Archiver > GENEALOGY-DNA > 2012-03 > 1333051203

From: (John Chandler)
Subject: Re: [DNA] Zhivotovsky's effective mutation rates - was Pressrelease: Genographic Project study of Afghanistan
Date: Thu, 29 Mar 2012 16:00:03 -0400
References: <><0bb101cd0d43$58869f80$0993de80$@org><20874C8256A446D8A7CB42263BC814D8@elizabethod>
In-Reply-To: <20874C8256A446D8A7CB42263BC814D8@elizabethod>

Elizabeth wrote:
> My own layman's viewpoint has always been to wonder how such unknowable
> factors like bottle-necks, back mutations, etc. can ever be adequately
> compensated for

That "etc." is exactly the difficulty. I'll point out in passing that
back mutations are automatically accounted for in the variance method,
and true bottlenecks are quite rare (since total extinction is the
usual outcome of a steep decline). However, variable fecundity, whether
systematic or random, introduces an unknown distortion into any
statistical method based solely on the sampling of the current
population. In other words, the "coalescence time" is necessarily a
biased estimate of the TMRCA -- the bias direction is known, but the
amount is not.

> I expect that the academic community are continuing to use Zhivotovsky's
> rates in their papers because these have been peer-review published and the
> pedigree rates that the citizen scientists use have not.

That's not really an adequate summary. Zhivotovsky et al. postulated
a uniform and constant fudge factor to be applied to the measured
pedigree rates (which had not yet been adequately measured for many
markers). The people who apply the fudge factor nowadays tend to do
so using subsequently published pedigree- or father-son-based rates.
As such, there is no such thing anymore as "Zhivotovsky rates" --
just the fudge factor.

> My understanding is that Zhivotovsky was studying isolated populations with
> known MRCA's, and his use of pedigree rates calculated times far too recent
> to be accurate,

Actually, the populations didn't start from a single ancestor, but
rather from substantial starting populations at a known (or assumed)
time. The study made various assumption about the unknown starting
diversity of those populations.

> For Zhivotovsky to be wrong, there must be
> something wrong with his method and conclusions. Can anyone tell me what is
> incorrect about it that makes you all so sure that he is indeed completely
> inaccurate?

Basically, the problem is in the proposition that a single fudge
factor can adequately compensate for the unknown demographic history
of any population under study. Z et al. hammered on three different
case studies until they got all three to give roughly the same fudge
factor, but they made no attempt to explain why human reproduction
would always average out the same way. In particular, all three
cases are rather recent, and so the extrapolation to pre-agricultural
times is reckless, to say the least.

When you get down to actual cases, it is possible to apply sanity
tests to the various age calculations. The coalescence time is an
underestimate of the TMRCA, but the interclade variance age of the
next "upstream" node is necessarily an upper bound. If the Z factor
goes beyond the interclade limit (and I think it does in various
cases reported on this list), then you know the "one-size-fits-all"
approach is flawed.

John Chandler

This thread: