Announcement

Collapse
No announcement yet.

Deep Subclades ... Who Cares?

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

  • Deep Subclades ... Who Cares?

    Maybe the hardcore posters and data crunchers can educate me, but I don't see what the need is for SNP and deep subclade testing.

    First, while there is a certain degree of anthopological evidence that distinguishes R1b and I (e.g. western europe and central europe), there aren't any studies that firmly establish any anthropological claims if R1b1c1 versus R1B1c2. (or M65+ versus M126+ for that matter) So who cares whether or not someone can make the statement that they are one or the other? There aren't any studies that conclusively state that one subclade is more, celtic, pict, brytonic, etc. than the other.

    Second, the phylogenetic tree seems arbitrary in its selection of mutations (M) and polymorphisms (P). Maybe I'm missing something, but it seems to me that there could be any of a number of true phylogenetic trees depending upon which mutations one selects to partition the data. For example, it is generally agreed that 80% of all Europeans are R1b. If we find out that 90% of all R1b is R1b1c1, then what have we learned? Probably only that the mutations selected to subdivide the R1b data were not the ones most conducive to this kind of analysis.

    Third, any discussion of SNPs and deep subclades extends back beyond thousands of years. While this may be of some anthropological interest at some time in the future, it really doesn't provide any tangible information for the average FTDNA participant who is more genealogically focused.

    Don't get me wrong, I don't blame FTDNA. Other labs were selling all this SNP and subclade testing and FTDNA had to respond. All I'm asking is what value is it?

    Also, I would like to know what STR markers and values correspond (i.e. predict) these subclades? Just like FTDNA was able to predict an I, C3, or R1b from their STR tests, shouldn't their haplogroup predictor be extended to also make a SWAG at the subclades?

    I'm sure that questioning the value of this type of testing maybe considered heresy by many, but I just don't undertstand. Everyone out there who is M126+ (i.e., R1b1c3) is welcome to respond and tell me how they've made practical use of this knowledge.

  • #2
    Good questions. Looking forward to answers on this.

    Comment


    • #3
      Originally posted by K. Campbell
      Maybe the hardcore posters and data crunchers can educate me, but I don't see what the need is for SNP and deep subclade testing.

      First, while there is a certain degree of anthopological evidence that distinguishes R1b and I (e.g. western europe and central europe), there aren't any studies that firmly establish any anthropological claims if R1b1c1 versus R1B1c2. (or M65+ versus M126+ for that matter) So who cares whether or not someone can make the statement that they are one or the other? There aren't any studies that conclusively state that one subclade is more, celtic, pict, brytonic, etc. than the other.
      I see your point. Some people may appear excessively worried about such seemingly insignificant details. Obviously there isn't a single answer that could explain what drives each one in their search for anthropological answers, although I don't really see how this hobby is any different from other hobbies that people follow passionately.

      Second, the phylogenetic tree seems arbitrary in its selection of mutations (M) and polymorphisms (P). Maybe I'm missing something, but it seems to me that there could be any of a number of true phylogenetic trees depending upon which mutations one selects to partition the data. For example, it is generally agreed that 80% of all Europeans are R1b. If we find out that 90% of all R1b is R1b1c1, then what have we learned? Probably only that the mutations selected to subdivide the R1b data were not the ones most conducive to this kind of analysis.
      I wouldn't say that the selection of SNPs in the current phylogenetic tree is arbitrary but rather incomplete. The current phylogeny isn't carved in stone, (at least not yet) and it is a given fact that as new genetic discoveries are made it will be revised as many times as it is necessary.

      Third, any discussion of SNPs and deep subclades extends back beyond thousands of years. While this may be of some anthropological interest at some time in the future, it really doesn't provide any tangible information for the average FTDNA participant who is more genealogically focused.
      True. For genealogical purposes, STRs are best suited to establish it there's a genetic affinity between two or more persons.

      Don't get me wrong, I don't blame FTDNA. Other labs were selling all this SNP and subclade testing and FTDNA had to respond. All I'm asking is what value is it?

      Also, I would like to know what STR markers and values correspond (i.e. predict) these subclades? Just like FTDNA was able to predict an I, C3, or R1b from their STR tests, shouldn't their haplogroup predictor be extended to also make a SWAG at the subclades?
      In theory, a set of specific markers should have some predictive value to determine not only the haplogroup but also the subclade of a haplotype. In the E3b Project we have tested that idea with mixed results. In any case, enough haplotypes have to be "snipped" tested first to discover the motifs or patterns indicative of such and such subclades. It is possible also that the markers we're using are not the most appropiate for such purpose and we will never find the correlation.

      I'm sure that questioning the value of this type of testing maybe considered heresy by many, but I just don't undertstand. Everyone out there who is M126+ (i.e., R1b1c3) is welcome to respond and tell me how they've made practical use of this knowledge.
      Heresy? Not al all! We all have our own reasons to test or refrain from testing. In my particular case, although I'm not R1b, and my subclade turned out to be in an "undifferentiated paragroup", I can tell you that knowing that is helping me in a practical way by steering my genealogical research to areas where similar haplogroup/subclades as mine have been detected by the available population genetic studies.

      Regards.

      Comment


      • #4
        It's food for thought.

        Comment


        • #5
          Originally posted by K. Campbell
          Maybe the hardcore posters and data crunchers can educate me, but I don't see what the need is for SNP and deep subclade testing.

          First, while there is a certain degree of anthopological evidence that distinguishes R1b and I (e.g. western europe and central europe), there aren't any studies that firmly establish any anthropological claims if R1b1c1 versus R1B1c2. (or M65+ versus M126+ for that matter) So who cares whether or not someone can make the statement that they are one or the other? There aren't any studies that conclusively state that one subclade is more, celtic, pict, brytonic, etc. than the other.

          Second, the phylogenetic tree seems arbitrary in its selection of mutations (M) and polymorphisms (P). Maybe I'm missing something, but it seems to me that there could be any of a number of true phylogenetic trees depending upon which mutations one selects to partition the data. For example, it is generally agreed that 80% of all Europeans are R1b. If we find out that 90% of all R1b is R1b1c1, then what have we learned? Probably only that the mutations selected to subdivide the R1b data were not the ones most conducive to this kind of analysis.

          Third, any discussion of SNPs and deep subclades extends back beyond thousands of years. While this may be of some anthropological interest at some time in the future, it really doesn't provide any tangible information for the average FTDNA participant who is more genealogically focused.

          Don't get me wrong, I don't blame FTDNA. Other labs were selling all this SNP and subclade testing and FTDNA had to respond. All I'm asking is what value is it?

          Also, I would like to know what STR markers and values correspond (i.e. predict) these subclades? Just like FTDNA was able to predict an I, C3, or R1b from their STR tests, shouldn't their haplogroup predictor be extended to also make a SWAG at the subclades?

          I'm sure that questioning the value of this type of testing maybe considered heresy by many, but I just don't undertstand. Everyone out there who is M126+ (i.e., R1b1c3) is welcome to respond and tell me how they've made practical use of this knowledge.
          look i think of it as a road map to the historic past one my ancestors took.
          my dna split from others.as an e3b i came thru ethopia off the growing tree.
          this all sounds simple now.but vision is what is the real thing needed here.

          9,000,000 members in the database oh lets say victors e3b project has 1,750,000 members. that little tree he has now has many more splits and turns and the roadway is more complex. the real good thing is ftdna has said from now on it will update automaticly when they find a new roadway. shocked overwelmed what happens when its 25,000,000.

          this should be able to diferenciate between 37/37 matches

          Comment


          • #6
            To SNP or not to SNP

            Originally posted by K. Campbell
            First, while there is a certain degree of anthopological evidence that distinguishes R1b and I (e.g. western europe and central europe), there aren't any studies that firmly establish any anthropological claims if R1b1c1 versus R1B1c2. (or M65+ versus M126+ for that matter) So who cares whether or not someone can make the statement that they are one or the other? There aren't any studies that conclusively state that one subclade is more, celtic, pict, brytonic, etc. than the other.
            It's quite obvious that knowing your subclade, especially if it's part of R1b, will only tell you about your deep ancestry and not be of much immediate help in solving genealogical questions that usually go back a couple of hundred years or perhaps 500 years. However, if we take a long-range view of the research involving DNA testing, anyone now being tested for their haplogroup today is a pioneer who is establishing the conditions for better scientific understanding of population genetics and its application to genetic genealogy in the future. Hopefully, that will be in our lifetime.

            One question to think about is why should we assume that all SNPs occurred thousands of years ago. Isn't it possible that men have had SNPs occur on their y chromosome in the last few hundred years? And if that's the case, won't more SNP testing of people lead to a better chance of discovering these more recent SNPs? If that scenario is true, which I think we will find to be the case in the next ten years or less, then we are going to find SNPs which aid people in answering their immediate genealogical research questions.

            I'll add that we already have indications that this is the direction in which SNP testing is heading. EthnoAncestry, with whom FTDNA is trying to compete through the new deep clade tests, has dicovered several new SNPs in R1b and I in the last 6 months. This goes in the direction I mentioned above. So now plain vanilla R1b can be further divided into more subclades that represent a more recent time period. And population geneticists can fine tune their theories of human migrations in Europe, using these new SNPs, along with knowledge of history, archaeology and languages. That may not have any immediate relevance to specific genealogical questions, but it certainly tells us more about the ethnic groups and nations from which our ancestors came. I find that as fascinating as who my gggg-grandfather was and whom he married.

            Mike Maddi

            Comment


            • #7
              Actually, there are several deep SNPs or clades that CLEARLY show that you are of one ethnicity or another.

              It all depends on where your ancestors originate.

              If for example, you are from Britain, and you test as Hg I, you may assume you are of Viking descent. Indeed, many web pages tout this as gospel truth.

              The real truth is that Hg I has been dissected down to sub-sub-subclades, and most serious scientists do give credence to the ability to tell whether one is of the I subclade that indicates even which invading Germanic tribe (e.g., Anglo-Saxon vs. Norse) from which a British Hg I male derives.

              For more on this, visit www.northwestanalysis.net.

              If this same Briton who is generic Hg I takes a subclade test and turns up as I1b2, this person who assumed descent from Vikings would be dead wrong, as for some reason, the deep Hg I clade of I1b2 accompanied the original R1b settlers to Britain, and this stubborn Brit who saw no need to take a deep SNP test would actually have been derived fromthe people FIGHTING THE VIKINGS...

              Let's now move on to Italy...

              Let's say an Italian tests as J2. Great. In all likelihood, he is descended from the Neolithic farmers who made it to the S. Mediterranean. Nevertheless, if that Italian is J-M267, a deep SNP that evolved only among Arabs, that same Italian would be descended from Arab invaders, not Neolithic farmers.

              The same applies within E3b: M81 in Spain or Sicily is a clear indicator of recent Arab origin. (And incidentally, it is exceedingly rare in both).

              I remind you that five to ten years ago, E3b was considered an advanced "subclade" of E. Pretty big difference to be sub-saharan African (E3a) vs. Neolithic Mediterranean (E3b).

              Most of us don't care what we are ultimately, as long as the test is accurate. A deep SNP test does just that.

              I completely recommend the inexpensive and user friendly www.dnaheritage.com.

              And I am quite sure in the years to come, more details on the clades will emerge.

              Comment


              • #8
                Originally posted by Mikey

                The real truth is that Hg I has been dissected down to sub-sub-subclades, and most serious scientists do give credence to the ability to tell whether one is of the I subclade that indicates even which invading Germanic tribe (e.g., Anglo-Saxon vs. Norse) from which a British Hg I male derives.
                [/url].

                If this same Briton who is generic Hg I takes a subclade test and turns up as I1b2, this person who assumed descent from Vikings would be dead wrong, as for some reason, the deep Hg I clade of I1b2 accompanied the original R1b settlers to Britain, and this stubborn Brit who saw no need to take a deep SNP test would actually have been derived fromthe people FIGHTING THE VIKINGS...
                Yes. To me (I1a), a subclade test confirmed that my ancestors probably arrived Finland from Norway, although the REO matches strongly suggested it. If I had been I1b, it would probably have meant my ancestors came from eastern Europe instead of western.

                Comment


                • #9
                  Help me Understand

                  OK, Jim seems to be arguing that more information is better. That's valid but what good is the extra information and why would one want to use SNPs (older variations) instead of STRs (younger variations) to differentiate a 37/37 match? What we need is more markers instead of a different kind of test.

                  Mike Maddi makes essentially the same point -- that SNP testing adds to the greater body of knowledge and may be of use at some point in the future. Maybe at some point SNP testing may be used for recent ancestry, but as of now, all SNP analysis is solely for deep ancestry (thousands of years).

                  I disagree with Mikey, as I said in my original post -- at the coarsest level (HG1, HG2 or the newer designations of R1b and I) some anthropological conclusions have been drawn. Mikey goes on to reference Ken Norveldt's site as an example detailed subclade analysis with anthropological conclusions.

                  The fact is that Ken is NOT using SNP testing and the phylogenetic tree to try to discern population groups in a "top-down" way. Instead Ken is subdividing empiric STR data form the SMGF data to try to infer population tendencies. Ken's work is more of a "bottoms-up" data mining approach. Look at Ken's I varieties and notice that they are all defined on the basis of STR values not SNP values.

                  Mikey goes on to state that "most serious scientists do give credence to the ability to tell whether one is of the I subclade that indicates even which invading Germanic tribe (e.g., Anglo-Saxon vs. Norse) from which a British Hg I male derives."

                  To this statement I have the following response.
                  1. People trying to distinguish Anglo-Saxons from Norse are using STR values and not SNP values.

                  2. These populations are muddied and overlapping. Analysis shows that these populations have 60/40 or 40/60 tendencies instead of being able absolutely and definitively distinguishable based on a SNP mutation. I challenge you to try to relate the 8 groups that Ken defines for I1a-AngloSaxon (1 thru 8), I1a-Norse, and I1a-UltraNorse in his figure 1 to any of the defining I1a SNPs on the FTDNA web site (M21, M72, M227)

                  3. With all due respect to the hard-core data miners out there, the work published on the internet has to be considered suspect. While these fellows are smart and diligent, the fact remains that their work is based upon questionable data. For example, (1) lineages and oldest proven ancestors are voluntarily provided by participants and are not validated; (2) the overall data population is biased in favor of North American participants and not necessarily uniformly representative; (3) when the population is subdivided into small enough loci to draw conclusions, the validity of the sample size has to be questioned. Above all, these speculations are not rigorous academic works that have been vetted and juried by their peers. Instead, I'm reminded of the old statistical quip -- "If you torture the data long enough, it will tell you whatever you want.". I think that I can hear the data screaming!

                  Keep the feedback coming -- tell me again what being M72+ (I1a3) tells me

                  Comment


                  • #10
                    Originally posted by K. Campbell
                    Keep the feedback coming -- tell me again what being M72+ (I1a3) tells me
                    According to this website, I1a3 tells you that your ancestors most likely came from Denmark or Norway. I1a2 would have told you they most likely came from The Netherlands, Switzerland, Germany or Austria:

                    http://freepages.genealogy.rootsweb....i1a_part_1.htm

                    As far as I know, STRs can only predict to which subclade you belong, but SNPs can verify it. I've seen STR values of people belonging to I1b that are close to my I1a values. A haplogroup predictor gave only about 40% probability for me to belong to I1a and about 30% probability for I1b.
                    Last edited by Eki; 9 March 2006, 08:39 AM.

                    Comment


                    • #11
                      Originally posted by Eki
                      As far as I know, STRs can only predict to which subclade you belong, but SNPs can verify it. I've seen STR values of people belonging to I1b that are close to my I1a values. A haplogroup predictor gave only about 40% probability for me to belong to I1a and about 30% probability for I1b.
                      I believe that K. Campbell's essential argument is that SNP tests will not help him or anyone else in a genetic genealogy sense. That is, he's saying that a SNP test will not tell him if he and another person share a common ancestor in the last few hundred years. That's essentially correct, since SNPs, at least the ones we currently know about, are thousands of years old. I don't dispute his argument there, although there may be rare cases of a close match at 37 markers where a SNP test will rule out common ancestors if both men are SNPed and have different haplogroups or subclades.

                      I stated in my first posting where I do disagree with him. There won't be any advance in scientific knowledge and its application if the tests aren't performed and analyzed and compared. What would have happened, if 5 or 6 years ago FTDNA had decided that there weren't enough people testing to have a meaningful database they could use to connect people with distant ancestors? Obviously, the database would not be here today to accomplish what it does and all of us would not be debating SNPs on this board. However, FTDNA and enough inquisitive genealogists got together and the testing went on without any immediate prospect of success. My contention is that the same situation exists with SNP testing today.

                      Mike

                      Comment


                      • #12
                        Originally posted by MMaddi
                        I believe that K. Campbell's essential argument is that SNP tests will not help him or anyone else in a genetic genealogy sense. That is, he's saying that a SNP test will not tell him if he and another person share a common ancestor in the last few hundred years.
                        There are more than one reason to take a DNA test. Church records tell me who my ancestors were 300 years ago and that they all lived in Finland. What I want to know is where my ancestors lived 500, 1000, 1500, etc. years ago.

                        Comment


                        • #13
                          Originally posted by MMaddi
                          I believe that K. Campbell's essential argument is that SNP tests will not help him or anyone else in a genetic genealogy sense. That is, he's saying that a SNP test will not tell him if he and another person share a common ancestor in the last few hundred years. That's essentially correct, since SNPs, at least the ones we currently know about, are thousands of years old. I don't dispute his argument there, although there may be rare cases of a close match at 37 markers where a SNP test will rule out common ancestors if both men are SNPed and have different haplogroups or subclades.

                          I stated in my first posting where I do disagree with him. There won't be any advance in scientific knowledge and its application if the tests aren't performed and analyzed and compared. What would have happened, if 5 or 6 years ago FTDNA had decided that there weren't enough people testing to have a meaningful database they could use to connect people with distant ancestors? Obviously, the database would not be here today to accomplish what it does and all of us would not be debating SNPs on this board. However, FTDNA and enough inquisitive genealogists got together and the testing went on without any immediate prospect of success. My contention is that the same situation exists with SNP testing today.

                          Mike

                          hey i have matches that look like serious matchs cause the numbers match but snps look at the 18 385bs and see the way it is. maybe they arent quit the same. when matches are more frequent i want a 37 match who travels the same roads as me . why do people whose matches have the same number yet they test differently on snps. because the different numbers repeat differently. agian you need to see the vision 25,000,000 people and 800 matches 25/25 400 37/37 with paperwork. thats the future

                          Comment


                          • #14
                            Originally posted by K. Campbell
                            OK, Jim seems to be arguing that more information is better. That's valid but what good is the extra information and why would one want to use SNPs (older variations) instead of STRs (younger variations) to differentiate a 37/37 match? What we need is more markers instead of a different kind of test.
                            more marker will help but thats extra info too. at this level of testees we would have more with no matches. whats we need is more testees look what happened after the national geo has happened to group membership
                            more testees then both more markers and snps and a better way to compare mtdna more geo projects

                            all this will come in time

                            a year and a half ago i thought my denning project might have a third memnber by now we have 19

                            Comment


                            • #15
                              The devil with y-str testing is convergence. People descended from different. but similar, haplogroups: R1b1c1 & R1b1c2 can have their y-str markers converge to make it look like they are a close match. Deep SNP testing separate the chaff from the wheat. If the y-str results are close, but the tested subject belong to different sibling subclades, as shown above, we know that they are NOT related in any genealogically relevant way.

                              Moreover, as results accumulate, geneticists may note that one deep R1b clade is concentrated heavily in Ireland, while another is in Wales.

                              My patriline is from Switzerland and is R1b. I don't know if they are Helvetii or Alemanni, but hope that y-DNA testing will some day tell me.

                              Timothy Peterman

                              Comment

                              Working...
                              X