Hello List,
"By George, I think I've got it!" This regarding the Carpenter Y-DNA
testing.
So with out messing with it ... here it is. See below.
John R. Carpenter
La Mesa CA
----- Original Message -----
From: "John F. Chandler" <JCHBN(a)CUVMB.CC.COLUMBIA.EDU>
To: "John R. Carpenter" <jrcrin001(a)cox.net>
Sent: Saturday, November 09, 2002 8:16 PM
Subject: Re: [CARPENTER] Re: DNA-Y testing
John,
You wrote:
> Let me see if I get this straight - the variance of close (first, second
and
> third) cousins should be almost nil in a difference in the Y-DNA
25
marker
> tests and zero for the 12 marker test.
Right. The math is fairly simple: count the generations from one person
back to the common ancestor and then forward to the other person (e.g.,
for 1st cousins it would be 4); then multiply by the number of markers
tested (either 12 or 25 or whatever); then multiply by the average
mutation rate (nominally 0.002). That's the "expected" number of
discrepancies between the two testees. It's not exactly zero for any
case at all, but it's pretty small for the cases above. For 1st cousins
and 12 markers, that comes to 0.1. This translates into a set of
probabilities that the actual number of discrepancies will be 0, 1, 2,
or whatever.
> Using a baseline of well documented descendants of the "Founders" of the
> different Carpenter Branches (MA/RI/PA) in North America - we use these
DNA
> samples to reconstruct the relationship (if any) to a common
ancestor.
I just want to emphasize here that you can't skip the step of
reconstructing the DNA pattern of each known founder -- which means
several judiciously chosen and well documented descendants of each
must be tested. You want to be quite certain of the pattern as of
the 1600's before trying to reconstruct the 1500's.
> Those with less well documented ancestry to the "Founders" that match or
> closely match the baseline have a higher probability of being related to
the
> same common ancestor. This will allow them more confidence in
claiming a
> relationship to the "Founders."
Of course, if two or more of the founders turn out to match each other,
then the unattached who in turn match *both* founders will still have
to do some head-scratching. Then again, I will not be surprised if
the three Carpenter lines do what my three BLANCHARD lines did --
come out with thousands of years of separation time from any
common ancestor. You just can't predict these things for most
surnames.
> Since most of us are about 11 or 12 generations from the "Founders" on
our
> different Carpenter branches and assuming the
"Founders" (especially
William
> & William of the MA/RI Branches) are related within 3 or 4
generations.
> Also assuming that the PA branch (Samuel) are related to them about 2 or
3
> generations back ... This leads us to 16 to 19 generations back
to a
> "common ancestor."
That's why you want to pin down the founders first.
> FAQs - "Using our new 25 marker test the 50% likelihood drops to 7
> generations and the 90% to 19.8 generations."
Don't get tangled up with those calculations of common ancestor
probabilities. That figure of 19.8 generations is the 90% limit
for the case where two 25-marker patterns match *exactly*, but only
when you have *no* *other* information about the test subjects.
This is the other reason why you want to pin down the founders --
we have various amounts of genealogical information linking people
all the way back to the founders, so we can get virtually 100%
certainty back that far, provided we get agreement between the
paper trails and the DNA evidence.
> By comparing Carpenters that claim a common ancestor:
> We may be able to determine IF all or some have a common male ancestor
some
> 30 generations (plus or minus a bit) back.
Yes, we may.
> All of this is dependent on getting those living male Carpenters with
good
> documentation to the "Founders" tested. Ideally two
samples from each
> descendant son of a "Founder."
Good target, but even just one testee from each son of a founder would
make a convincing result.
> While this may take some time, it will be invaluable to many others.
When
> this project is concluded it should give us
"definitive" results one way
or
> the other.
Yes, indeed.
> This is excluding random Y-DNA matches.
Well, that could be a problem if two or more founders turn out to fit
the "Atlantic Modal Haplotype" (the most common DNA pattern in Europe),
Random matches are a real headache for the folks with AMH. We'll have
to see what happens.
> Do I understand this better now?
Yes, I think you've got it.
John