GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-02 > 1107527584


From: Doug McDonald <>
Subject: Re: [DNA] Extinction Rates for new mutations from a Founder
Date: Fri, 04 Feb 2005 08:33:04 -0600
References: <002501c5088e$f887e140$eb409145@Ken1> <REME20050201201207@alum.mit.edu> <016901c508e3$f512de20$eb409145@Ken1> <REME20050203000042@alum.mit.edu> <42024FD9.4090900@scs.uiuc.edu> <REME20050203193642@alum.mit.edu>
In-Reply-To: <REME20050203193642@alum.mit.edu>


John Chandler wrote:

>
> That's a plausible scenario, but it introduces too many variables.
> Every arbitrary choice in the above scenario is an unknown that must
> be investigated with the Monte Carlo approach if the simulation is to
> satisfy everybody, or even satisfy anybody other than the proposer.


That's incorrect. If a few runs with reasonable parameters shows a wild
difference between "real" and "effective" mutation rates then we know
that you can't use the method to study population times without
a very good population growth and splitting model. That's my
hypothesis: the madel really matters. Once proven by simulation,
its up to somebody else to show that in teh real world all
actual real historical models give the same results.

Doug McDonald

This thread: