Archiver > GENEALOGY-DNA > 2003-11 > 1067817595

Subject: Re: [DNA] Marker Mutation Rates
Date: Sun, 2 Nov 2003 19:00:04 -0500 (EST)
References: <>
In-Reply-To: <> (messagefrom David Faux on Sat, 1 Nov 2003 13:14:10 -0800 (PST))

David wrote:
> I hope that someone will step forward and offer a rebuttal to our take
> on this situation.

Actually, you and Ron seem to have very different takes. Ron's
contention is simply that the accepted nominal mutation rate is
way too low, while yours is that the statistical nature of MRCA
calculations makes them too undependable to use seriously.

I think people do tend to treat the MRCA calculation with more
respect than it deserves. I don't know if anybody has actually
adopted the policy you mentioned (drawing a cuttoff line below
23/25 match as a boundary of relatedness), but I'm sure that
fixed cutoff lines do figure in a lot of the analysis that goes
on. This can be a serious mistake if the lines are too tight,
but it can also be a useful triage tool for deciding which
near-matches to scrutinize as possible indications of previously
unknown connections. The statistical nature of the calculation
makes it very natural to place odds on a possible connection, based
on the DNA comparison -- and the same is true when evaluating a
previously believed connection that seems to be denied by the
DNA. If you calculate the probability as being above 99% one way
or the other, then it goes beyond mere guesswork. Just be sure
that the calculation is done correctly.

John Chandler

This thread: