No announcement yet.

Big Y Non-Matching Variants

  • Filter
  • Time
  • Show
Clear All
new posts

  • Big Y Non-Matching Variants

    I still do not understand how to interpret non-matching variants. I have two project members who are proven second cousins. Their common ancestor is their great grandfather (born 1830) and they descend from two of his sons. Their Big Ys show nine non-matching variants--six named and three not. I thought I read somewhere that a SNP occurs on the average of 100 years. Even if the variations occurred equally in the two lines that would be 400-500 years back to the MRCA, but proven paper trail places that time to just 100 years. Where is my understanding wrong?

  • #2
    EVEN IF the "average SNP" mutates with that frequency, which doesn't sound plausible to me, the mutation rate is still only an average, first, over the entire universe of (Y chromosome?) SNP's (but we should expect different SNP's to have different mutation rates, shouldn't we?), and second, there is no reason to suppose that every pedigree will turn out to have the same number of SNP mutations over a period of 100 years (about 3 or 4 generations?). Some will have more, some will have less than the average. There is also a distinct possibility of uncertain data for some SNP's, unless the Big Y data files have been read very carefully. If the two members are proven to be second cousins by virtue of having the appropriate amount of total shared cM of autosomal DNA, I don't see any basis for concern here.


    • #3
      Excel probability calculations using the binomial distribution function
      MRCA 1830
      Donor BY (est.) 1950
      Years lapsed 120
      Avg. years to mutation (est.) 144.41
      Non-matching SNPs, total unallocated 9
      Allocation Scenarios
      Donor A Donor B Hypothesis C
      SNPs 4 5 9
      Probability 0.84% 0.14% 0.00%
      Odds 119 733 5,648,909
      Confidence 0.16% 0.02% 0.00%
      Not sure I understand it, but the Big Tree makes some reference to a difference of opinion in the community as to the reliability of certain regions of the Y chromosome for this sort of testing. The 'CombBED' regions?


      • #4
        you must consider the quality of each mismatch on both sides
        My brother and I have 2 named SNPs that do not match I think both are errors in scoring


        • #5
          IIRC, one blog(real weighty authority there, unnamed random blogs on the internet) I recall considering very well done cited one calculation which put the "average mutation rate" for the Y chromosome in general at 1 mutation every 1.5 generations. And that was just based on applying the number of SNP's in the chromosome and using generic observational data on "similar" things that have been observed previously.

          Some circles claim that certain labs are error prone due to one reason or another and the reported mutations don't actually exist, but are instead "laboratory errors" instead of valid data even without evidence(additional testing, third party or otherwise) to further support their position. (Great science work there /sarcasm)

          This also ignores the matter that the High Resolution DNA analysis of humans is often "problematic" to say the least, and tends to be underplayed by certain (marketing) interests.

          Even with perfect lab methodologies, depending on various factors in what gets sampled and how, you can end up testing and not exactly matching your own self. Biology is messy.


          • #6
            For Big Y matching criteria see Roberta Estes article here:
            Roberta works closely with FTDNA . Note that unnamed variants aka singletons are included in the matching criteria. She explains the logic behind the matching algorithm.


            • #7
              I've seen conflicting claims on just how detailed the BigY is, or if it ventures into areas where the probability of failing to get an EXACT match on yourself becomes more than slightly probable. Although with as many people as they've tested at this point, even without going "all in" they have almost certainly encountered many people who have accomplished just that.

              Do note: Exact Matching is not to be confused with "not matching" entirely. Generally speaking while that inexact match might happen, the probability of it not being "identifiably" you, and not a close sibling for example, is going to be on par with being struck by a (identifiable) falling meteoroid from orbit, not just once, but multiple times.