Archiver > GENEALOGY-DNA > 2012-04 > 1334266647

From: Keith Britton <>
Subject: Re: [DNA] Out of Africa
Date: Thu, 12 Apr 2012 17:37:27 -0400
References: <><>
In-Reply-To: <>

The methodology associated with A. Klyosov and the recent
Rozhanskii/Klyosov paper command wide interest and need critical
review. That's off topic here, so deserves its own subject title.
Didier's post is directly related, in that PHYLIP use is specified by
Anatole and in the paper, but theory, precautions and limitations for
STR analysis seem hard to find. Again there is wide interest, so
again another specific subject title seems indicated. Accordingly,
I'm posting a comment to John Chandler under "Rozhanskii/Klyosov Paper
and Methodology", questions and a response to Didier's note under
"PHYLIP Questions".


On Wed, Apr 11, 2012 at 6:46 PM, John Chandler
<> wrote:
> Anatole wrote:
> calculations, since they all embody the same deep-seated flaw --
> you have ignored the fact, established by the father-son DATA, that
> the fundamental time unit in genetic genealogy is the literal
> generation, not the year.  The result of this flaw is that each case
> you take up is hit by a calibration error that is both unknown and
> unrelated to the errors in other cases. ...



... : A. Klyosov is first making a detailed tree, using PHYLIP
software ; this IS the step were, in fact, a distance matrix is made
and the main idea is that between the nodes the timing is best
established by a linear relationship. So, rather than fighting on the
mutation rates the only question I would debate is whether trees with
linear inter-nodes can be settled for very early events. It depends on
factors such as the exact way PHYLIP can handle the rare kind of
data typical of the tree near the root.

In my opinion, the approach promoted by A. Klyosov deserves a wider usage.

This thread: