GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2006-04 > 1146269462


From: (John Chandler)
Subject: Re: [DNA] extension to 59 markers
Date: Fri, 28 Apr 2006 20:11:02 -0400 (EDT)
References: <4452372E.3060308@charter.net>
In-Reply-To: <4452372E.3060308@charter.net> (message from John Lerch on Fri,28 Apr 2006 10:39:26 -0500)


John wrote:
> IOW IMO the calculation
> below shouldn't use just the 22 new ones.

That is true, and Ann pointed out that she did indeed take the first
37 markers into account in setting up the calculation she did. There
is an optimal way to use all the information available, in case anyone
wants to take the trouble of figuring the odds and perhaps making a lot
of money on all the bets flying around. The only problem is that the
average mutation rate for the 22 "new" markers is very uncertain, so
beware of the risks! The method: take the table of probabilities that
a 37/37 match would be separated by 1, 2, 3, etc. transmissions and
just shift it over to start at 11 transmissions (4th cousins once
removed being the closest possible allowed relationship). For each
separation, multiply the likelihood by the probability of getting no
mutations in the 22 new markers (e.g., using Ann's calculator), and
add up all these products. The sum is the total probability of getting
no additional mutations. I leave this as an exercise for the budding
bookmakers out there...

John Chandler


This thread: