DNA-R1B1C7-L Archives

Archiver > DNA-R1B1C7 > 2007-08 > 1186761643


From: "David Ewing" <>
Subject: [DNA-R1B1C7] Network.exe
Date: Fri, 10 Aug 2007 10:00:43 -0600


John, I'm not sure I follow your reasoning and I haven't used this feature
of Network, but let me suggest what I think is a much easier and more
transparent approach. Just calculating a TMRCA without remaining
excruciatingly aware of the confidence intervals is sure to mislead.

I really like the results I get from the little TMRCA calculator based on
Bruce Walsh's method and prepared by B.G. Galbraith, available for free
download at www.clangalbraith.org/DNATesting/MRCA.htm. Rather than giving
you a simple number, it gives you a set of probability distributions
displayed in three different ways.

Another cool little calculator is Ann Turner's Mutation Calculator that will
calculate the probability of finding a specific number of mutations in a
given number of generations. This can be downloaded for free through a link
on www.genetealogy.com/resources/html/cat10.html. There are a number of
other tools available through links on this page as well, but I think you
are probably already familiar with most of them.

As regards your calculations for a MRCA for Ewings and McLaughlins, know
that if we use these calcuators on just the 12 men in our project known to
be descendants of John Ewing of Carnashannagh b. c1648, (ranging between 6
and 9 generations from him), and make the assumption of an average .002
mutation rate per marker, we get results that don't square very well with
what we know based on conventional genealogy.

For example, two of these men are known 6th cousins twice removed (so one of
them 9 generations from John of Carnashannagh and the other 7 generations)
are at genetic distance 6/37 from one another; the calculation suggests that
there is (round numbers) no chance of them having a MRCA w/in the last 100
generations. On the other hand, we have two other men who are known 5th
cousins once removed (one 6, the other 7 generations from John of
Carnashannagh), who are at genetic distance zero (both matching the Ewing
modal exactly), and the calculation more plausibly suggests that there is a
52.3% likelihood that the MRCA is 5 generations or less ago, and a
64.5%likelihood that the MRCA is 7 generations or less ago; however,
the 95%
confidence interval for MRCA is 0.2 to 24.9 generations.

The last time I calculated the observed mutation rate in this kindred, it
came out about .004, but we have added another couple of men since then and
it may be slightly different than this. Most of the men in this kindred are
at genetic distance 2 or 3 from one another. If we use the .004 mutation
rate estimate and genetic distance 2/37, the calculation gives us a median
MRCA at 9.3 generations and a 95% confidence interval of 2.1 to 25.1generations.

Bottom line: be wary of TMRCA calculations.

David Ewing

This thread: