Archiver > GENEALOGY-DNA > 2008-11 > 1226528484

From: (John Chandler)
Subject: Re: [DNA] Trying to get a better feel for MRCA calculations
Date: Wed, 12 Nov 2008 17:21:24 -0500
References: <>
In-Reply-To: <> (message fromJonathan Day on Wed, 12 Nov 2008 10:11:06 -0800 (PST))

Jonathan wrote:
> I've finally extended my family tree back to the 1600s and have
> found that there are living pure-male-line descendants of a common
> ancestor from 6 generations back. This would seem like a reasonable
> distance to see some mutations. Provided I can persuade any of them
> to actually do the testing (which is by no means certain), is it
> possible to use that information to "tune" MRCA calculations at all?

That depends entirely on what you mean by "tune".

> (eg: If I can find two people willing to be tested who connect to me
> through this individual, can I reliably infer what his markers would
> have been, and thus perform the calculation with him as starting
> point?

If you can test descendants of three different sons of this individual,
then you can indeed infer his haplotype, unless you are unlucky enough
to have a marker on which no two descendants agree. Even then, if you
have three different, but contiguous, values, you will have a good
probability of being right if you pick the middle value. In principle,
these inferences could break down through parallel mutations along two
lines -- but the probability of that is very low in only 6 generations.

> Or use one other person's data to recalibrate the curve?)

What curve? You will certainly not be able to revise the probability
distribution of TMRCA for a given pair of haplotypes on the basis of
testing a handful of people.

John Chandler

This thread: