Archiver > GENEALOGY-DNA > 2005-03 > 1111190716

From: "Ken Nordtvedt" <>
Subject: Re: [DNA] Finding new USEFUL SNPs?
Date: Fri, 18 Mar 2005 17:05:16 -0700
References: <> <004401c52c10$7a20dc80$eb409145@Ken1> <030901c52c14$17a5ff50$fda1bf43@YOURF8387228BF>

The formulas genetic statisticians have developed to relate spread of repeat
values to age --- diversity or variance are the buzz words --- take the
back (up and down) mutations into account. That is why it is the average
squared distance which is related to age or number of generations. Only for
a short time does this averaged squared distance approximate the average
distance. It soon squishes down to achieve roughly what you are talking

Then there is the issue of repeat patterns getting too long or too short,
and then something about the rate of mutation must change. This may happen
sooner in some STRs than others. A repeat pattern can't go down to zero
repeats! And the biggest repeat we have seen so far is about 40. This may
be a factor for the very oldest of the haplogroups. Does anyone know what
is the most diverse haplogroup population of all which has been driven by
mainly mutation and not splintered into various founder populations by

But for some of the comparisons between father/son rates and applications to
ancient populations, none of the factors you mention seem to explain
anything like the 1/3 factor. Both situations are working with the same
range of repeat values as the other.

----- Original Message -----
From: "Phil Goff" <>
To: <>
Sent: Friday, March 18, 2005 4:42 PM
Subject: Re: [DNA] Finding new USEFUL SNPs?

> Ken,
> I am not a mathematician, but don't up and down mutations ultimately
> the effective mutation rate over very long periods of time?

This thread: