GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-08 > 1123263328


From: OrinWells <>
Subject: Re: SMGF [was Re: [DNA] FTDNA Genetic Distance Calculations...]
Date: Fri, 05 Aug 2005 10:35:28 -0700
References: <IGEOKAGLHNEKPCKPADIGOEIKKEAA.bbailey.lowedna@baileyconnection.com><REME20050803190044@alum.mit.edu>
In-Reply-To: <REME20050803190044@alum.mit.edu>


There are a couple of flaws in this "science". First, it is just as
possible for a single marker to mutate two steps as it is for it to mutate
a single step. So when two individuals differ on a single marker, without
supporting evidence (a descendent with only one step descended from a node
before the two step mutation occured) you can not really be certain what
you are seeing. It is also possible that each descendent has a single step
mutation away from the REAL original DNA value. For example it may have
originally been 14. In sample A you see a 15 and in sample B you see a
16. If you assume that this is a result of a two step mutation without
supporting evidence you would again be wrong. Now it is probably pretty
rare for either of these to happen normally, but it is still a finite
possibility.

Further, it is also possible for an individual to experience two mutations
on different markers. Or for a mutation at one node and a further mutation
on a second node as a particular branch progresses. We have several
examples of this in our study. We also have one example where we know that
an individual had two mutations, however one of them was not passed on. We
would not have known this if we hadn't coincidentally tested individuals in
three generations in this family.

Now THIS introduces an interesting possibility. Most discussions on
mutations have been predicated on the belief that once a marker mutates
that this mutation will be passed on from there on out. This may not be
true. We really have no studies to show this. You would need three
generations and quite a sizable sample to see if this is indeed true. We
know this happens. Why does it happen? How often does it happen? When we
see these mutations would it be useful to find out if these are simply
transient mutations? That is they only appear in one generation. If this
were more prevalent than anyone suspects, what would it do to the MRCA
calculations?

At 04:03 PM 8/3/2005, John Chandler wrote:
>To be a little more explicit, suppose we are comparing two 25-marker
>haplotypes, and 21 markers agree while 4 differ by one step each. We
>would call this a 21/25 "match" (or "near-match"), and the genetic
>distance is 4. In a second example, let's suppose that the same 4
>markers differ by two steps each instead, while the other 21 markers
>still agree. This is, therefore, still a 21/25 match, but the genetic
>distance is now 8.


This thread: