GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2005-06 > 1117692332


From: "Andrew and Inge" <>
Subject: RE: [DNA] Middle Eastern ancestral markers on new Euro 1.0 test
Date: Thu, 2 Jun 2005 08:05:32 +0200
In-Reply-To: <REME20050601154559@alum.mit.edu>


Dear John

It could be that either you are not reading me carefully, or I am not
reading you carefully, and/or we are writing badly, so I'll just summarise
the position using a lot of quotes.

Firstly, though it seems to shock you, I did really want to see that maths,
and it does show that it does not confirm what you said - at least the way I
interpret it. All you give is a model (all the numbers need to be filled in)
that as you yourself say, could be solved by making X = 1. In any case it
does not say, as you did at the start, that...

"If there are more males who survive from one surname or haplogroup than
from others, that surname or haplogroup has a selective advantage and will
"soon" come to dominate the population completely to the exclusion of all
others."

(As far as I can see this means, which you deny, that you say that "you say
that any surname which is in a majority can be assumed, at least within
reasonable probability, to have more than chance on its side - that it must
mean that, something, for example the Y DNA of that family, must give better
ability to breed males," or "within any population, being a majority Y-type
(or mt-type or surname) [necessarily means] that there must be a selective
advantage".)

To touch a couple of side points (assumptions in your model and qualitative
versus quantitative differences in survival rates)...

(a) You say that your model needs to specify no other givens such as the
number of generations, but your discussion about it shows that it does
involve, at least in your mind, something about population size and growth.

(b) On the other hand, what you say is a qualitative difference is not one
as far as I can see and I suppose you might have misunderstood what I meant.
(You do seem to give a good example when you say "The extinction of an
individual or a group is exactly the same thing as the extinction of a
species or a genus. The only difference is the number of individuals
involved". Perhaps I misunderstand your point here though. It sounds an odd
statement.) In any case, a qualitative difference is between an apple and a
pear or a random event and a non-random event; while a quantitative
difference is between a big apple and a small apple or a likely random event
and an unlikely random event.

As far as I can see, your model adds up to almost 1 no matter what p values
you give, and agrees better with what I said, and what you originally
disagreed with: "if it is just random loss of male lines [which we are
talking about] th[e]n more common surnames and haplotypes will take LONGER
to die out" and that this longer time needed to expect a reasonable
probability of extinction in a large surname rather than a small one is
"just a probability thing" (i.e. requiring no other explanation such as
natural selection of advantageous traits). But perhaps you actually agreed
with that in the first place?

And just to complete the logic, what we can say from the above is that no
amount of commonness of a particular trait in any animal can ever be
assumed, just on that basis, to be a result of natural selection rather than
random drift. Evolution, which is in fact what you are talking about, is not
teleological and can even lead animals down dead ends paths to extinction.

I also disagree that you have made any case that justifies you saying that,
when it is present, "natural selection ... operates quickly and completely"
(compared to random extinctions). But again, perhaps I do not understand
what this odd sounding statement actually means.

Best Regards
Andrew


This thread: