Archiver > GENEALOGY-DNA > 2005-06 > 1119876709

From: Robert Stafford <>
Subject: Re: [DNA] HELP Please - I've Got a Headache
Date: Mon, 27 Jun 2005 05:51:49 -0700 (PDT)
In-Reply-To: <000801c57ac4$b6fb5330$80004b0c@cicadaneu0kf7n>

I would start with the assumption that common
mutations on markers with rates over .004 are
independent, unless a well-documented paper trail
indicates otherwise. However, we are dealing with
ballpark figures, so it doesn't matter too much. We
don't know the rates for 607, 576, 570 and CDY.
However, with the possible exception of 607, they seem
to be very fast.

With that assumption, I count 14 independent mutations
in your gray group. 464 is sometimes a likely tricky,
but you don't seem to have any problems. The modified
birthday solution gives a probability of about 73%
that there are one or more parallel mutations on any

Using a more complex calculation, I get approximately
a 4.5% chance that there could be three coincident
mutations on the same marker. Four would be

We have two mutations in parallel at 390 and three at
439 in our project. We have about the same total
number of mutations. Ours, fortunately, occurred when
the documentation was good. The families were widely
separated at the time of the mutations, so
non-paternity events are extremely unlikely as an

Bob Stafford

--- Sue Waite-Langley <> wrote:

> OK all you experts on the list. Please help
> me...I've stared at my
> descendency chart for two months and cannot come up
> with a good working
> hypothesis of what's going on with my line. I need
> some unjaundiced eyes
> and people without a vested interest in their family
> lines being changed
> ;o)
> Please visit
> to view my
> results. I'm concerned right now with the light blue
> lines at the bottom
> of the chart.
> Please visit
> to
> see the descendency chart.
> The cross hatching represents samples with a
> DYS437=16. All other
> samples have DYS437=17.
> The turquoise color represents a CDYb=37. All other
> samples have
> CDYb=36.
> Of the 12 samples received, 7 are exact 25-marker
> matches (DYS437=17)
> Of the 12 samples received, 1 is a 24/25 match with
> a mutation in DYS464
> but (DYS437=17)
> Of the 12 samples received, 4 have the DYS437=16
> marker
> So on the face of it...I'm working under the
> assumption that the DYS437
> ancestral value =17.
>, look at my descendency see my
> problem!
> I've been back and forward and around and over the
> paper lines of these
> men. I cannot find any errors or gross leaps of
> faith.
> I need to get more subjects...but can't get my
> update written to send
> out with the request. No matter how I look at
> it...IF the lines are
> correct, I have three independent parallel mutations
> at DYS437. I doubt
> it! So, some of the lines must be wrong...but even
> working with this
> hypothesis...I can't make sense of what might be
> going on. The presence
> of the CDYb mutation which also appears in the same
> family group (we are
> trying to raise funds to have all the kits upgraded
> to 37-markers) seems
> to also indicate that these men belong in the same
> portion of the
> tree...but again if the lines were right I'd have to
> be looking at
> additional parallel mutations. I know this can't be
> what's going on.
> I also know that there must be a problem somewhere
> in the paper trails
> that I haven't found yet. But, how do I explain all
> of this to my test
> members so that they can understand it and not get
> angry about the
> possibility that their paper trails are wrong?
> Specific Questions -
> 1. Does anyone have a good feeling for what the
> actual mutation rate of
> DYS437 might be?
> 2. What is the minimum number of possible bad lines
> that might be here
> (everyone can't be wrong;o)
> 3. What in your opinion should my hypothetical
> ancestral haplotype be
> with the samples I have to work with?
> Thanks so much in advance for your help.
> Sue
> ==============================
> New! Family Tree Maker 2005. Build your tree and
> search for your ancestors at the same time. Share
> your tree with family and friends. Learn more:

This thread: