Archiver > GENEALOGY-DNA > 2010-02 > 1266244311

From: Vincent Vizachero <>
Subject: [DNA] Intraclade usefulness (was FW: : variance of S116*)
Date: Mon, 15 Feb 2010 09:31:51 -0500
References: <>
In-Reply-To: <>


Right, though the primary concern with intraclade variance as an
estimator of age is not sample size but rather the underlying
structure of the phylogeny.

If you knew the full details of the tree you could correct the
intraclade variance to get a decent TMRCA estimate. On the other
hand, if you knew the full details of the tree you'd just do an
interclade estimate anyway.

However, intraclade variance is easy to calculate and can be done in
the absence of downstream SNP testing - and therefore cheap -which
makes it a popular option for some studies. And if you have a
haplogroup in a bunch of different populations you should be able to
use intraclade variance to infer relative age among the populations
under a simple condition: that is, as long as the shape of the clade
phylogeny was similar in all the different populations then the
distribution of intraclade variances should be an unbiased reflection
of relative age.


On Feb 15, 2010, at 7:28 AM, Alan R wrote:

> As far as I can see, despite the many issues of the intraclade
> variance method, there is simply no other show in town when
> attempting a within-clade study with a view to establishing the
> origin area and direction of spread of a clade. Other than that you
> just have the phylogeny of the clade defining SNP - what is up and
> downstream of it and distribution maps. However, this method is
> still very vague and there still seems a very blurry (if not
> missing) link between the upstream forms in SW Asia and SE Europe
> and the downstream of P310 forms in the west. I think there is
> little choice but to look at intraclade variance when the aim is the
> finer detail of a clades history. The major concern I have is not
> the method but the sample for each area which (if I am understanding
> this correctly) is very very small. I wonder if it would be better
> given the small number of 67 markers to look at 37 markers too.
> Alan
> Gary said
> Once again, this method as a measure of variance is suspect.
> -------------------------------
> To unsubscribe from the list, please send an email to
> with the word 'unsubscribe' without the quotes in the subject and
> the body of the message

This thread: