GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2012-02 > 1328124951


From: (John Chandler)
Subject: Re: [DNA] Question regarding Markers DYS459 and DYS464
Date: Wed, 1 Feb 2012 14:35:51 -0500
References: <20120129.111054.14832.0@webmail07.vgs.untd.com><810DF391F3E8473880DEDB1C182B329A@kenPC><254E7666-5C2B-4D8C-B30C-DFB6BAE6AEFB@earthlink.net>
In-Reply-To: <254E7666-5C2B-4D8C-B30C-DFB6BAE6AEFB@earthlink.net> (messagefrom Bill on Tue, 31 Jan 2012 09:36:52 -0500)


Bill wrote:
> Multi step mutations could be the result of either RecLOH mutations,
> or of several single point mutations occuring in different ancestors.

You need a more precise nomenclature for this discussion. The word
"mutation" is commonly used to mean two quite different things:
(A) any discrepancy between two haplotypes that are mostly the same, or
(B) an event whereby a marker changes from one generation to the next.

Meaning (A) is imprecise and should be avoided in a technical
discussion. Hence, the above statement needs to be rephrased:

Multi-step discrepancies can be the result not only of direct
multi-step mutations, but also of RecLOH mutations, or of several
single-step mutations occuring in different ancestors.

It is important to note that direct multi-step mutations have been
observed in father-son studies, and they form a non-negligible
portion of the total inventory.

> I've noticed that multistep mutations are rarely found when there is
> only a single marker that has changed, and the frequency of multistep
> mutations increases as the total number of changed markers increases.
>
> Does that imply that most multistep mutations are due to a series of
> single point mutations, and that RecLOH's are relatively uncommon?

This is two, or even three, separate questions. The unstated question
hiding behind the imprecise nomenclature is whether multi-step mutations
are the major source of multi-step discrepancies. The answer is: it
depends on the number of generations of separation between the two
haplotypes being compared. The frequency of pairs of coincident mutations
goes as the square of the mutation rate times the square of the number of
generations. The frequency of two-step mutations is very likely a fraction
of the single-step mutation rate depending mostly on the length of the
repeat unit. Thus, two-step mutations are likely to dominate in closely
related individuals, but coincident pairs will be more and more frequent
as the separation grows (and, obviously, as the observed number of
discrepancies grows). For single-copy markers, that's the whole story,
since recLOH events don't affect them. For multi-copy markers, the
recLOH rate probably depends on the palindrome and perhaps on the
position within the palindrome, but it should bear a constant ratio to
the direct two-step mutation rate. However, the effect of recLOH
depends on the starting difference between the copies of the marker
-- if that difference is zero or one, the result won't be a multi-step
discrepancy after all. In general, though, recLOH and direct multi-step
mutations should have the same relative frequency for all separations.

> FTDNA makes adjustments in the calculation of genetic distance based
> on the rates at which different markers mutate. If RecLOH's are
> easily recognized, I'd think they'd account for this as well. Do they?

Yes and no. The only adjustments made by FTDNA are to multi-copy markers,
but they don't adjust *all* of the multi-copy markers, and they don't
make the extra adjustment for the possibly double effect on DYS464.

John Chandler

This thread: