Archiver > GENEALOGY-DNA > 2010-02 > 1266290473

From: "Anatole Klyosov" <>
Subject: Re: [DNA] TMRCA assessments
Date: Mon, 15 Feb 2010 22:21:16 -0500
References: <>

>From: Robert Stafford <>
>In its genealogical usage, a parallel mutation is the same mutation
>observed in two different lines in the same family. They are very
>significant in genealogical testing, since people often assume that the two
>belong to the same branch in the absence of a paper trail to their common
>ancestor or might question their genealogy.

Dear Robert,

I still do not get a thing about a significance of "parallel mutations" and
why to single them out compared to other mutations. Then, I still do not get
why folks here typically mention "back and parallel mutations" in one
sentence. What is common between them? To me, "parallel mutations" are just
trivial mutations, as any other mutations. It can happen any time anywhere.
In fact, I view "parallel mutations" as independent events, which might
occur in very different time frames. And someone looks at them, and - "look!
Parallel mutations!!" Big deal. One has happened two thousand years ago,
another in the past generation. Are they still "parallel mutations"?

I do not see a clarification of that matter in your definition above. You
said - "the same mutation observed in two different lines in the same
family". So what? There are VERY many mutations that happen in two different
lines in the same family, many of them happen in different loci, some of
them happen in the same locus but in two different lines. Is that what makes
them "parallel"?

The difference between "parallel mutations" and back mutations is a
fundamental one, in my view. "Parallel mutations" are trivial mutations, can
happen in any generation, and should be counted as anything else. Back
mutations take a long time to make a noticeable contribution, because they
are in a way similar - statistically - with two-step mutations. First a
"forward" mutation should happen, it takes time. It takes - on average - as
many as 45 generations for one mutation to occur in the 12-marker haplotype.
Then that mutation has a "choice" - either to mutate back, and we are not
going to see it anymore, or to mutate "forward" again, resulting in two
mutational steps. In other words, those two-steps mutations in a haplotype
dataset can serve as a readout of how many back mutations could have
happened. That is why the "quadratic" method does not need a correction for
back mutations.

Now, if you take a haplotype dataset with a TMRCA of 26 generations or less,
you are not going to see many of those two-steps mutations. If they are (in
a large dataset), their contribution in the total mutation count is
negligible This is another way to explain why a contribution of back
mutations is minimal (next to nothing) during the first 26 generations, and
more, up to 40-60-80 generations they still will be within a margin of
error. David mistakenly assumed (using a "simple maths") that 50% of all
mutations should be back mutations (at least I understood him that way), but
it is incorrect. In the same vein, John suggested that in his series he saw,
or "assumed", or "believed" a significant contribution of back mutations,
however, there were no many double mutations in his dataset. Actually, there
were two, as I remember, but they belonged to a different branch. Are they
considered as "parallel mutations", I wonder??

On the contrary, as I see it, "parallel mutations" are very common.

Andrew wrote: "But, why do you make a distinction between parallel mutations
and back mutations? Or is this a misunderstanding? Aren't you just saying
that you
ignore both?"

Amazing. Our friend continues to confuse issues non-stop. First, I make a
distinction between them because they two are completely different things,
as I explained above. Second, how can I ignore back mutations when they -
for ancient series - contribute 20%, 30%, 50% to the final, corrected TMRCA.
I ignore back mutations for recent TMRCAs, because my handy table of
corrections for back mutations does not call for any correction for the
first 26 (or so) generations. Third, I ignore "parallel mutations" in a
sense that I count them as all other mutations when in the same branch. How
can I ignore mutations if they are mutations in a dataset?

>(Robert) The main relevance to the discussion is that someone apparently
>told David that they were rare in a genealogical time span. This is a
>common misconception among those who have not crunched the numbers and/or
>do not have many test subjects per family.

I completely agree. Back mutations are almost non-existant in a genealogical
time frame. Other mutations are common, including those what you call
"parallel mutations".

One more question. In some R1a1 series, particularly English and Irish,
there are many DYS388=12 and DYS388=10. Are those "parallel mutations"
according to your definition? (1) Yes, (2) no, (3) perhaps (?)

Thank you

Anatole Klyosov

This thread: