GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-04 > 1019624149


From: Brett Miller <>
Subject: Re: [DNA] Relative Genetics/Ancestry.com results
Date: Tue, 23 Apr 2002 21:55:53 -0700 (PDT)
In-Reply-To: <103.141fe35e.29f5be9c@aol.com>


I knew the mutation rate thing would come up as soon
as RG's results were out.

When I was preparing some of the product related
literature, I asked Diahan at RG what mutation rate
they were using. She said .004x (I can't remember
what the last digit was she told me.) I asked her to
explain why RG was using a rate double that of the
others. I pressed her pretty hard to explain it. I
told her the people on this list were not going to let
the number go unchallenged :-)

In a nutshell, she said that this is what Dr. Woodward
has observed in the many extended family studies he
has done. I do not think anything has been published
by Dr. Woodward to prove this. This is an in-house
number based on research. She explained that because
he does so many tests on large families with well
documented genealogy, he is able to include people
with 3 mutations that might otherwise get tossed
because of too many mutations from the others. After
her explanation I was pretty convinced RG knew what
they were doing.

Also, as I alluded to in an earliler post, I think
much of the studies which use a .002 mutation rate are
forensics driven (one or two generations). It makes
sense to me that comparing the mutation rate of 1000
father-son events, all taking place within about a
decade, is not the same as comparing mutations rates
of 100 10-generation family lines. Same number of
meiosis events, but not over the same time periods.
Just for the sake of argument, lets say mutations
increase during the peak of the 11 year sun-spot
cycle. That would throw off a lot of studies that use
lots of one generation events within a short time
period. It makes sense to me to actually USE
genealogy, as Dr. Woodward does, to find the
"many-generation" mutation rate that will then be
applied back to determine genalogical unknowns.

I hope Dr. Woodward will publish some of his findings
(although I'm not sure how many labs will make the
effort to reproduce his findings).

I don't know if RG has done anthing with the
calculator. I did use your calculator to provide some
sample results since I don't think some of our
customers could figure out how to use it themselves.
I can include your name (Ann) near the results on the
web page if you'd like. Just let me know how you'd
like it to appear. And yes I used .004 as the rate in
those calculations. I actually rounded the RG rate
down, but it doesn't really matter since we are
talking fractional generations at that point.

Brett

--- wrote:
> In a message dated 04/22/02 10:37:30 AM Pacific
> Daylight Time,
> writes:
>
>
> > Ancestry has some nice verbiage associated with
> the test results of the
> > type that visitors to their web site will
> recognize right away. There
> > is one thing that is worth pointing out under the
> heading of "Most
> > Recent Common Ancestor" that raised some questions
> in my mind based on
> > statistics that I have seen earlier. It is:
> >
> > # of Mismatches - 0
> > Median Number of Generations - 3.6
> > Possible Range in 95% Interval - 0.1 to 19.2
> >
> > # of Mismatches - 1
> > Median Number of Generations - 8.9
> > Possible Range in 95% Interval - 1.3 to 29.7
> >
> > # of Mismatches - 2
> > Median Number of Generations - 14.5
> > Possible Range in 95% Interval - 3.4 to 39.3
> >
> > Compared to the numbers that we have seen reported
> here earlier, these
> > are way too small. Assuming that this is
> incorrect, this is a serious
> > error and problem for Ancestry users.
>
> I checked these numbers against my Most Recent
> Common Ancestor (MRCA)
> Calculator
> http://members.aol.com/dnacousins/MRCA.exe. In fact,
> I wonder if
> they are using my calculator. Scott Woodward asked
> for my source code a while
> back, saying he wanted to extend the calculator to a
> higher number of
> mutations. But if Relative Genetics only goes to two
> mutations, perhaps they
> had a harder time extending the calculator than
> they expected. I stopped at
> two. because the equation even for two mutations is
> quite complex.
>
> Anyway, it appears that Relative Genetics is using
> an average mutation rate
> of .004 per locus, while Family Tree DNA is using
> .002. Bruce Walsh has some
> discussion of mutation rates at
>
> http://nitro.biosci.arizona.edu/ftdna/TMRCA.html
>
> The .004 number has been reported by Holtkemper,
> based on a study of sperm.
> If you search our archives for Holtkemper or sperm,
> you can probably find
> some discussions on whether it's appropriate to use
> .004. However, Brett
> intimated in a message a while ago that Relative
> Genetics is using mutation
> rates derived from their in-house database. There's
> been very little reported
> in the scientific literature on mutation rates for
> markers beyond the core
> group that everyone uses.
>
> Ann Turner
> GENEALOGY-DNA List Administrator
>
http://lists.rootsweb.com/index/other/Miscellaneous/GENEALOGY-DNA.html
>
>
>
>
>
> ==============================
> To join Ancestry.com and access our 1.2 billion
> online genealogy records, go to:
>
http://www.ancestry.com/rd/redir.asp?targetid=571&sourceid=1237
>


__________________________________________________
Do You Yahoo!?
Yahoo! Games - play chess, backgammon, pool and more
http://games.yahoo.com/

This thread: