GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-08 > 1028411199


From:
Subject: Re: [DNA] Test Variability
Date: Sat, 3 Aug 2002 17:46:39 EDT


In a message dated 8/3/2002 11:56:30 AM Eastern Daylight Time,
writes:


> > Thanks Ann for the instructive calculations. Please more: Ken and I match
> > counts at 21 of 23 loci except:
> > Counts Me Ken
> > DYS 390 23 24
> > DYS 392 11 13
> >
> > From these RG estimated our common ancestor lived around 1690. Which
> factors
> > are instrumental? The two loci that count different or the two count
> > difference in the one locus or are both equally important? What would
> > conclusions have been if Ken's DYS 392 count had been 12 or for that
> matter
> > if his counts for both loci were identical to mine. Is there reading or
> > examples on these calculations?
> >
>
>
> Most mutations are one-step, but there is debate on how to handle two-step
> mutations. The mechanism of mutation is called "replication slippage,"
> which
> doesn't really explain it, just describes it. The idea is that the enzymes
> responsible for replicating a strand of DNA lose track of where they are in
>
> the process, and gain or lose a repeat section (typically 3-5 bases in
> length). Two-step mutations have been observed to occur in father-son pairs
>
> (incidence 1/14), so in that case, it would count as one mutation. However,
>
> it's also possible that two separate mutations occurred at different times,
>
> adding up to two steps.
>
> Bruce Walsh, the population geneticist who consults with Family Tree DNA,
> says you need to take the context into account. If the two-step mutation is
>
> the only one, then it should probably be counted as one. If there are many
> other mutations, then you should count it as two -- that is, the common
> ancestor lived so long ago that there's been plenty of time to have two
> mutations at the same locus. He has created some tables at
> http://nitro.biosci.arizona.edu/ftdna/TMRCA.html -- you could probably use
> the table for 25 markers to get an approximation of your RG test (23 or 24
> markers?).
>
> You will notice that the results are given in generations, not years. That
> is
> another point of controversy -- how many years in a generation? Some
> authors
> use 20 years, which we feel is far too short for the average length in the
> past few hundred years. There is a research paper by Heyer which supports a
>
> longer generation length of 30 or so years -- search the archives for Heyer
>
> if you want the exact numbers.
>
> But any way you count it, you'll notice that FTDNA gives a longer time to
> the
> MRCA than RG. We have concluded that RG is using a higher mutation rate,
> but
> we don't have the details on how they are calculating things.
>

I've been watching this thread for a while and wonder if there isn't a much
more simple explanation. While it is exciting and exotic to believe that one
has found a mutation in an allele between two brothers, there is a much
higher probability of a "false positive reading of 24 for Ken" or a "false
positive reading of 23 for the other brother" or a "clerical error in the
lab for either brother", or a combination of all these possibilities. Have
the tests been repeated for both brothers? This exciting variability may be
a simple matter of quality control in the lab.

Paul Wood

This thread: