Archiver > GENEALOGY-DNA > 2008-07 > 1215676601

From: Beth Long <>
Subject: Re: [DNA] calibrating for common ancestor - QUESTION?
Date: Thu, 10 Jul 2008 00:56:41 -0700 (PDT)
In-Reply-To: <001701c8e255$76805ac0$6401a8c0@alfap43400ak>

Hi Anatole,
The reason I asked was that we have a set of (24) R1a people in our Bukovina Hungarian project, and I was curious if you could apply your method to it.
Since I asked the question, I sent you the data, and you were able to give me a reply.
The 24 are a mix of recent and non-recent relationships, as you were able to determine; the two Bereti-s are second cousins, and the two Kozsan-s are fourth cousins.
The two Koka-s are related sometime between 1550 and 1780. I say that because surnames were adopted in this area between about 1550 and 1600, and the Bukovina Koka
left Transylvania for Bukovina about 1780. But there is no actual paper trail.
Likewise with the two Bogos', the two Csoboth-s and (possibly) the two Salamon-s. The second Salamon is in the process of being upgraded to 25 markers.
This leaves only one pair (Pal/Marton) which is a 24/25 match but does not share a surname.
The other ten are "singletons" (so far).
Last night, I got back markers 13-25 on the second Csoboth, which revealed that he had a couple of extra DYS464 (e and f). According to Bennett Greenspan, the U of Arizona has a new process which is better at picking out "extra" DYS464. They are re-running the first Csoboth now to see if he has the same pattern (Bennettt expects that they will).
Thanks for having a look at my R1a excel chart, and for your thoughts on the possible migration patterns. We have about forty surnames left to test, and I will let you know if more R1a turns up (I expect it will)

--- On Wed, 7/9/08, Anatole Klyosov <> wrote:

From: Anatole Klyosov <>
Subject: Re: [DNA] calibrating for common ancestor - QUESTION?
Cc: "Anatole Klyosov" <>
Date: Wednesday, July 9, 2008, 11:23 PM

Beth Long wrote:

>In your example, you used a group of 40 individuals. What would you say is
>the minimum group size for which your calculation method is relevent? I
>assume that there is a higher confidence level as the group size increases?

My response:

It depends. Generally, of course, the larger the group, the higher the
confidence. However, if you have only five 12-marker haplotypes and all of
them are identical to each other, you can safely say that a common ancestor
was a rather recent one, no more than 8-9 generations ago. More so, if you
have only five 37-marker haplotypes, and all of them are identical, you can
safely say that a common ancestor lived no more than 3-4 generations ago.
And you can be damn confident about it, that he did not live, say, 2000
years ago.

If in a group of, say, twenty 12-marker haplotypes as many as 10 haplotypes
are identical to each other (10 base haplotypes), then ln(20/10)/0.024 = 29
generations = 725 years to a common ancestor. You can safely say that it is
not 4000 years and not 2000 years to a common ancestor, Often it is all you
want to know. Hence, a conclusion: your "confidence" is determined by
intention, what actually do you want to know? What question do you adddress?


Anatole Klyosov

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: