GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-12 > 1135445002


From: "Sasson Margaliot" <>
Subject: Re: [DNA] Rules of thumb for mutations in 1000 and 2000 years
Date: Sat, 24 Dec 2005 19:23:22 +0200
References: <BAY5-DAV6C0E895940ACDB357911AB3330@phx.gbl> <REME20051223190445@alum.mit.edu> <000001c6083c$13051d00$9d6660cb@Marsh> <REME20051224012633@alum.mit.edu> <7faf853a0512232338q71c458f7g894f342c1b1d24ca@mail.gmail.com>


Andreas O. wrote:

> Is there a formula that relates:
> 1) The number of markers tested
> 2) The number of mutational steps
> 3) The expected number of generations since the common ancestor
> 4) The mutation rate

If you consider the function R(s, p, t), where

R is mutation rate,
s is an STR marker,
p is a population, and
t is time

there is no reason to expect R to stay constant. In other words, R is not a
variable, it is a function of three variables. Without knowing the function
R(s, p, t), there is simply no way to calculate the number of generations
based on STR data.

Using the average value of R measured today incorporates an assumption that
R was the same in previous generations, which is unwarranted. The "standard
deviations" given in published results reflect only the purely statistical
uncertainty (assuming some particular "average" value of R as a constant),
and ignore the possible variablility of R(s, p, t), rendering meaningless
any attempt to estimate age of SNPs.

Sasson Margaliot


This thread: