GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-07 > 1183676352


From: "Bob Bootle" <>
Subject: Re: [DNA] R1b1c7 as a calibration tool - Critique ?
Date: Thu, 5 Jul 2007 23:59:12 +0100
In-Reply-To: <REME20070705174221@alum.mit.edu>


John,

Many thanks for a superb response.

The upfront fixable one will be done pronto. Square root it is.

The " population structure " problem, I can see but can't solve.

On " uneven sampling of the population".
Did you mean our uneven sampling of current population ?

Or

Did you mean some uneven sampling process in the earlier population ?

Regards

Bob
.....................................................................
John Wrote :-
If I understand what you are doing, there's an obvious fixable problem
up front. You seem to be assuming the characteristic genetic distance
is growing linearly with time, but that's not so. It grows as the square
root of time.

There is a more subtle and hard-to-fix problem as well. The distribution
of genetic distances within a population is fairly broad after some time
has elapsed, and therefore the mode is very much subject to displacement
by uneven sampling of the population or the presence of structure within
the population.

John Chandler



This thread: