GENEALOGY-DNA-L ArchivesArchiver > GENEALOGY-DNA > 2006-04 > 1145934866
From: (John Chandler)
Subject: Re: [DNA] Genealogy as we knew it
Date: Mon, 24 Apr 2006 23:14:26 -0400 (EDT)
References: <email@example.com> <1f5c01c665cd$477615a0$022aa8c0@davros> <REME20060424020537@alum.mit.edu> <031801c66770$bb17b910$022aa8c0@davros> <REME20060424124726@alum.mit.edu> <040a01c667e0$a835a120$022aa8c0@davros> <REME20060424175302@alum.mit.edu> <054101c66809$a61e8450$022aa8c0@davros>
In-Reply-To: <054101c66809$a61e8450$022aa8c0@davros> (firstname.lastname@example.org)
> Is it a bias towards higher mutation rates in the modern population?
No, the father-son studies are the ones that sample the modern
population. FTDNA/UofA used a pedigree-based study and should have
been sensitive to some weighted average period of a century or more
ago, and yet they came out with rates that are too high.
> Or do you believe it's just a flawed method of deriving
> mutation rates in general?
Other pedigree-based studies, including my own from a few years ago,
have obtained results consistent with the father-son results.
Therefore, it isn't that the modern mutation rate is *lower* either.
The problem is most likely due to selection bias, as far as I can tell
from the outside. This problem has been discussed on this list off
and on ever since the partial results were announced. The flaw seems
to be that the FTDNA study was announced, and then mostly sat back
waiting for volunteers to come forward with data. Self-selected
volunteers who already know the impact of their contribution cannot be
expected to give an unbiased sample. There are at least two reasons
to expect skewed results from this study: first, the publicity did not
emphasize that data would still be useful even if no mutations were
found; second, projects with more mutations than average would be
sensitized to the importance of understanding mutation rates and would
therefore be more likely to contribute data to the study.
> If it's the latter, I'd like to understand not only why that might be the
> case but also why FTDNA and the UofA lab would think otherwise.
The last public comment from that quarter that I know of was from
Bruce Walsh at the 2005 convention. At that time, he expressed
confidence in the FTDNA study, but he did not know about (or at least
had not seen) the pre-publication release of the Gusmao results, and
he apparently did not take the Dupuy results seriously either.