GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2011-06 > 1308595099


From: "Kenneth Nordtvedt" <>
Subject: [DNA] oops in: frequency of SNPs
Date: Mon, 20 Jun 2011 12:38:19 -0600
References: <015901cc2d69$fa3783e0$eea68ba0$@dgmweb.net><SNT139-w65D338D1D98E031C57FD10C86F0@phx.gbl><004701cc2efa$5bda8110$138f8330$@dgmweb.net><F6CBB475CBF3452EAE36ACBB26F7C1F1@kenPC><004f01cc2f03$6020c730$20625590$@dgmweb.net><F33C42C3046B4D80A0ED2B5EDD57D386@kenPC><4DFF4F13.1020803@mail.utexas.edu><CE651443B22247EC9D39F9F84068187F@kenPC><BANLkTikAbeteVKdcfv7bcavBrDP8wFaZbg@mail.gmail.com><55F97A3CD32A4360BEC0446991A68984@kenPC><BANLkTinSqeeNmvk5e0p51VLvF-rDY08Jug@mail.gmail.com><30546A860FCD47F78B4D71269B9A2335@kenPC>
In-Reply-To: <30546A860FCD47F78B4D71269B9A2335@kenPC>


-----Original Message-----
From: Kenneth Nordtvedt


-----Original Message-----
From: Keith Britton

If you treat the population as "monotonically rising", it's not true that
"the bulk of the sum over n(g) comes from the recent generations".

[[ I took care to change your use of actual historic population to the "y
tree population". The latter is the relevant population and necessarily
must be monotonically growing or flat. How would you possibly construct a
male line of descent tree starting from a common ancestor and leading to
today's population? {{{{ oops, I left out the key phrase ".... and which
was decreasing anytime in population?" }}}}}}

During most of the history since the male MRCA, the actual male population
was factors of hundreds or thousands larger than the y tree population which
contains only the males whose pure male lines of descent did not go extinct.
Extinction of male lines has been extremely close to 100 percent until
"recently", and even now it is probably pretty high.

If I am correct and the y tree had a population of only a few thousand as
recent as 7 or 8 thousand years ago, and it has a population of billions
today, I stand by the assessment that sum over n(g) is dominated by the
recent generations.

KN ]]

More to
the point though, doesn't the severe non-linearity of general population
growth differentially affect fast and slow marker statistics?

[[You might be able to invent some statistics so affected. But I bet they
are rarely used in any serious manner. But variance growth between two
mileposts on a branch route through the tree --- no! Fast versus slow
mutation rate only changes the slope of the linear relationship: Var = mG.
The development of variance along a branch line of the tree does not care
how bushy or lean the overall tree is being during the generations --- it is
just a function of the mutations during the passing of the batons from
fathers to sons on the branch line of interest. That's probably why we
build so many useful statistics from this basic branch line variance.

But modification of the mutation rules of a marker, making them dependent on
repeat value, differing in up rate versus down mutation rate, etc., will
alter the linear relationship between variance and generational elapsed time
along a branch line. KN]]



SNPs and other way points giving
intermediate MCRAs ought to provide tractable subsets for at least
comparative analysis by statistical methods. Comment?

[[If I understand this sentence as you meant it, yes, snps can help you
isolate sub-populations of haplotypes so as to better understand features
unique to their history since occurrence of their unique snp tags,
excluding from considerations things happening including noise on other
branches of the y tree. KN ]]







-------------------------------
To unsubscribe from the list, please send an email to
with the word 'unsubscribe' without the
quotes in the subject and the body of the message

This thread: