No announcement yet.

Apparent nonsense on the Big Tree - FGC23343 subclades

  • Filter
  • Time
  • Show
Clear All
new posts

  • Apparent nonsense on the Big Tree - FGC23343 subclades

    The Big Tree is without a doubt a terrific resource regarding the phylogenetic position of a lot of R1b subclades. But I think I've found some huge errors in the calculation of age distribution for two particular subclades, FGC28370 and FGC23344.

    They don't seem to share any SNPs outside of the FGC23343/2 block, so you'd think the distributions would both have their origin point as the expected age of the youngest SNP within the FGC23343 block--about 563 B.C. per the Big Tree. Given the information in the Mutation matrix, and the novel variants listed on the pages for each individual donor, the MRCA birth year of the FGC28370 people should be about 1247 A.D., which is a little older then previous back-of-the-envelope calcs I have done using STR Haplotype data, but not by much. And all the data on the FGC23344 people should show them having a common ancestor in the last 1 to 3 generations. Instead, here is what the 95% two-sided confidence interval shown by Big Tree:

    FGC28370 - 507 B.C. to 743 A. D.; 50%=196 A.D.
    FGC23344 - 198 A.D. to 1304 A.D.; 50%=828 A.D.

    Again, looking at the mutation matrix, donors average 14 and 12 SNPs within and under these blocks, which is about what you'd expect for them to have the same origin point, plus or minus normal deviation. So the age distribution curves displayed by the Big Tree just make no sense to me.

    Can anyone explain this to me? If I'm getting it wrong, I'm certain a lot of other people are, too. If someone within the FGC28370 block thinks their common ancestor with the most remote donor under the block is more likely to be 196 A.D. rather than 1247 A.D., they're in for a big surprise. Likewise for anyone who thinks the origin of the FGC23344 block extends to only 828 A.D.

    The information on Y Full seems better in some senses, but worse in others, so maybe it's not a complete substitute for what I thought I was getting at the Big Tree.

    Y Full methodology is spelled out more explicitly, but there are fewer donors to reference, resulting in a significantly more recent estimate of the MRCA birth year for FGC28370 (aka FGC28368), although I'd bet the MRCA birth year estimate for FGC23344 (aka FGC19339) is pretty close. Plus it doesn't have the pretty pictures from a fully fleshed out distribution curve, and I like pictures.
    Last edited by benowicz; 29 March 2020, 06:52 PM.