GENEALOGY-DNA-L ArchivesArchiver > GENEALOGY-DNA > 2005-04 > 1114555011
From: "Lowe DNA" <>
Subject: RE: [DNA] Sometimes seeing pink
Date: Tue, 26 Apr 2005 17:37:02 -0500
Ken and David...
I would like to see 50+ markers and SNPs down to the 5th-6th level....
then we might be able to start "pigeon-holing" some of these SURNAMES.
We have Stephens that are matching Hoopers and Vaughans that are matching
on 33 or more on 37 markers and who also match on haplgroup.
For example: See X3TH6 at YSEARCH
We must start ramping up on our DNA tests to narrow down the differences
that we are beginning to see across SURNAMES... Perhaps more markers and
deep level SNP testing.
When YSEARCH gets to 50,000+ entries, we are going to need all the help
we can to "sort the grain from the chaff".
Sent: Tuesday, April 26, 2005 5:09 PM
Subject: Re: [DNA] Sometimes seeing pink
Nothing could be further from the truth (ok, a bit of hyperbole). Clearly
you are not aware of what is going on there in the world of SNP searches -
some very promising discoveries have been made and are now being examined.
You extole the work of Y-STR analysis in relation to geographical stucture.
These markers mutate so while the discoveries you and others make may
pertain to a sizeable percentage of, for example, the R1b 25/11/14 group or
whatever, you will always be faced with the stark reality that some 25/11/14
exist in that form due to convergence and the relationship is coincidence
and the person with that haplotype is in danger of being misled. SNPs do
not carry that risk.
By all means carry on with your work, but you need to publish your data so
that the world beyond this list are aware of your advances in the haplogroup
I arena. I know one geneticist who knows of your work and is following it
with great interest, but you need a wider audience.
-------------- Original message --------------
> Unless there is a technological breakthrough in the search methods for
> the population of the world will more and more be carved up into objective
> varieties with separate histories of descendancy from different founders
> using the techniques of organizing haplotypes of many STRs by their
> structure and geography, with the SNPs falling more and more behind the
> frontiers of interesting research.
> If we could do what you imagine doing with orders of magnitude more SNPs
> strategically located (that is not being private or redundant ones), then
> you are right. With essentially all SNPs that probably exist discovered,
> could probably dispense with STRs.
> ----- Original Message -----
> From: "gareth.henson"
> Sent: Tuesday, April 26, 2005 12:47 PM
> Subject: Re: [DNA] Sometimes seeing pink
> > Ken
> > presumably only results whose haplogroups have been SNP verified can be
> > considered, otherwise there is no evidence for or against convergence.
> > excludes quite a large proportion of available results (are there any
> > for SNP testing available from the testing companies?)
> > At 6 markers you get overlaps between the major haplogroups, at 12 the
> > occasional overlap between subgroups or closely related haplogroups, I
> > suggest that if we could SNP test down to the "few thousand years ago"
> > we would also see overlaps at the 25 marker level.
> > Gareth
> > ----- Original Message -----
> > From: "Ken Nordtvedt"
> > To:
> > Sent: Tuesday, April 26, 2005 4:25 PM
> > Subject: [DNA] Sometimes seeing pink
> > > Often I see pink (not red, this hobby of affection does not warrant
> > red about anything) at word sound bytes use in list messages, in surname
> > projects, and yes, originally from the testing companies --- which
> > misleading sense of the truth. Yes/no statements about people being
> > or "not related", depending on how many mis-matches their haplotypes
> > one thing that triggers my pink. Being "related" is a sliding scale
> > it is not a yes/no thing.
> > >
> > > Another often heard term is "convergence", like in "they only match by
> > convergence", etc. The picture connoted is of two ancient and different
> > haplotypes mutating through the generations to arrive at a common form.
> > >
> > > The better picture for two identical but only very remotely related
> > haplotypes is "lack of divergence" --- not convergence. These haplotypes
> > descend from a common founder, and through the many generations since
> > examples at the extreme statistical low end of the distribution of
> > of mutated differences we expect the haplotypes to have accumulated.
> > are the most likely number of mutated differences just as in the TMRCA
> > curves the testing companies impress on their customers, and there are
> > percent and 95 percent confidence intervals way below and way above the
> > likely. There are the 1 percent aned 99 percent confidence intervals
> > and they are even lower and higher yet in generation number.
> > >
> > > So given a founder for the R1b haplogroup, for example; his descendant
> > population today consists of haplotypes which if picked from random
> > most commonly have 10 out of 25 markers different. But there will be
> > random pairs with more differences and some with less, even if this
> > descendant population has only been shaped by the pure mutation
> process ---
> > no developments of sub-populations due to isolation, etc. As one
> > more and more from the most likely 10 differences between a random pair
> > R1b haplotypes, one eventually reaches the probability of there being
> > mis-match out of 25 or none. These probabilities are very small but not
> > zero. Such pairs match not because of "convergence" but from "lack of
> > divergence". On statistical grounds, given the mutation process, this
> > occur at some frequency.
> > >
> > > Most talk about haplotypes matching because of "convergence" are
> > cases of haplotypes which did not diverge.
> > >
> > > My image of "convergence" is a pair of haplotypes --- one from
> > A and the other from Haplogroup B which has a decent modal difference
> > (reflecting the difference of their founder haplotypes). If the
> > being looked at are short --- the 12 basic markers, for example --- then
> > two haplotypes may have mutated from their original different forms to a
> > common form today. I think this is the scenario we should save the term
> > "convergence" for. There may be a few examples of this genuine
> > among the G, I1b, and J haplotypes of only 12 marker lengths in the
> > databases.
> > >
> > > If the haplotypes were 25 markers long, the chances of "convergence"
> > this sense become so extremely low. For the fun of it, I challenge
> > to produce a case for 25 or longer haplotypes which have converged.
> > Logically, if not practically, they exist; so you have a non-zero chance
> > success. But it is no fair looking after the fact at some haplotypes and
> > selecting out the 25 markers of your choice. The set of markers should
> > one of the commonly sold panel of markers which were chosen by the
> > company and none can be thrown away.
> > >
> > > Ken
> > ==============================
> > Search the US Census Collection. Over 140 million records added in the
> > last 12 months. Largest online collection in the world. Learn more:
> 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: http://www.ancestry.com/s13966/rd.ashx
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: http://www.ancestry.com/s13966/rd.ashx