GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-02 > 1267219566


From: (John Chandler)
Subject: Re: [DNA] Re Clerical mutations and lab errors
Date: Fri, 26 Feb 2010 16:26:06 -0500
References: <159618.90854.qm@web24507.mail.ird.yahoo.com>
In-Reply-To: <159618.90854.qm@web24507.mail.ird.yahoo.com> (message fromAnders Pålsen on Thu, 25 Feb 2010 23:00:22 -0800(PST))


Anders wrote:
> So could these lab errors contribute to a overestimated mutation
> rate versus Zhiv rates?

The value I gave was for errors that were apparently detected and
corrected within a period of slightly under two years. I have no
obvious way of extrapolating that to a count of undetected errors,
and so the only thing I can logically do is call that 1/800 figure
the error rate itself. Dividing by 67 to get the average per-marker
rate of 1.8e-5 shows that this is negigible compared to the measured
mutation rates of "average" markers.

Note that the error rate in high-resolution genealogical tests is not
necessarily the same as the error rate in the father-son studies that
mostly piggy-backed on commercial paternity testing.

Also note that, in principle, genotyping errors would lead to a biased
(high) estimate of the present-day variance of the populations studied
by Zhivotovsky et al and thus to an overestimate of their mutation
rates as well. In practice, of course, an error rate as low as the
one I found would be negligible for Z et al, too.

John Chandler


This thread: