Re: For Sean Pitman: Some real nested hierarchies

On Mar 14, 9:19 am, Seanpit <seanpitnos...@naturalselection.
0catch.com> wrote:
On Mar 13, 7:44 pm, John Harshman <jharshman.diespam...@xxxxxxxxxxx>
wrote:



If you took the person that was 22 substitutions different from a
Neanderthal and compared that person to every other human being, are
you really suggesting that you would be unlikely to find any other
human being with more than 22 differences to this person? I find this
notion of yours to be statistically unlikely given a Poisson
distribution of a few billion people.

Distribution of what? And why would it be Poisson?

I believe he's claiming that sequence differences are randomly and
independently sprinkled among individuals. Apparently he doesn't know
that mitochondrial genomes form a single linkage group.

Mitochondrial differences are indeed randomly produced in individual
generation lines.

Ok, I see what you have in mind. To be more explicit, let's imagine a
progenitor individual mitochondrian M, and it's collection of
descendants at some fixed number g of generations later. It's a
reasonable approximation that the differences between M and the
descendants at distance g have a Poisson distribution.

But why would the distribution remain Poisson if you consider all
descendants in the tree rather than only those at a fixed generation,
or more importantly if you measure the distance not from M but from
some other individual in the family tree?

That is why there would be a Poisson distribution
of differences relative to any one person ...

That is certainly false -- just think of any person who is an outlier
relative to the other people.

... with a range of more than 22
regardless of the person chosen as a reference relative to all other
people in the world.

And that is simply fabricated. Remember, the distance of 24 arose as a
maximum from the set of distances between any *pair* of people.
There's no mathematical reason that the distribution of pairwise
differences (albeit from a limited sample) would bear any relationship
to the distributions of distances (even across the entire population)
from a particular individual. Nor, as best as I can imagine the
evolutionary situation is your claim likely in practice. The modern
human mitochondrian (call it N) that is only 20 differences from a
Neandertal isn't the end result of a long sequence of mutational
wanderings which happened to bring it close to the Neandertal mtDNA
population; more plausibly it is a primitive sequence which never
wandered very from the human-Neandertal common ancestor. The
descendants of that ancestral sequence wandered in various directions
through mutation, hence the model that suggests itself to me puts N
near the middle of the human descendants.

.

Relevant Pages

  • Re: Howard Hersheys Challenge of Sean Pitmans Assumptions
    ... Poisson distribution describes the probability of exactly k events ... likely to be found within a fixed distance." ...
    (talk.origins)
  • Re: nearest neighbour of a point
    ... sets is such that points at a certain distance are concentrated and on ... having non-negative co-ordinates in each of the dimensions. ... all I'm saying is that the number of points in distribution X & Y ... regular grid embedded in that high-D space, and it may be a grid of ...
    (comp.graphics.algorithms)
  • Re: Planar Linkage Problem
    ... uniform distribution on [0,2pi) ... then of course the distance between endpoints is just 1. ... because there's zero probability that distance ...
    (sci.math)
  • Re: Delaunay distances of random 2D poitn distribution
    ... distribution with respect to area on a flat surface of area A. ... > the point is not within distance r of the boundary. ... What you propose is the distribution of distances to the one NEAREST neighbor. ... This can be explained by the fact that in the Delaunay model, we are not only counting the nearest neighbor, but a few of the closest neighbors. ...
    (sci.math)
  • Re: 1c+1c Closing Velocity...,answer to Henri Wilson
    ... Paul Andersen's hot air engine has it all. ... The distribution drawn below is wrong, ... >>This time is proportional to the distance. ... >>the star, but not very much, it will be of this order ...
    (sci.physics.relativity)