GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-09 > 1128133046


From: (John Chandler)
Subject: Re: [DNA] STR mutation rates
Date: Fri, 30 Sep 2005 22:17:26 -0400 (EDT)
References: <1f9.1337e671.306f2a10@aol.com>
In-Reply-To: <1f9.1337e671.306f2a10@aol.com> (DNACousins@aol.com)


Ann wrote:
> But... isn't the purpose of Charles' log to assemble data about *differences*
> in mutation rates between various Y-lines? A few transmission events isn't
> going to be sufficient to make any statements about that.

Nobody is expecting to find in the log a family study that stands out
so far in terms of mutation rate that it can single-handedly prove a
statistically significant difference. In fact, if such an entry were
to surface, the immediate suspicion would be that the study was using
flawed selection criteria -- either achieving a zero mutation rate
by rejecting as "unrelated" anyone with even one mutation, or achieving
a very high mutation rate by accepting as related anyone who thinks he
has traced his ancestry back to the founder.

Indeed, the "flag" for serious, systematic variation in the mutation
rate, if it shows up, will be a skewed distribution of many small
studies, each of which is plausible as an outlier by itself. The
trick will be to eliminate various types of selection bias as
alternative explanations for such a skewed distribution.

John Chandler


This thread: