GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2007-03 > 1174314153


From: "Alfred A. Aburto Jr." <>
Subject: Re: [DNA] Cruciani 2007 dating method and the demic diffusion modelof E3b and J2
Date: Mon, 19 Mar 2007 06:22:33 -0800
References: <BAY105-F34E3B3B699C0EE3993D9A3CC760@phx.gbl>
In-Reply-To: <BAY105-F34E3B3B699C0EE3993D9A3CC760@phx.gbl>


Steven,

> Steven Bird wrote:

> "To estimate the time to the most recent common ancestor (TMRCA) of
> haplogroups we used the seven tetranucleotide loci and applied the
> average square
> distance (ASD) method (Goldstein et al. 1995), where the ancestral
> haplotype was
> assumed to be the haplotype carrying the most frequent allele at each
> microsatellite
> locus. We employed a microsatellite evolutionarily effective mutation
> rate (Zhivotovsky
> et al. 2004). However, since the loci used here and those used by
> Zhivotovsky et al.
> (2004) do not overlap completely, we calculated the microsatellite
> mutation rate as
> follows: we obtained the mean and standard deviation of the
> father-to-son mutation rates
> reported by Gusmao et al. (2005) for the same loci here used, and
> reduced them by a
> factor 3.6 [i.e. the discrepancy between the rate estimate obtained
> from population data
> and that obtained from father-to-son transmissions (Zhivotovsky et al.
> 2004)]. This
> resulted in an evolutionarily effective rate ù = 7.9 x 10-4 (SD = 5.7
> x 10-4 ), a figure that
> was also used in recalculating the E-M215 coalescence age (data from
> Cruciani et al.
> 2004). Recently, Zhivotovsky, Underhill and Feldman (2006), showed
> that reduced loss
> of diversity in an expanding population brings the evolutionarily
> effective rate closer to
> the germ-line rate than in constant-size populations. Thus, in the
> case of expanding
> populations, we used a correction of the 7.9 x 10-4 value, that was
> calculated as follows.
> With reference to figure 2 in Zhivotovsky, Underhill and Feldman
> (2006), the values of
> accumulated variance in 200-300 generations for the scenarios of i) a
> single rate for
> exponential population growth and ii) growth with four distinct
> consecutive rates, were
> compared with the amount accumulated in constant size populations.
> This resulted in
> evolutionarily effective mutation rates decreased of factors 2.4 and
> 2.8, respectively
> (instead of 3.6), i.e. 11.9 x 10-4 (SD = 8.5 x 10-4) and 10.2 x 10-4
> (SD = 7.3 x 10-4),
> which were applied to haplogroups E-V13 and J-M12 found in Europe.
> C.I.s for the
> ASD (and TMRCA) were obtained as follows: mutations on the
> microsatellite genealogy
> were simulated using a Poisson process, in which the total number of
> mutational events
> was calculated based on branch length and assuming that mutations at each
> microsatellite were gamma-distributed with mean and standard deviation
> calculated as
> above. Each mutation increased or decreased allele length by one step
> (each with
> probability 0.5). ASD was then evaluated for the simulated data and
> the whole process
> repeated 1000 times, to quote the central 95% of values. This method
> represents a
> refinement of that by Thomas et al. (1998) and Scozzari et al. (2001),
> as it also takes into
> account heterogeneity of mutation rates across loci. An independent
> dating method (ñ
> statistics; Forster et al. 1996; Saillard et al. 2000) was also used
> to assay how robust the
> time obtained is to choice of method.



[snp,snip]


> I would greatly appreciate any comments concerning any potential
> issues or problems with Cruciani's approach to establishing the TMRCA
> date range.
>
>
> Steve


I don't know if this is relevant, but I noticed that in Supplementary
Table 3 the DYS461 values appear too low by a count of 2. It appears
that the values given for DYS461 are really A7.2 values (which need to
be adjusted upward by a count of 2 to equal the DYS461 definition (Bosch
et. al. (2002), Forensic Science International 125 (2002) 42-51)). I
don't think "allele length" is a factor, but I though I would add this
comment in ...
Al

This thread: