GENEALOGY-DNA-L ArchivesArchiver > GENEALOGY-DNA > 2010-03 > 1268619615
From: "Ken Nordtvedt" <>
Subject: Re: [DNA] Percentages of males and females who reproduce-compare to the "one third rule"
Date: Sun, 14 Mar 2010 20:20:16 -0600
----- Original Message -----
From: "Steven Bird" <>
If so, then you are also saying (and please correct me if I am mistaken)
that population history is irrelevant when calculating a coalescent? That
would suggest that the germ line mutation rate is in fact correct for
estimating the time to the coalescent and there is no need to be concerned
about reduced diversity due to founder's effects. This is assuming of
course that the germ line rate is in fact correctly estimated for each locus
of the Y.
> This is helpful, thanks for your response.
There are three kinds of variances floating around in in use by different
1. Interclade variance between two separate clade haplotype sample
populations whose common ancestry is at an earlier node. (you will use this
interclade variance to estimate the age back to that early node ancestral to
2. Variance of a clade sample population of haplotypes measured from
assumed founding haplotype.
3. Self variance (Coalescence) of a clade sample population of haplotypes.
The first two variances have expected values UNaffected by tree structure
(bottlenecks, subclades, variation of growth rate, etc). Those variance
expected values are only a function of generations to the TMRCA.
The third variance expected value is influenced by all the details of the
tree structure complications (bottlenecks, subclades, variation of growth
rate, etc), because it is an average of the haplotype pair-wise TMRCAs which
is necessarily less than the overall TMRCA. That's why special names of
coalescence age or expansion age are used for what it estimates; it is
something other than TMRCA.
The shape of the distribution of the variance values you will get for all
the three cases, this shape determining the statistical confidence interval
of the age estimates, is definately determined by the population history,
i.e. tree structure in all its gory details.
The final caveat is that you are always estimating the ages of the pruned
tree for the sample of the haplotypes you use. This is not the tree of the
whole population that was sampled. If one's sampling was of quality and
sufficient size, your estimated ages for the tree for the haplotype samples
will be close to the age for the tree of the whole population.