GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-04 > 1175440384


From:
Subject: [DNA] Average Mutation Rate/Generation vs. Average MutationRate/Transmission
Date: Sun, 1 Apr 2007 11:13:04 EDT



Hi Steve,

I estimated Oppenheimer's assumed mutation rate using Capelli's R1b
database, assuming Oppenheimer's R1b-9 haplotype was the origin, and using his date
of 15,600 years ago for their expansion out of the refuge. The average
genetic distance is 1.56. The result was 3.3x10-4 mutations/20 years, fairly close
to your value of 2.6x10-4.

However, using more realistic values for mutation rates for these types of
calculations generally result in estimates which are much too low and don't
agree with other non-DNA sources. I believe one of the sources of error in
these calculations is that the average mutation rate which is being used in
origin dating should not be the same as those determined from father-son
measurements or family data. The latter are "average mutations/transmission (am/t)."
However, the rates which should be used in calculations of origin times
are "average mutations/generation (am/g)." The relationship between the two is
a function of the descendant tree structure from the originator to his
descendent testees. If all the testees have a unique connection back to their
originator, which would require the originator having at least the same number
of sons as the number of testees, then (am/g)/(am/t)=1. For normal trees, the
ratio is less than one. For one small, 5 descendant, 300 year tree in my
project, this ratio = 0.8. For a case where the number of descendents
increases at the same rate per generation, the ratio is about 0.3 over a 30
generation period. Genghis Khan effects would reduce this ratio. Of course, we
could never tell what the exact tree structure is for these time calculations,
but perhaps some sort of deep tree structure could be developed to approximate
some average ratio, so we can get a better handle on these deep time
estimates.

Bruce Baird


In a message dated 3/31/2007 7:52:25 A.M. Central Daylight Time,
writes:

>Digging under the surface my understanding is Oppenheimer used Forster
>et al's (2000) "effective" mutation rates which are 10x lower than
>observed father son rates and Cruciani 2007 used an estimate from
>Zhivostovsky et al 2006 which is ~3X lower. Ignoring which, if any, is
>"better" this single fact alone explains a ~3 fold difference in
>estimated ages from the two sources.
>
>It also highlights a key underlying assumption about the whole dating
>schema used for Oppenheimer's book. Often we can't see the wood for the
>trees, or in this case 534 fact filled pages. It took me some time and
>leads from others before I stumbled on this tucked away as a reference
>listed on page 440. It is only on reading in some detail the reference
>itself that everything fell into place about the age estimates. To quote
>from the abstract of Forster et al 2000 "we estimate a mutation rate of
>2.6x10-4 mutations/20 years for slowly mutating Y STRs, ~10-fold slower
>than the published average pedigree rate".







************************************** See what's free at http://www.aol.com.

This thread: