GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-02 > 1266197633


From: "Alister John Marsh" <>
Subject: Re: [DNA] TMRCA assessments.
Date: Mon, 15 Feb 2010 14:33:53 +1300
References: <mailman.4572.1266045490.2099.genealogy-dna@rootsweb.com><A597BD8569E24290A10B37465204E6B5@anatoldesktop><525460709C6A493EBD3D6BECD514B319@john><REME20100214190716@alum.mit.edu>
In-Reply-To: <REME20100214190716@alum.mit.edu>


John,

You said...
>>>>>>>
You do NOT know that there will be 99 mutation opportunities in a tree
with nine descendants after 11 generations. This is one aspect of the
point that Ken has been making so forcefully. It's unlikely that the
MRCA had nine sons all represented in the sample of testees, and so
the number of mutation opportunities will be reduced from the
theoretical maximum by some amount. This reduction is a problem,
unless you have paper evidence to pin down the exact structure of the
tree.
<<<<<<<

I agree completely.

Interestingly, the suspected ancestor of "most" of the Marshes in my example
I was referring to was born 1596, and had 9 sons!! But as it happens, a
couple of his sons were likely responsible for most of the descendant lines.
But I suspect that perhaps 3 or 4 of the kindred lines in my example had
separate lines which joined the rest several hundred years before 1596.
Some may have had a common ancestor as far back as 1100, so by roughly
assuming a comb shaped tree 330 years old, it may have roughly given an
equivalent number of mutations to the "real tree", lop sided or whatever
whatever it was, which may have been 900 years old, not 330 years old.

If the tree included only 2 lines going back to about the year 1100, that
would be 26 mutation opportunities in each of those lines, ie 52 mutation
opportunities per marker. Having only two much longer lines, significantly
increases the proportion of back mutations likely to have occurred, compared
to parallel mutations.

My example was a somewhat hypothetical example to put some numbers up for
discussion. If the numbers were 50% out, it still indicates a reasonable
probability of parallel or back mutations.

For the record, from known parts of 9 descendant lines in the example tree I
gave, I counted 60 known mutation opportunities in the past 330 years.
However, a good amount of the early years of the tree is unknown, and no
matter how you look at it, there must at the very least be say another 20
mutations to fill in the unknown gaps. If my suspicions that the common
ancestor of the 9 lines was 900 years ago, instead of 330 years ago, then
there would be many more mutation opportunities than 100.

John.





This thread: