GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-01 > 1105568951

From: "Whit Athey" <>
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 17:29:11 -0500

Charles,

See my post of a few days ago (Jan. 2) when you and John started this latest
round. I showed the distribution that you would expect of mutation rates
from 100 surname projects. No one is disputing the fact that if you plot
the average mutation rates from 100 surname projects, you will get a
normally distributed histogram. I noted that for projects similar to mine
and yours, in 100 projects you would expect to see about 1 with no mutations
at all and about one with 10 mutations (with most falling in between). That
much of a "spread" you get from chance alone. It is only if there are more
projects at the extremes than you would expect due to chance alone, or if
there are other significant departures from the predicted distribution, that
the data would support your hypothesis. Here is the predicted distribution
(from chance alone) for 100 projects with the parameters I discussed the
other day:

Number
Mutations No. of Projects
0........... 1
1........... 4
2........... 9
3........... 15
4........... 18
5........... 17
6........... 14
7........... 10
8........... 6
9........... 3
10.......... 1

That would make a nice histogram, and it's all due to chance alone assuming
a constant rate in each project. What is it that you're predicting that is
different from this distribution???

In any event, you have started a log that I hope that many administrators
will contribute to. Soon we should know if there is anything remarkable
about the distribution of mutation rates. I'm sure that you would agree
that the proof is in the pudding.

Whit

-----Original Message-----
From: Charles [mailto:]
Sent: Wednesday, January 12, 2005 4:55 PM
To:
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

John,

The tyranny of the average and statistics. ;-)

It will be a "normal" distribution histogram plot, imo. We may get to the
same average rate for the overall male population but it will show that the
surname project average rate varies dramatically from one surname project to

another. Most of the mutation rate studies done up until now were
studies of a few thousand male father-son pairs or other methods or
analysis of sperm cells, etc. This new method I am proposing will
calculate rates within a male family line (a Y chromosome line) and then
plot those on a histogram chart. We should come out with the same
overall average. But still each family lines can be significantly
different from one another.