DNA-R1B1C7-L ArchivesArchiver > DNA-R1B1C7 > 2012-01 > 1325874650
From: Bernard Morgan <>
Subject: Re: [R-M222] analysis
Date: Fri, 6 Jan 2012 18:30:50 +0000
References: <email@example.com>,,,<SNT128-W6485989739F5F4A04CE66EBB940@phx.gbl>,, , <000001cccb7f$007a4540$016ecfc0$@com>, ,<SNT128-W65685B15B0EECB820BB181BB940@phx.gbl>, ,<000001cccc72$601b4640$2051d2c0$@com>,<SNT128-W31B47488651E68332695E4BB950@phx.gbl>,<000001cccc9c$3a1e4ad0$ae5ae070$@com>
He offers three method under heading "Base Haloptypes Remaining After t Generation" based on the assumption that a certain frequency/percentage of the original common ancestor haplotype will remain unmutated in later generations.
The common ancestor haplotype will be the most frequent haplotype within his ancestors, others having randomly mutated away from it.
Now the frequency of this base ancestor haplotype can be used to compute the number of time/generation it would have taken to shink down to this level of frequency/percentage.
His offers three methods to do this:
Logarithmic method, i.e. same as the Poisson distribution equation with the occurance of the event at zero (i.e. no mutation from base ancestor haplotype).
All three are basically are Poisson distribution.
I cannot promise that it will work, for the first problem Anatole identifies is the danger that the tested ancestors do share a common ancestor and we have instead created a phantom base ancestor. However we can double check the assumed base ancestor haplotype by do a separate mutation based calculation based on it.
My interest is in the possiblity of identifing base ancestor haplotypes for surname founders.
> Date: Fri, 6 Jan 2012 17:54:24 +0000
> Subject: Re: [R-M222] analysis
> I may have read it, but if so, it was about three years ago. I do remember
> him writing about first-order kinetics.
> [He refers to a logarithmic approach where transition of the base haplotypes
> into mutated ones is described by first-order kinetics ln(B/A)=kt, which is
> a Poisson distribution.]
> The Poisson distribution is a distribution of frequencies. I may be missing
> something, but what is it about ln(B/A) = kt that leads you to the
> conclusion that it is a distribution? Can it be turned into a distribution