PIE vs. Proto-World (Proto-Language) (fwd)

[Dr. R. McMahon]

[ Moderator's note:
  Dr. McMahon is not a subscriber to the list, so CC: to his address if you
  wish him to see your response.
  --rma ]

Rob McMahon (PhD)
Clinical Scientist
Box 158, Level 1
Addenbrooke's Hospital
Cambridge.

> ---------- Forwarded message ----------
> Date: Sat, 31 Jul 1999 11:00:11 +0100
> From: petegray <petegray at btinternet.com>
> Reply-To: Indo-European at xkl.com
> To: Indo-European at xkl.com
> Subject: Re: PIE vs. Proto-World (Proto-Language)

Hi,
The  message above was passed to me for interest, and I hope no-one
minds me posting a few comments. Although I'm not working on this stuff
myself, I'm a molecular geneticist with a background in population
genetics and a long standing interest in human diversity.  So the comments
that follow are my own interpretations of a rather diverse database of
information.
Firstly, the study of the genetic patterns underlying human population and
individual diversity is really just in its infancy and there are more
issues of debate in the genetics community than can be easily covered in a
few lines (it really needs someone to sit down and do a full scale review
article, but.....).
Having said that, here are a few points that might be relevant.
1. There are several different sets of genetic data that could be used to
make statements about the relation between human groups, and by inference
the languages spoken by those groups; these split into two types.
	a. the frequency of rare, and not so rare, variants in proteins
and cell surface antigens in different populations, eg the ABO and other
blood group data.
 	b. Molecular data tracing the history of individual stretches of
the DNA that carries the genetic information.

The molecular data is what people have been focusing on in the past few
years as it has the potential to be least affected by very recent changes
in population demographics and interpopulation marriage, and, perhaps most
importantly, can be analysed without reference to a superimposed data
structure.  To make that a bit clearer, if one wants to say anything about
population frequency data (data type'a') one first defines the populations
to be sampled, a process that usually means selecting human samples from a
geographically defined, isolated population or using some marker of
'tribal affinity' (usually language) which could confound the subsequent
analysis and make any conclusions drawn suspect.  Any stretch of DNA
studied at the molecular level will retain a 'memory' of its own history
independent of the population that the current carrier finds himself in.
The question is then merely one of how closely the history of individual
parts of the genome (molecules) reflect the histories of current human
groups.   And obviously, different parts of the genome, eg the
Y-chromosome (determines male sex and is only inherited from father to
son), the mitochondrial DNA ( a small DNA ring passed only from mothers to
their children) and the rest of the chromosomes (which exist as pairs,
one from father and one from mother), could give completely different
stories about the individuals in a population.

It's perhaps encouraging then, that in a broad sense at least, the three
types of molecular data are beginning to give similar patterns, that
don't conflict too badly with those previously obtained from data type a.

2.  The mitochondrial picture, a thumbnail sketch.  To keep things simple
I'll just tell a just-so story without dwelling on the potential problems,
error bars or areas of uncertainty--but remember THIS IS A GROSS
SIMPLIFICATION.

Take a group of individuals from around the World and type their
mitochondrial DNA by sequencing short stretches that are known to undergo
rapid rates of change.  Arranging these into a set of aligned sequences it
can be seen that some changes occured before others.  You can then draw a
'tree' relating the individual molecules, gradually the groups will become
bigger and start to link up at deeper and deeper nodes.  At some point all
of the groups will join up (coalesce) at the 'root' of the tree, that is
all the sequences can be seen to share a common ancestor. The age of this
common ancestor can then be estimated by counting the average number of
changes down the divergent 'branches' and scaling that by estimates of the
rates at which new variation is accumulating per year.  Similar estimates
can be made for each higher order node in the tree and then one can look
at the tips and see if any of the groups correlate with other factors,
such as language and geography.  For example, a particular node might
connect 15 samples and be dated at 12,000 years; if all of these
individuals came from North America one might suggest that North American
had been isolated from other groups for at least that long.....  Using a
bit of hand waving and some pretty good guesstimates, the pattern that
then
emerges is as follows.

a. All human mt Variation originated from a single molecule present in
Africa at about 250-150 thousand years ago.  This was NOT the only
molecule present in the human species at that time, just the lucky one
whose descendants always managed to leave female offspring.

b. There are 7-8 old coalescent groups with ages estimated to be around 80
thousand years ago, and of these only one is represented in significant
numbers outside Africa, MEMBERS OF THIS GROUP ARE ALSO WIDELY DISPERSED
WITHIN AFRICA.

c. There are many later signatures of expansion both within and outside
Africa, some of which could indicate later expansions from a similar group
of African populations to the initial 'big bang' group, and or other
groups. While others indicate recent events such as agriculture.  Most
indigenous human 'Tribes' seem to have coalescent dates that are
relatively recent being in the order of 10-40 thousand years.

So one might be tempted to say that Humans arose and divesified within
Africa.  These population underwent a dramatic expansion about 80 thousand
years ago and representatives from a limited number of these diversified
groups then subsequently moved out from Africa colonising the rest of the
World.  Subsequently, other African groups may also have moved out with
later expansions, while on the other hand there is some indication that
some of the original Asian colonists may have expanded in later periods
back towards Europe and North Africa, particulary around 50-30 thousand
years ago.  However, in purely quantitative terms more genetic variation
exists within Africa, despite this being perhaps the least well studied
genetically of the major continental masses (excepting Australia), than in
the rest of the World.

I hope that this little summary is useful to the ongoing debate, but if
anything requires clarification, I'd be happy to try.

Rob McMahon

> Rick said:
> >. DNA studies obtensibly show that non-African
> > humans seem to go back a single population distinct from Africans.

> "Distinct" from Africans?   Not in my reading of the texts.  I don't wish to
> raise a non-linguistic topic, but your implication - that languages split
> into African and non-African - is a language topic, and we need some
> scientist out there to give us the DNA truth.

> Peter