genetic Genealogy is a new field. As such, we are still learning the best methods and practices and for this reason, you will find little consensus in terms of recommendations. I remember answering your question about three times in the recent past and made the same recommendation to you the three different times. It is up to you whether or not you should follow anyone's recommendations, but let me explain to you where I am coming from. You will see why, though you may get consistent recommendations from specific people, you still will not find much consensus.

      There are two main lines of researchers in Genetic Genealogy at the moment (speaking only of Y-DNA): those who claim they can guess SNP statuses based on STR values; those who believe that STR's and SNP's are complementary. I fit in the second category and, as such, I believe that one always needs to know his terminal SNP if he is to apply his genetic results to genealogical research.

      Let us use P95 as an example. If you look at our project, you will find 5 individuals (four different surnames) who have tested positive for P95. Of all of them, only one has a match in the 67 markers (different last name, though). And another P95+ duo (same last name) is the only family to have matches at the 37 markers (which... not only have a different surname but also are P95-).

      Let us now suppose we didn't know of the existence of P95 and that the duo I mentioned above found this match with another surname at the 37-marker level. What would they think? Since they have no other advanced-marker matches, they'd be convinced that they are related to these matches at some point in the recent past (because of their similarities in STR values). Well, but we do know that P95 exists and we do know that the duo is P95+ whereas their 37-marker match is P95-. So they cannot possibly be related in the recent past, as P95 is much older than the use of surnames in Europe. What does this tell us? P95+ individuals do not even match themselves. There are so many differences in terms of STR values, and they are such a diverse group, they get false matches with individuals outside of the P95 group all the time. Why? Because STR values converge, which gives us the false impression of shared ancestry. Please see: http://en.wikipedia.org/wiki/Convergent_evolution . SNP's are much more stable and, as such, much more reliable than STR's.

      When I took over the project as administrator in March of last year, there was a huge number of individuals who were believed to be P95+ but who hadn't actually tested. Some said the tests were unnecessary because the STR values already indicated their positive status. Well, I need to see things to believe them. So upon testing, we learned that only one of those believed to be P95+ was actually positive for this SNP. Had they not tested, they'd be under a false impression all this time - and their genealogical research could've gone after ghost Most Recent Common Ancestors. This taught me a few things:

      1- P95 is old, much, much older than the use of surnames in Europe. Its estimated age today is based solely on the diversity of haplotypes that we know are P95+. In other words, since we have found so few haplotypes that are P95+, P95 could be a lot older than "guesstimated" and we simply cannot gauge it.

      2- P95 does not exist in large numbers today. Added to the age, this means it is very difficult to find true common denominators in the haplotypes. Similarly to the type of convergence that occurred with P95- individuals, when we have such a small sample of P95+ individuals and their MRCA is so ancient, it is hard to tell whether their common denominators (similar marker values) are due to shared ancestry or to convergence. There are too few samples to create any reliable rules. 456=15, 458=14 (many have these values and yet are P95-). Similarly, I would expect there to be P95+ individuals with different values at DYS458 and DYS456 (they differ at pretty much anything else). 640=13 (there is a huge number of individuals within Cont2b who have these values and yet are P95-). Again, not reliable. Just because we haven't found it, it doesn't mean there couldn't be a P95+ individual with 640 different than 13. Too much time has passed since its emergence and as 111-marker results show us, the P95+ population is very diverse today, though they are not numerous. They are probably even more diverse than our limited imagination allows us to guess.

      3- P95 is an elusive SNP. The large amount of time since its emergence has allowed P95+ individuals to diverge amongst themselves in terms of their STR values - look at the two 111-marker results we have for P95+ individuals and see just how much they differ. I wonder how these 111-marker results affected the age estimation of this SNP. Simultaneously, P95+ individuals converged STR-wise with others who are P95-, creating the false impression of shared ancestry.

      For these reasons, rather than guessing, I always recommend that Cont2b project members spend $29 to know for sure whether they are P95+ or -. If they are not interested in testing more markers at the moment, at least they will be able to immediately apply the results they have so far to their research if they know their terminal SNP. Separately testing 4 individual STR markers to still have to guess whether one is positive or negative for P95 is actually more expensive than testing the SNP itself. This guessing will also have damaging effects on one's genealogical research. We all know that research time is precious and we'd rather not waste it on Ghost Most Recent Common Ancestors who never actually existed.

      In general, I make SNP recommendations at the headings of specific subgroups. There are two SNP's known in the Cont2b population today: P95 and L1201. The latter is even more rare than P95. So if you are negative for P95, go to the upstream SNP, which is L801 or Z76 (they seem to be phyloequivalent). Simply follow the heading backwards until you find your terminal SNP: M223>Z161>L801/Z76>P95. L1290 is confined to Group 6 of Cont2a, not Cont2b. Only individuals who share your terminal SNP should be considered in your genealogical research. Knowing your terminal SNP will go a long way in terms of determining who is actually related to you.

      As researchers, we tend to get impatient because we don't get any close matches. That is not anyone's fault - it simply has to do with how rare a specific line of descent is and with how many children belonging to this line have survived to the present. If you cannot find matches, you can either wait until individuals related to you eventually test or you can draft them. We all have similar frustrations and have to find ways to work around them.

      That said, our project has about 1,2000 members. As you know, most of these members have different surnames so it is not humanly possible for the project administrators to really aid in private genealogical goals. We guide members in terms of establishing their terminal SNP's, then applying this to their matches and finding clues to their genealogical puzzles this way. And we try to get SNP's that are useful to our community accepted by ISOGG and FT DNA, so they are added to the list of generally-recommended SNP's and so more connections between haplotypes, lineages and surnames may be found.

      You are new to this project and I know how easy it is to criticize and how hard it is to see value in the work of others. Please believe that we are doing our best to make everyone's genealogical research easier by organizing these SNP's and understanding the history of the I-M223 haplogroup. Eventually we will arrive at a point where we can match specific SNP's to specific surnames going all the way back on the SNP tree to the emergence of M223. We will then have both a historical and geographical understanding of this haplogroup and all the human beings who have belonged to it. Only then our work will appear useful. Because I know how valuable all this information is, I cannot be dissuaded from this goal.

      Though some use very rabble-rousing arguments against it, I know this project is in the right path. The I-M223 tree is incredibly old, and our surviving lines are much, much older than in the I1 haplogroup, for example. In other words, we can expect a much more complex tree. When I took over this project in March of last year, we had only six known SNP's downstream from M223. Look at how many we have now: http://www.familytreedna.com/public/M223-Y-Clan/default.aspx?section=results . Our work with ISOGG and FT DNA is simply to get this new knowledge officially recognized. It will be beneficial to everyone who is M223+. Again, I think we are in the right path to accomplish our goal: we will arrive at a point where we can match specific SNP's to specific surnames going all the way back on the SNP tree to the emergence of M223, several thousand years ago.

      I hope this has explained some of the work that we do here - though I know there will never be a consensus on how to get this goal accomplished, and I am happy to be criticized for doing the right thing.

      Aaron Torres

      --- In I-M223@yahoogroups.com, "ca.noodle" <ca.noodle@...> wrote:

      > Thanks to Wayne and Mike for the info and Gene for the support.
      > Yes, I imagine that many people find the proceedings in this project
      > just too daunting to try to get a handle on. I have been trying. It is
      > simply not like the usual surname project and I had not known what to
      > expect when I signed up, based purely on the predicted haplogroup for my
      > kit.
      > I do think that the project itself, as an FTDNA project, perhaps needs
      > to accommodate individuals' private genealogical goals better than it
      > seems to be doing, by addressing what we should be doing for our own
      > purposes and not just for the ISOGG-related goals.
      > I'm at a loss to understand your reply, Aaron, since I
      was clearly
      > asking what test(s) to order, and nothing I have said can in any way be
      > interpreted as rejecting suggestions that further testing should be
      > done. You may be happier if you just ignore my questions.
      > As just this short thread shows, I have in fact received conflicting
      > advice/information, in particular as to whether P95 is the necessary
      > first step and whether there is any point to testing for the Zs. this
      > was the reason for trying again. I was hoping for some sort of consensus
      > on a 'simple' 'DO THIS NOW' reply.
      > It strikes me, since there is no financial advantage to ordering two at
      > once, that it makes sense to do 1290 first, since it *might* obviate the
      > need for P95 testing. But I'm still having trouble reconciling:
      > Wayne: "A test for P95 will tell you which branch you are on P95- or
      > P95+ and what to do afterwards.
      Test for L1290 will also do the same but
      > a negative will mean you still have to test P95."
      > Aaron: "All Cont2b individuals should test P95. ... If you are negative
      > for P95, then you should test L1201 or L801 (terminal SNP's available to
      > your group)."
      > (and I've had even different advice privately, also that this kit will
      > just be redundant for the project's general purposes)
      > Apart from the conflicting advice/information I've received directly,
      > FTDNA says, for what to order for this kit -- i.e. all of the items I've
      > copied below marked '!' are shown in orange, "test available", with no
      > guidance as to whether all should be ordered or which to order --
      > https://my.familytreedna.com/snp-order_v2.aspx
      > predicted haplogroup: L59 P221 L36 P222 L34
      P220 M223 P219 P223
      > 1. I2b1*
      > a. M284 ! I2b1a*
      > 1. L137 L126 ! I2b1a1
      > b. M379 ! I2b1b
      > c. P78 ! I2b1c
      > d. P95 ! I2b1d
      > So I'll just make an executive decision. P95 for now. And then I'll come
      > back and ask: What do I do now??
      > Many thanks to all.
      > BTW, in case it's of interest: waiting for others of the same surname to
      > test is hopeless. The surname is not hugely common and there is no
      > reason to think that various clans of the surname, either in England or
      > in the US, are related to one another. The FTDNA surname project has
      > only a couple of dozen members, about equally divided into I1 and
      > R1a1b2, with my kit a total outlier; the same is true in a related
      > surname project with 500+ members, where my kit is again a total
      > outlier.
      > My kit has a reliable paper pedigree to
      pre-1600 (the father named on
      > the c1615 baptism of the grx7 grfather), all in a little English village
      > up to c1800, with the men having sons in their 40s on average. I believe
      > the grx4 grfather c1735, grx3 grfather c1895 and grx2 grfather c1820
      > were the only males in their generations,and in fact the only children
      > of their parents who reproduced at all; many descendants of the c1820
      > ancestor are then traceable. There would only be fellow male-line
      > descendants from earlier generations if they had left the village
      > pre-1735, or if the pre-1600 ancestor came from elsewhere where others
      > remained. I am in touch with a few descendants of the c1820 ancestor,
      > but none is a son of a son of a son; I should probably try to find one
      > and ask him to test, to make sure there has been no NPE in my kit's line
      > since 1820, but I very much doubt that possibility (the options
      > that the testee or his father, grandfather or great-grandfather had a
      > non-surname father, and all are unlikely).
      > I have been informed by FTDNA of one 37-marker match; the kit has a
      > surname that is of no known relevance to me, and all other info is
      > totally private and the person has apparently not uploaded to Ysearch,
      > so I have no way of knowing what the match looks like or whether
      > anything about the kit, like ancestral place of origin, lines up with
      > mine. Ysearch produces 5 genetic-distance-0 matches, all of whom have
      > tested only 8 or 12 markers.
      > So I am one of the whining 'no match for me' crowd, I'm afraid. And I
      > was hoping that the M223 project might produce something. To be
      > perfectly frank, I would really rather have a new television than $500
      > worth of genetic testing, but I will make a start with

    Your message has been successfully submitted and would be delivered to recipients shortly.