GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-09 > 1031233230


From: "Roy Johnson" <>
Subject: Re: [DNA] Best DNA Test Labs
Date: Thu, 5 Sep 2002 08:40:53 -0500
References: <JCHBN.020904.235418.RC0@CUVMB.CC.COLUMBIA.EDU>


John Chandler wrote:
> > and since a 25 marker Y chromosome test can
> > bring the probability down to 4 generations or less, to indicate which
sub
> > branch would be the best area of research for such persons.
>
> You are hoping for too much here. The 95% confidence interval on the
> number of generations since the most recent common ancestor is still
> almost 40 generations wide for a perfect 25/25 match. To put it
> another way, if you have a collection of ten test subjects from line
> #1, you might expect about 100 transmission events connecting them
> all to the progenitor, meaning 2500 mutation opportunities, meaning
> five mutations in all, and some of those mutations might be quite
> recent. It will take a LARGE number of tests before you can build
> up a general directory of mutations pinpointed to when they occurred.

Information I looked up indicated a perfect match with 25 markers indicated
a 50% probability at 3.6 generations, with the bell curve running down from
there to a 95% chance at 40 generations. Given the nature of a bell curve,
I would think you would get a 60-70% probability pretty quickly, at maybe
5-7 generations. Still, that might not be worth it. Thanks for the advice.

Now, how about some advice regarding the labs? I am looking hard at
FamilyTreeDNA.com right now, since they give a lit of help in coordinating
the whole thing. Do you know if they are any good, or if not, who is best?

Roy Johnson

This thread: