GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2003-02 > 1044492515


From: "Lowe DNA" <>
Subject: RE: FW: [DNA] mutations and paper trail
Date: Wed, 5 Feb 2003 18:48:44 -0600
In-Reply-To: <JCHBN.030205.175733.RC0@CUVMB.CC.COLUMBIA.EDU>


John...

Thanks...well spoken

-----Original Message-----
From: John F. Chandler [mailto:]
Sent: Wednesday, February 05, 2003 4:57 PM
To:
Subject: Re: FW: [DNA] mutations and paper trail


Bill wrote:
> In our LOWES, we have documented each descendant and found
> on men with a 1 marker mutation (389-2) out of 12 in the last 142 years
> .not 500 years as predicted...... At this rate we would have about
> 3.52 mutations in 500 years in that time....

You're missing a very important point. There is no single steady rate
of mutation in your study or anywhere else. There is an ACTUAL rate
that happens to be exactly zero most of the time, but it jumped to one
for that particular marker at a particular meiosis in that line. If
you had slipped back in time and tested that father-son pair, you would
have had an "astonishing" average mutation rate of one every 25 years.
The lesson here is that statistics predicts not only that the average
rate you find will be "close" to the true average, but that it will
inevitably be DIFFERENT from the true average. You mustn't be
astonished when that happens, and you mustn't assume that the departure
from the true average will hold up in the long run. You just calmly
calculate the probability of the outcome you see and decide on that
basis whether the paper trail is probably right or probably wrong.

> Dennis is correct....we need a great number of records, maybe 2-3
> million ?, from all of the "H" groups before we can begin to understand
> real-world MRCA's for each of the 25+ NRY markers.

No, it doesn't work that way. With a large database, you can make very
precise statements about the average behavior of the total population,
but a "real-world MRCA" has to be a single instance. You're back to
calculating the probability, given a pair a "real-world" haplotypes
fresh from the lab. We already know what the probabilities are within
a moderate tolerance.

John Chandler


==============================
To join Ancestry.com and access our 1.2 billion online genealogy records, go
to:
http://www.ancestry.com/rd/redir.asp?targetid=571&sourceid=1237




This thread: