[RASEY-DNA] FW: Adding members from other labs

[Diana Gale Matthiesen]

Hi Jim,

These estimates to MRCA (most recent common ancestor) are based on an average
frequency of mutation derived from very large sample sizes.  The problem is
applying that average (which is a *descriptive* statistic, not a natural
constant)  to individual cases.  In my opinion (speaking here as the retired
biologist), it shouldn't be done for genealogical purposes, and I can give you a
parallel example that makes the reason easy to understand.

The average life expectancy in the U.S. (based on a HUGE database of age at
death) is, I believe, currently 77.9 years.  So, how certain are you that you
will die at age 77.9?  Not very, right?  You could die as a child or live to be
90.  Average age at death is merely a descriptive statistic, not a constant of
nature.  Same thing with regard to mutations.  

Mutations are RANDOM events and random is not just "un-even," it is the OPPOSITE
of even.  You cannot treat the frequency of mutations as if it were a natural
contant, especially not in situations like these, where we are talking about a
few dozen generations in a few families.  The sample size is to small.  Just as
you cannot apply an average age at death in a population to an individual
person's life, you cannot apply an average mutation rate in a species to an
individual line, except in the most gross way (e.g., it's soooo improbable that
ten mutations could arise between father and son that you can consider it proof
of non-paternity).  

<end of lecture - whew!>

If you are matching 64/67, I'd guess FTDNA is estimating that you have about an
80% chance of having an MRCA within the past 12 generations.  I wouldn't argue
with that, but keep in mind there is also a 20% chance that the MRCA is not
within 12 generations, and the tails on this probability curve are long, that
is, 12 could be waaay off.  That said, if you are matching this other TYLER line
64/67, I would not dispute that you could have a close common ancestor, possibly
as near as 6 generations and easily 12, but there is absolutely no way to
*calculate* the distance between you or the generation in which the mutations
occur.  The "slop" in the "mutation rate" statistic is *greater* than the
distances you are trying to measure.  

What you need to do, at this point, is determine the most probable haplotype of
your mutual progenitor, and to do that you need to test more cousins.  You see,
what you don't know at this point is which of the mutations you carry is an
actual change.  It may be that the other Tyler carries a haplotype unchanged
from the progenitor, but you have had three mutations in your line.  Or you may
match the progenitor and he has had the three changes.  Or neither of you
matches the progenitor, and you have had two changes and he has had one -- or
you have had one and he has had two.  It's a case of testing as many lines of
descent as you can, then logically deducing when each mutation must have
occurred based on who carries them based on your paper pedigrees.

I have begun to reach this point with my mother's STRAUBs.  We have eight
individuals tested who have a 63/67 or better match.  Two match 67/67, and they
carry the most common value for the group *at every marker*, so have the "modal"
haplotype for the group.  The modal haplotype is considered to be that of the
progenitor, and none of the eight differs from it by more than 2.  Note that
while several individuals in the group have a genetic distance from each other
as high as four, no one differs from the progenitor by more than two.  The
tables and discussion on the STRAUB site may help:

http://dgmweb.net/genealogy/DNA/Straub/StraubDNA-results-HgI.shtml#I1a-AS5 

Please note also that I have begun to construct a "node chart" for this group:

http://dgmweb.net/genealogy/DNA/Straub/StraubDNA-NodeChart-I1a-AS5.shtml 

The node chart doesn't show much at this point because too few cousins have been
tested.  We have only one case where we have identified the individual with a
new mutation (two brothers who differ by one).  It will take a lot more testing,
but eventually we can deduce where existing mutations in the family occurred,
and that can support or refute some paper connections.  This kind of limited DNA
testing is about supporting or falsifying hypotheses of relationship, and the
conclusions are based on logical deductions, not mathematical calculations of
distance.

And, yes, I hope to eventually construct a node chart for RASEY, though it will
likely be years before we have enough individuals tested.  If the TYLER admin
isn't creating one for TYLER, you can create one for yourself, assuming the
lineages of the test subjects are supplied.  And now you know why I harp on
having participants supply their lineages to my projects.

Hope this helps,
Diana




> -----Original Message-----

> From: rasey-dna-bounces at dgmweb.net 

> [mailto:rasey-dna-bounces at dgmweb.net] On Behalf Of Jim Tyler

> Sent: Wednesday, February 28, 2007 11:56 AM

> To: rasey-dna at dgmweb.net

> Subject: Re: [RASEY-DNA] FW: Adding members from other labs

> 

> Hi Diana,

<snip>

> 

> Now a question about DNA, if you don't mind. I had mine analyzed and

> compared to another Tyler of known lineage back to the immigrant Job Tyler.

> In my case I am brick walled at my gr-gr-grandfather. We have a 35-37 match

> and 64-67 match. For the 35-37 match Family Tree suggests the most recent

> common ancestor is within 12 generations. The mutations are at loci 16 and

> 34 where mine are one number less than his (11-12 and 35-36). 

> At locus 55 it is the other way, I am 17 and he is 16. Is there any data 

> that can explain in what generation the mutations took place?

> The reason I chose the person analyzed is that when my gr-gr-grandfather

> died in 1851 he was buried next to a 6 year old Tyler girl whose father is

> an direct ancestor of my comparison.

>

> Any thoughts? Thank you.

> Best regards,

> Jim

>