Announcement

Collapse
No announcement yet.

Autosomal DNA Test 2

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

  • #31
    Square-rooted to find the actual frequency of what?

    Square-rooting a result for a homozygotic entry to get a frequency for the single allele implies that the single allele is less frequent than the homozygote. That can't be true.

    Any population that has a frequency for the homozygote must also have a frequency for the single allele - a different frequency to be sure, but the frequency measure for the homozygote is good enough for comparison to other populations.

    There is no other way to research a individual allele on the online db's other than by running it as a homozygote - it seems to be a bias that has to be accepted

    Tom

    Comment


    • #32
      Calc

      EXAMPLE:

      Find allele frequency for TH01=8.

      Do the allele homocygotes search in ENFSI DB with the value 8.

      The output for the "actual matching probability" for Norway be 0.00884717

      Do the square root of 0.00884717 and you will get 0.094059396

      This approach seem correct to me because of the formula

      Pm=Pj^2

      where Pj is the actual frequency for the allele in the population.

      The probability of getting two 8 alleles in TH01 should be the probabilty times the probabily.

      Noaide

      In the Norwegian population DB the frequency for this particular allele is 0.0884, but it is slightly different because of different sample size in these different samples.

      Comment


      • #33
        Thank you Noaide, that is useful information.

        Tom

        Comment


        • #34
          this math stuff is so over my head.

          I have a question, for those of you that did ABDNA or DNATribes and got Native American, did this also show up in your FTDNA as Haplogroup Q???

          Comment


          • #35
            Q is Native y-DNA. Believe most posters in these threads so far have only done mt-DNA or didn't get a Native haplo for either y or mt-DNA.

            Comment


            • #36
              Originally posted by Noaide
              EXAMPLE:

              This approach seem correct to me because of the formula

              Pm=Pj^2

              where Pj is the actual frequency for the allele in the population.

              .

              Noaide,

              Would you check the formula on the csfs site and see if we should follow the same square-rooting procedure for homozygotic entries on that site.

              Thank you,

              Tom

              Comment


              • #37
                CSFS plain vanilla

                Originally posted by tomcat
                Noaide,

                Would you check the formula on the csfs site and see if we should follow the same square-rooting procedure for homozygotic entries on that site.

                Thank you,

                Tom
                It is even more easy, inputing only homocygotes TH01=8 will give you the plain frequency allele. I checked the Omnipop "RCMP Saskatchewan Native" population frequency and its identical to the one in the output result for "RCMP Saskatchewan" under the label "value". You may check yourself the value should be for TH01=8 be 0.034 in both, no rooting neccessary.

                CSFS

                Noaide

                Comment


                • #38
                  You find most of the freq table for the CSFS DB here, they even have the raw profile data!

                  http://www.csfs.ca/databases/index.htm

                  Noaide

                  Comment


                  • #39
                    Learn probability calculations!

                    You should all study some probability calculations, it has some very informative use with autosomals, finally you can do statistics for a fun and interesting purpose!

                    One example for homocygotes alleles, the probability for having identical alleles in one loci is P*P = population frequency times population frequency. To have homocygotes alleles strongly indicate that this allele is common among your ancestors and you should look for populations with high frequency of your allele. If the population you believe you come from have high frequency of this allele it is no big deal to get homocygotes alleles, however if it is rare in the population the chance of having it can be over 1:300 if its freq is less than 10%.

                    If your into more advanced stuff, what is the probability of you having two different alleles compared to different populations? This way it is possible to get closer to your population if these two alleles are common in pair among your ancestors.

                    All these probability calculations is the reason why these databases return so incredible small numbers for your match, it is probabilities calculated against each other for 26 alleles.

                    Noaide


                    Noaide
                    Last edited by Noaide; 5 September 2006, 03:33 AM.

                    Comment


                    • #40
                      Originally posted by tomcat
                      Of the 8 autosomal STR's that I can cross-reference those that score highest as Native American on an American database score highest as Finnish and Estonian on the European database. Might be evidence of the "Siberian" roots of the two populations and/or population bottlenecks both experienced leading to reduced heterozygosity.

                      Tom
                      May I ask what particular loci and allele values you are talkning about? To me the Finn, Swede and Norwegian allele distribtion look quite similar.

                      Noaide

                      Comment


                      • #41
                        Originally posted by Noaide
                        May I ask what particular loci and allele values you are talkning about? To me the Finn, Swede and Norwegian allele distribtion look quite similar.

                        Noaide
                        See Native American 'Tribes thread Page 4. Comparison of results from four sources. The ENFSI numbers have not been square-rooted.

                        Tom

                        Comment

                        Working...
                        X