Announcement

Collapse
No announcement yet.

Extent of DNA database

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

  • #46
    Originally posted by MMaddi
    A haplotype refers to STR counts, such as a 12 or 37 marker test result. A haplogroup, whether predicted or SNP-tested, refers to a broad division of human males who all share the same SNP set - in other words, those who are R1b1b2 all have the M269 SNP, plus the full set of SNPs going back (R1b1b, R1b1, R1b, R1, etc.).

    That is how I'm using the terminology and what FTDNA means by haplotypes. From what I can tell, when Deirwha refers to haplotypes, I think he's using the same definition too.

    Usually when ancient DNA (from graves) has been tested, only mtDNA results can be obtained. However, in the last year or two, there are a few cases where they've been able to obtain yDNA results. These are haplotypes (STR results) as far as I can recall, not SNP-tested haplogroups.
    I'm well aware of the distinction. My query was "why" the two terms were being used as each other on the thread? No what the two terms mean.

    Sorry if my query muddled things. I was trying to point out that the use of either is not right. Next time, I'll just come out and say that.
    Last edited by rucksack; 14 December 2008, 08:56 PM.

    Comment


    • #47
      Originally posted by rucksack
      I'm well aware of the distinction. My query was "why" the two terms were being used as each other on the thread? No what the two terms mean.

      Sorry if my query muddled things. I was trying to point out that the use of either is not right. Next time, I'll just come out and say that.
      Actually, you helped to unmuddle the discussion in this thread. It turned out that Deirwha was badly misunderstanding the terminology involved. If you hadn't noticed something was amiss, I wouldn't have bothered to define the terms.

      Once that was done, we got to the core question that he was trying to discuss, which dealt with haplogroups/SNPs, not how many 37 marker haplotypes FTDNA has in their database.

      Comment


      • #48
        Originally posted by MMaddi
        If you use haplotype in a loose way, then you could say that the HVR1 and HVR2 mutations are your mtDNA haplotype. But it is not the same as a yDNA haplotype, which is counting repeats at a marker, whereas an HVR1 or HVR2 mutation is a change in a specific nucleotide base from one to another, which is analagous to a yDNA SNP.

        So the strict answer to your question is no, although I've seen some good genetic genealogists use the term "haplotype" for mtDNA mutations, in a loose sense.

        Thank you MMaddi

        I had wondered about that.
        What about the PBS tv program that matched dna from a present day Mongolian girl to that of ancient dna extracted from the remains of a supposed Amazon (female warrior in Eurasia)? They had said it was a match (the girl was/is a descendant of the Amazon). Would the match be just a haplogroup (maybe with the same mutations)?

        http://www.pbs.org/wnet/secrets/prev...zon/clues.html

        http://www.pbs.org/wnet/secrets/prev...zon/index.html

        http://www.pbs.org/wnet/secrets/prev...zon/about.html

        And they didn't even announce what the haplogroup was.
        Last edited by rainbow; 14 December 2008, 09:33 PM.

        Comment


        • #49
          As I've posted before, the GP will do backbone SNP testing, at least if necessary. When I received my own 12 marker report from them, they told me I was yDNA haplogroup "J". Immediately (instantaneously) upon uploading those results to FTDNA, I found out that I was indeed J1 - my results page showed the requisite SNPs, M304 & M267. I then had to order further deep clade tests to get to J1e*. Likewise, for my ggf's haplotype (through my distant cousin), GP tested SNPs to get the results to E1b1b1 because his 12 marker haplotype is so unusual. I've since ordered the further deep clade, and I'm still waiting for the full results which were due 11/19 and are apparently backed up by all the new first time customer orders, but that discussion's for another thread...

          As for the stats, the bottom line is that all of the FTDNA testing is done through the process of self-selection - we've all chosen to test - it is not done through scientific random sampling of any specific population. That is why the database is skewed toward the British Isles; so far, it has been people with origins in that area of the world who have most chosen to have DNA testing done. Likewise, I believe, the same is true of the GP except for some areas of the world where they (or other researchers) go and specifically test particular groups of people. However, I have not read any papers that have been based on true random sampling of any populations; of course non-random samples can be adjusted to better fit the profiles of the larger populations from which they are taken, which might allow some generalization of the results. If anyone knows of studies based on true random sampling, please let me know.
          Last edited by vinnie; 14 December 2008, 10:37 PM.

          Comment


          • #50
            I don't see how scientifically "random" testing can ever be accomplished when it comes to dna testing.

            I know of no country on earth that would allow dna testing to be done on individuals that have not consented to it.

            Once consent is achieved, you throw out the random possibility, and you arrive at the same open minded people that tested in the first place.

            Comment


            • #51
              Originally posted by rucksack
              I don't see how scientifically "random" testing can ever be accomplished when it comes to dna testing.

              I know of no country on earth that would allow dna testing to be done on individuals that have not consented to it.

              Once consent is achieved, you throw out the random possibility, and you arrive at the same open minded people that tested in the first place.

              The National Geographic takes no names. They test people that have lived in the area for long generations only.

              Comment


              • #52
                Random sampling has nothing to do with anonymity; those are two different issues. Additionally, it could be possible to achieve random sampling of a population, even with the requirement of participant consent. For example, suppose you have a defined population of 5,000 people (e.g., a tribe in Oceania), and you draw a 10% random sample of 500 because that's the number of participants you need to provide enough data to perform the particular stats you want to do. After contacting all the randomly chosen potential participants, you find that 200 don't want to participate, so you then draw another random sample of the population in order to identify 200 more willing participants; this is called random sampling with replacement. Granted, there are a lot of other issues involved such as government consent, and how well the final sample represents the original population in terms of age, gender, etc., but conducting such DNA testing is theoretically possible. After all, random sampling simply means that each individual in a given population has as much chance of participating in a study as any other individual in the same population.

                Comment


                • #53
                  Random

                  Random relates to the method of selecting participants. It is distinguishable, for instance, from self selection, a process which risks skewing the results toward a specific sub segment of the population.

                  Many forms of scientific study rely upon the taking of random samples. These samples need not be extensive so long as they are representative and taken under controlled conditions.

                  It is an ethical and often legal requirement that participation be secured by fully informed consent. Getting to the requisite sample may take considerable time as potential participants selected in accordance with pre determined criteria are given the opportunity to consent and participate or turn down the opportunity.

                  Among the features of such a testing protocol is a full disclosure of the protocols and procedures followed in selecting, screening, and obtaining consent; and also of the records of that process. All this is designed to eliminate skew and to ensure replicability of the experiment. I expect that random sampling is a routine aspect of the protocol followed by those who conduct tests that lead to academic papers concerning population genetics. There is always considerable debate about how large a sample must be and the protocol one has followed in sampling. That is a part of the peer review process.

                  Similar protocols are followed in professional polling and marketing studies. However there the polling companies opt not to share the records on their protocols for commercial reasons. In an academic setting not sharing the records on the protocol is a reason to suspect the result. Scientific experiments should be replicable.

                  I think probably the greatest obstacle to conducting random testing in our field of interest is not the possibility of securing volunteers within the framework of an acceptable protocol. I suspect that the two greatest obstacles are who will fund and the sharing of trade secrets, which includes compilations of data. Someone from FT DNA feel free to correct my understanding. I am not a scientist. Beyond college level exposure and participation in some experiments my thinking on this has been shaped by political polling and market studies.
                  Last edited by Deirwha; 15 December 2008, 03:50 PM.

                  Comment

                  Working...
                  X