GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-09 > 1284041719


From: "Ken Nordtvedt" <>
Subject: Re: [DNA] 2 routes for P312 and U106 ?
Date: Thu, 9 Sep 2010 08:15:19 -0600
References: <N1-eZsWo6u__D@Safe-mail.net><84303C27-22C7-4056-935C-14D1348BA825@vizachero.com>


----- Original Message -----
From: "Vincent Vizachero" <>


> The maps are simply graphic presentations of the Td estimates from
> Myres Table S4.

> http://vizachero.com/ht35/U106.png


The real concern is that the Estonian tail is wagging the global map tail.
If I read some table correctly Estonia is the real variance stand out of the
regional calculations. It was based on 10 haplotypes, two or three of those
haplotypes showing some real oddities. Then we have the black box map
drawing and coloring algorithm which smooths out discrete geographically
connected numbers which themselves are composites (variances) of quite
variable number of underlying haplotypes from different regions. I suspect
the algorithm has no default (a-priori) structure free map which then allows
the discrete inputs of variance from different locations to carry different
weight factors in influencing the map and properly determined from the
sigmas that would be associated with the variance numbers?


So the final product --- a visually potent map --- folds in a lot of
questionable massaging along the way in its production.

One reality check might be to redo the map drawing/coloring algorithm with
Estonia data point excluded.

Since U106 and P312 appear to both have ages very close to their interclade
common node age, that would tend to favor migratory scenarios for the two
descending populations that began fairly close together. Ken




This thread: