GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-07 > 1184852068


From: "Eric Olson" <>
Subject: Re: [DNA] Why the mutation rate is not same for all markers ?
Date: Thu, 19 Jul 2007 06:34:28 -0700


Ahem.........

It's not that people don't have vocabulary as much as it is definition of
terms within that vocabulary. I suggest the term random is poorly
understood by any except the mathematicians, who have a certain world view.
The same for the term, or concept, of confidence intervals. That is a
concept unknown to traditional genealogy. Preferred is the concept of
provisional proof, that is, accepted as proof until and if new
documentation is found. If I were to construct a family tree based upon
confidence intervals alone it would become ludicrous after a few
generations. The accumulating uncertainties would make it worthless. Yet
that is what seems to be advocated by the proponents of the probabilistic
mathematics of Y-DNA STR mutations, but with the caveat that it is true
only within certain (huge) bounds.

Legitimate Heritage Societies would laugh you right our of their
auditoriums, and rightly so..

I think we are the victims of an advertising blitz by FTDNA and other
commercial interest groups marketing DNA tests. SMGF and National
Geographic excluded. All the hype about finding ancestors through DNA
testing is not proving to be very true in practice. Instead, we have lots
of data for population genetics...which is good. No argument there. But
for personal family trees, don't think there is much utility in it.... The
confidence levels are unacceptably large..

Eric Olson
(running for cover) :-)


> [Original Message]
> From: Vincent Vizachero <>
> To: <>
> Date: 7/19/2007 3:46:35 AM
> Subject: Re: [DNA] Why the mutation rate is not same for all markers ?
>
> The observation that random events are random should come as no
> surprise to anyone paying attention, or to anyone with a vocabulary.
>
> The observation that knowing the frequency of mutations is not
> useful, on the other hand, appears to be provably false. You are in
> effect bemoaning the fact that a screwdriver is not a very good
> hammer, and so is not a useful tool. It may not be good for
> hammering nails, but is quite useful for turning screws.
>
> More to the point: if the goal of genetic genealogical testing is,
> as you contended earlier, to test a hypothesis then knowing the rate
> of mutation is not only useful but is absolutely critical. And it
> matters a great deal whether the rate is 0.003 or 0.0003, so clearly
> some level of precision is required.
>
> If my hypothesis is that two tested men are fourth cousins, then a Y-
> DNA tests only informs that hypothesis if I know the rate of STR
> mutations for the markers tested. Without knowledge about the rate of
> mutations, I can reach no legitimate conclusion.
>
> You make good points at the margins (mutations are not a continuous
> function, pay attention to confidence intervals, etc.), but the
> central contention (that knowing the rate of mutations has "no
> practical use") is unsupportable.
>
> Vince
>
>
> On Jul 19, 2007, at 2:42 AM, Diana Gale Matthiesen wrote:
>
> > No matter how precisely you calculate the frequency of mutations in
> > a large
> > sample, that frequency will not tell you with any useful precision
> > (except in
> > the broadest terms) how closely you are related to someone in
> > genealogical time.
> > To put it mathematically, the confidence intervals (which everyone
> > blithely
> > ignores) are too broad.
>
>
> -------------------------------
> To unsubscribe from the list, please send an email to
with the word 'unsubscribe' without the
quotes in the subject and the body of the message



This thread: