Archiver > GENEALOGY-DNA > 2006-08 > 1155403870

From: "Ken Nordtvedt" <>
Subject: Re: [DNA] IMPORTANT Was 34/37 ... now being rechecked
Date: Sat, 12 Aug 2006 11:31:10 -0600
References: <>

Just want to add that points A) and B) are not quotes from me. Different
peoples' email quoting system seems to work differently

----- Original Message -----
From: "Bonnie Schrack" <>
To: <>
Sent: Saturday, August 12, 2006 9:56 AM
Subject: RE: [DNA] IMPORTANT Was 34/37 ... now being rechecked

> Lawrence,
>> > From: Ken Nordtvedt [mailto:
>> <mailto:>]
>> > Educate yourself about the broad
>> > parameters of what is reasonable to expect in the various
>> > haplogroups you come across (if you are a project administrator).
>> This task is, of course, made more difficult by
>> A) Incorrect haplogroup predictions recorded in Ysearch, and from there
>> incorporated into Athey's predictor (if I understand correctly). Hg Q
>> seems
>> to be particularly notorious in this regard.
>> B) Testing of ethnic groups heretofore underrepresented, for which one
>> does
>> not really know what the acceptable ranges of results will be. This is
>> illustrated by customers whose 37-marker haplotypes result in Athey
>> scores
>> below 50.
> I would like to strongly second Gareth's comments, especially as I have
> worked closely with Whit. Point A is not really valid, as first of all,
> Whit is very careful in the data he uses, not to include haplotypes that
> have been falsely attributed to haplogroups. But more to the point, a
> lot of people on this list don't seem to grasp that Q and R are extremely
> closely related cousin haplogroups, and their haplotypes simply don't give
> a lot of distance on which to judge and distinguish between them. There
> will always be a good deal of overlap in these predictions, I would
> expect, although Whit is working on some major enhancements of his
> program.
> As to the second point, it's true that there are major holes in the data
> for many ethnic groups and geographical regions. This is indirectly
> reflected in the under-50 scores on Whit's predictor. Lowish scores would
> indicate a haplotype that's an outlier within our current dataset, which
> does tend to be skewed toward the British Isles and northwestern Europe.
> What the very low (I would say considerably under 50) scores often
> indicate is membership in a haplogroup that Whit has not yet been able to
> include in the predictor, due to a lack of published STR data on that
> haplogroup. There may be a bit of scientific data out there, but it may
> include only 6 markers, for instance. As time goes on, and more extended
> haplotype data is shared or published, he will eagerly add it to his
> database and include those haplogroups in the predictor.
> I certainly support Ken's larger point, that by becoming aware of the
> consistent patterns in STR haplotypes, we can get a fairly good idea of
> what to expect, and what is new and unusual. Not only that, but as Ken
> wrote,
>> From the distributions of the extended haplotypes you can segregate the
>> subgroups.
> Thus we can find significant clusters that can be studied to gain insights
> into the ethnic and geographical structures within haplogroups. Using the
> abundant STR data that's being rapidly generated by the genetic genealogy
> phenomenon, many sub-groupings can be discerned which have not yet
> appeared in the literature.
> Bonnie
> ==============================
> Search Family and Local Histories for stories about your family and the
> areas they lived. Over 85 million names added in the last 12 months.
> Learn more:

This thread: