GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2008-07 > 1216197905

From: "Sandy Paterson " <>
Subject: Re: [DNA] calibrating for common ancestor - QUESTION?
Date: Wed, 16 Jul 2008 09:45:05 +0100
References: <mailman.94547.1216063954.8264.genealogy-dna@rootsweb.com><002f01c8e614\$38c79f00\$6401a8c0@alfap43400ak><000001c8e65e\$4668d990\$0201a8c0@owner8151f88a9><9F81AB3A-83E8-4994-A2BB-9DD497B53A96@vizachero.com><REME20080715165538@alum.mit.edu>

Hi John

Population growth comes from the mean number of sons per father which
is an input.

Let's call it m.

So m = 1.06 could be expected to result in a population growth rate
of 6% per generation.

When testing the software, I built in checks for both mutation rates
and population growth rates. By this I mean I calculated 'observed
mutation rates' and 'observed population growth rates' and compared
them to the values targeted.

Even after many simulations, you find that the observed mutation
rates display some statistical variation, but not much, and they
are quite tightly grouped around the targeted rates.

Resultant population growth rates worried me for a while because
they tend on average to be higher than what one would expect from
the mean sons per father used. Not much higher, but there was a
clear bias.

After a while I came to the conclusion that what I was measuring
was the population growth rate of the survivors, which would, on
average be higher that that of the lines that go extinct, and that
the two groups combined would exhibit an average of the target rate.
I haven't worked out a way to verify that though.

For this particular simulation, I did use a much higher population
growth rate than I would normally use. It was an attempt at simulating
particular grouping, they were thought to have been prolific breeders,
hence the high assumption. I think I used m = 1.2

The fact that the resulting growth rate was less than 20% is probably
just statistical variation due in part to the small number of generations
used.

Sandy