GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-11 > 1133233306


From: (John Chandler)
Subject: Re: [DNA] YSNP testing, hobby scientist forces, convergence, and back mutation issues -- to YSNP or not to YSNP
Date: Mon, 28 Nov 2005 22:01:46 -0500 (EST)
References: <IGEOKAGLHNEKPCKPADIGCEHMKOAA.bbailey.lowedna@baileyconnection.com>
In-Reply-To: <IGEOKAGLHNEKPCKPADIGCEHMKOAA.bbailey.lowedna@baileyconnection.com>


Bill wrote:
> Until proven different, I still believe that STR markers can change 1 or 2
> values between father and son; and, then can revert back the next generation
> or two.

Well, let's see. The probability of reverting is certainly nonzero.
The chances of reverting in the next generation or two would be about
1/500. Yes, that can happen, but it's highly unlikely. Did you have
in mind a process that would be less subject to chance?

> All we now have at our disposal for recent comparison are the STRS, when I
> find more difference STR marker differences between a 1st and 2nd cousins
> than 3rd or 4th cousins; or, to another man of the same surname who has no
> recent genealogical lineage affliations with 4 sets of cousins, then I start
> wondering how to explain these findings.

Why do you need to "explain"? Mutations are rare events, but they do
occur. Any time they occur, they are a surprise, but there's no need
to explain them.

> On STPs as I understand, leaving the other million SNPs or so aside, looking
> at the YCC chart going right from M168 to say for example HAP-Q on bottom
> right is a measure of time passed thru SNP creation toward the present date.

Emphatically, not a "measure" of time. The arrangement from left to
right is a time ordering in any one branch, but it is *not* to scale
and cannot be compared with other branches.

> in time we can extend the chart to the right with even more newly discovered
> SNP that are more recent.

Undoubtedly. Beware, however, that new SNPs will increasingly be found
in the middle of the tree as the branches elongate.

John Chandler


This thread: