Re: [tied] Nostratic's New Guinea Home

From: Michal Milewski
Message: 13614
Date: 2002-05-02

----- Original Message -----
From: <x99lynx@...>
 
> First of all, haplotype 87 (unmutated M90) MUST be older than any of the
> other varieties of 09
haplotypes.
 
Yes, this haplotype is older (as it appeared earlier), but it still doesn't mean that the location of modern populations (chromosomes) bearing this hyplotype should be considered as defenitely associated with the location of the initial split (although I admit, that in some cases such situation is more likely, and I will explain it later). Let me show the following example:
 
Let's assume that we have a population with defining X mutation. Let's further assume that this population was moving  at certain rate when splitting in two and a new mutation Y appeared in one of the branches shortly thereafter. Does it mean that the branch without the new Y mutation (the more "ancestral") suddenly stopped moving (and stays in this teritory till today), whereas the other group could move further? Or let's assume that we have a population that didn't move intensively and split in two or three. Of course, the people in those populations couldn't have known which branch had older haplotype (and who should stay at given place forever instead of spreading around). And it is also not very likely that the migration rate is correlated with the mutation rate (at least not at this basic level).
 
The only situation, when you can assume that the haplotype lacking subsequent mutations is more likely associated with the teritory of the split, is when you suspect that the original population was relatively large and not very mobile (of course, it is very hard for large populations to move without splitting) and the new splitting branches were represented by small migrating populations. In this case, the genetic drift could affect the rate at which the new mutations become visible (established/fixed) in different populations (it has nothing to do with the true mutation rate).
 
Let's go back to the real problem we are dealing with, which is the question of 09 haplogroup (its origin and split). You suggested that we should assume that the highest frequency of the founder haplotype should indicate the center of the spread. In this case the center would be New Guinea. I think that this kind of assumption is justified only when we have a reason to suspect that the original population was indeed quite large and stationary. And I don't think it is very likely (though not impossible) that this requirement was met. Firstly, since the sister branches of 09 haplogroup are qite frequent mostly in Near East, Cucasus and Europe, it is not likely that all these branches originated in New Guinea. More likely, the common ancestor of those sister branches lived close to Near East (especially if we remember that his ancestors came from Africa). Secondly, it is also not very likely that the 09 mutation originated not far from the Near East and then moved to New Guinea, where the population became much larger and subsequently split in daughter branches. The route from Near East to New Guinea is quite long, and it's not very likely that such long migration was accomplished by a large population, so the population was rather relatively small during the migration. In this case, we should expect the genetic drift will enable "accumulation" of new mutations, which we don't see. Again, this variant is still possible, but (in my opinion) less likely than for example the Southern Asian center of 09 spread. Thirdly, the New Guinean spread of 09 subbranches would have to assume that people moved from Africa through Near East and Southern Asia to New Guinea, and then spread only nordward reaching Southern Asia, SE Asia, Central Asia, Siberia, and finaly America nad Europe. Again, this cannot be excluded, but if we use the Ockham's razor, we should stick to Southern/Central Asian spread.
 
 
 
>  There are NO sister branches with the 09
> mutation.
 
Of course not, but there are sister branches without 09 mutations (which can be helpful in reconstructing their history).
 
 
> Haplotype 87 IS the phylogenic "root" of the whole group. 
 
Yes, this hyplotype formed the phylogenic root, but at the same time it can form one of the branches in the cladistic tree (if it survives till modern times).
 
 
 
> This
> means that wherever
you find 87, no other form of 09 can be as old.  All
> other 09
"polymorphisms" are subsequent.
>
> So what HAS TO BE the oldest
form of the 09 mutation is found in the greatest > per capita concentration in modern New Guinea.  There is no way around that,
> if Underhill's
cladistic tree is correct.
 
 
I hope you agree with my thesis that your rule applies only to specific situations. And we are not sure that this situation existed in New Guinea. I should also mention that this theory may depend on just one additional mutation defining New Guinean 09 populations that maybe was not yet detected. Would you change the whole scenario in such case?


> How it got there is a
different question.  BUT subsequent 09 mutations in
> Central Asia
or any where else DO NOT COUNT in answering that question,
> BECAUSE they
are subsequent. 
 
I would agree with you if we were certain that 09 mutation spread from New Guinea. But since we can not be certain, such additional data are very important.
 
 
> The assumption is that a new mutation starts
> with ONE and ONLY one individual.  So a vast
amount of people with newer
> forms of 09 mutation does not tell you how
many original Haplotype 87
> unmutated 09's there were originally in the
same area, since you only need
> and you only get one person starting the
spread of the new mutation.
 
 
Right. Note, that the same rule has to be applied when discussing whether a place with low frequency of a "founder haplotype" could be a center of the spread. The spliting/migrating population could preserve the "founder haplotype", whereas ONE and ONLY one mutation in the remaining population could start the transformation into a new branch (assuming that this population was small enough).


> Haplotype 87, the original
unmutated 09, had only one AND ONLY ONE immediate
> predecessor. 
And according to Underhill's phylogenic tree, it was Haplotype
> 71
(m89).  And Haplotype 71 had one and ONLY ONE immediate predecessor - the
> node marked as mutation "186" - which is so old it is given no
haplotype
> number and has no modern occurence.  Take two steps down
from that super-old
> caveman era mutation and you are back at your m09 -
Haplotype 87.  THAT is
> the DIRECT ancestry of the unmutated 09
genotype. This is the ancestral line
> of anchor 09's descent diagrammed
in the Underhill tree.  It is a very, very
> old gene.
 
Yes I agree, and I even dated it's appearing roughly as 35,000-30,000 BC.
 

> So, where was H87(09)'s immediate
ancestor located?  The best answer is the
> current data for Haploid
71 (m89) is very statistically small per capita, but > America has the biggest ratio of these probably statistically insignificant
> numbers.
India et al and Siberia et al show even more minor traces.  The
>
direct grandparent of 87(09) is m186 and it gives no current evidence of
> location or even current existence (according to Underhill.)
>
> This means that the first time we meet a significant modern per capita
> concentration in the Underhill data on grandpa m168, pa m98 or daughter
m09
> is -- believe it or not -- New Guinea.  You can explain the
concentrations of > later 09 mutations as you wish, but if you are going to use them to prove
> origins, then New Guinea sure gives exactly the same
evidence of being m09's
> original origin place.  Or at least being
near it.
 
I think this "being near it" will satisfy me, so maybe we can agree that Southern Asia is a quite possible center (it's almost equaly close to Central and SE Asia, where 09 descendants are common)
 
 
> You really should read Ringe's work on a phylogenic tree for IE to see how
> that has been handled.  I'll
send you a copy when you have the time.  I don't > know anyone who claims that the Baltic languages have the largest number of
> PIE
features. 
 
This was just an example, but what is the consensus as to the more archaic modern IE language? (or at least what is your opinion?)
 

> Finally, it seems a little
inconsistent to find all these origin points in
> other per capita
concentrations and then balk when New Guinea shows up. 
This accusation is a little bit unfair, since I never used the per capita concentrations as the only (or even main) marker of branching centers. I already asked you to cite my statements that would support it.
 
I will have less time to write now, but I will read your comments with pleasure.(I would be especially thankfull for the most possible routes of M173 mutation to Europe).
 
Best wishes
 
Michal