GENEALOGY-DNA-L Archives
Archiver > GENEALOGY-DNA > 2005-01 > 1105563935
From: (John Chandler)
Subject: Re: [DNA] Male Line Specific Y-STR Average Mutation Rates -- the one size shoe/(haplotype mutation rate) fits all approach is not valid, imo -- corrected URL link
Date: Wed, 12 Jan 2005 16:05:38 -0500 (EST)
References: <41E5775C.7060302@kerchner.com>
In-Reply-To: <41E5775C.7060302@kerchner.com> (message from Charles on Wed, 12Jan 2005 14:15:40 -0500)
Charles wrote:
> A simple example: All the males in a related Smith family have a Y
> chromosome 37 marker haplotype that mutates at an average rate of .005.
> And the males in a related Jones family have a Y chromosome 37 marker
> haplotype that mutates at an average rate of .001. You test a 100 Smith
> and 100 Jones males and you get an average rate of .003. But is the
> average rate for either family line .003. No it is not.
This simple example has underlying complexities.
FACT: the real-world groups of related test subjects are typically 10,
not 100. The Kerchner group of 9 is a case in point.
FACT: if you take ten groups of ten, all sharing the identical average
37-marker rate of 0.005 applicable to the Smiths, and compute the
*observed* mutation rate for all groups, you will see a wide variety of
results. When I say "wide", I mean it's very likely you will see at
least one of the ten groups come out with an observed rate less than 0.002
and at least one with a rate greater than 0.008. Should these "oddball"
groups jump to the conclusion that their mutation rates are *not* 0.005?
No, because we constructed the example so that their rates are all
*exactly* 0.005.
FACT: if you decide it's time to get more data and pull in 1000 more
such groups of ten, you'll find that the variation is even wider. If
you plot a histogram of the rates seen in the groups, you will indeed
see a normal distribution, and the mode will indeed come close to the
value 0.005. This type of histogram is an indication that one size
does indeed fit all.
FACT: if, instead, you pull in 500 Smith-type groups and 500 Jones-type
groups, the histogram will look very different. It will be bi-modal,
with peaks at 0.005 and 0.001. This will be an earth-shaking discovery
and win fame all around. Good luck.
John Chandler
This thread:
| Re: [DNA] Male Line Specific Y-STR Average Mutation Rates -- the one size shoe/(haplotype mutation rate) fits all approach is not valid, imo -- corrected URL link by (John Chandler) |