New blogs

Leherensuge was replaced in October 2010 by two new blogs: For what they were... we are and For what we are... they will be. Check them out.

Wednesday, May 12, 2010

Neanderthal genome confirm very old divergence between both species


This is something that I had oversighted on my first reading of
Green-2010:

Assuming an average DNA divergence of 6.5 million years between the human and chimpanzee genomes, this results in a point estimate for the average divergence of Neandertal and modern human autosomal DNA sequences of 825,000 years. We caution that this is only a rough estimate because of the uncertainty about the time of divergence of humans and chimpanzees.
I must say that it is a pretty decent age that correlates well enough with the archaeological record. Acheulean technology, which surely reflects this divergent episode, spread from Africa to Eurasia c. 900,000 years ago or maybe even earlier (1.4 million years in Palestine).

I have mentioned in the past that the actual time of Pan-Homo divergence must be in fact somewhat older, at least 8 million years (ref. Caswell 2008), the difference with the date used by Green is enough to push the Neanderthal-Sapiens divergence to almost 1 million years, right in time to fit with the Acheulean expansion.

I have been advocating for this scenario for some time now and finally it seems that I was right in this. However I still see a lot of people happily talking of a mere 300 Ka, what is simply ridiculous.

This has of course implication for the mtDNA of the Denisova finger, which is doubly old and hence from c. 2 million years ago, from the very spread of H. erectus.

This is important because, against some speculations in the opposite direction, these ages match almost perfectly with the archaeological record: the puzzle fits well and that makes me happy.


The other age estimate is not Green's one

Contradictorily, in a later section they report, using a different method, a divergence age between 270 and 440,000 years. I can only shake my head before this apparent inconsistency. It's worth to read the confusing method they used to reach to this age at the Supporting Online Material (section 14, page 122) because the method produces very strange (and rather very short) age estimates for humans, such as 5000-60,000 for the divergence between Han and Yoruba. I think it's a rather suspicious method that is not better in anything to the point estimate above.

The fact that this procedure was made by co-author Heng Li (from MIT-Harvard, not the main Max Planck Institut team) and that the whole section seems strangely contradictory with the first one (whose SOM article is authored by Green himself - section 10, page 56) makes me think in possible disputes between the authors on which method and result is best, somehow resulting in this patchy double conclusion.

I understand that the estimate of 850 Ka is authored by Green and hence essentially the position of the European team and that the low estimate was introduced in the last minute to satisfy the Americans (after all Science is a US publication and its executive publisher has connections with the MIT).

I wonder which is the ideological or camarilla interest in promoting and even force-feed, as in this case, these low age estimates against all common sense, it is beginning to stink.

14 comments:

Dienekes said...

You have it wrong, autosomal divergence times are not species divergence times.

A good way to see why this is the case is to hypothetically split any population into two. The population divergence starts when you split it, but their genes coalesce to a much earlier time.

For example, modern Europeans have genes that coalesce within tens of thousands of years, but if you colonize an island with modern Europeans the population divergence will start at the time of colonization, but the genes will still coalesce thousands of years ago.

Maju said...

I understand your point but I don't see clear how those divergence times could be kept for hundreds of thousand years periods within small populations: drift should cause a strong tendency to fixation all the time only countered by novel mutations (ex-novo diversification). Otherwise all age estimates should converge to infinity or the very origins of life... and that does not happen, certainly.

Also if it's not "real years" why is it presented as such, even with a responsible notice on the uncertainty of such estimate but depending only on the accuracy of the Pan-Homo divergence age?

Dienekes said...

I understand your point but I don't see clear how those divergence times could be kept for hundreds of thousand years periods within small populations

In a constant population size of size N, the coalescence time at a diploid locus is 2N. So, in a small population of 10,000 individuals, this is about 250,000 years (assuming a generation of 25 years), with individual loci coalescing to either much older or much younger ages than this.

I have no doubt that the authors did their math right, and the two sets of numbers that you think are contradictory represent gene coalescence/population divergence times.

Your ~800 thousand year estimate IS NOT compatible with a split of the Neandertal-modern ancestors at that time.

Maju said...

But the effective population size for European Neanderthals was estimated by this same team in little more than 1500-3500 individuals (consistent with estimates for later H. sapiens population before the Magdalenian expansion), what is a whole order of magnitude smaller and gives a mere 50,000 years of difference or so, a very low margin of error for the figures being managed here.

Even 250,000 years do not make up for the difference between the two dates, which is of 400-500,000 years.

But at the early period of proto-Neanderthal (H. antecessor) migration they must have been even fewer. And founder effect behaves like a bottleneck, leading to fixation even faster.

Also it's clear that the two sections are disconnected, probably because the second bloc by Heng Li was added in the last minute. Read it: it really shocked me how unrelated and contradictory the two sections are: Li's estimate looks totally like a compromise attachment of last minute, without really reviewing the text again before publication to make parts fit. This project is really big academic prestige for whoever gets his/her name attached to it so it does not surprise me too much.

Dienekes said...

But the effective population size for European Neanderthals

Are you predicting population size of the ancestral population of Neandertals 800,000 years ago, by the population size of a bunch of late European Neandertals.

There is nothing contradictory in the two age estimates.

Maju said...

"Are you predicting population size of the ancestral population of Neandertals 800,000 years ago, by the population size of a bunch of late European Neandertals".

Yes. I understand that their population when they had already grown larger brains and optimized their adaptations and even expanded into parts of Asia (though we don't have DNA data on them yet) could not be larger at any moment but smaller if anything. Neanderthals coalesced and expanded in Europe with all likelihood.

"There is nothing contradictory in the two age estimates".

There is nothing connecting them at all. They are two different hunches by two different people not even apparently aware of each other saying right the opposite at different sections of the paper. A big error that probably will be ignored by most.

And they are clearly contradictory. You are just in denial, IMO.

Dienekes said...

And they are clearly contradictory. You are just in denial, IMO.

You are in denial about your lack of understanding of coalescent theory which was quite evident by this passage:

"Assuming an average DNA divergence of 6.5 million years between the human and chimpanzee genomes, this results in a point estimate for the average divergence of Neandertal and modern human autosomal DNA sequences of 825,000 years. We caution that this is only a rough estimate because of the uncertainty about the time of divergence of humans and chimpanzees.I must say that it is a pretty decent age that correlates well enough with the archaeological record. Acheulean technology, which surely reflects this divergent episode, spread from Africa to Eurasia c. 900,000 years ago or maybe even earlier (1.4 million years in Palestine)."

Maju said...

Well, we can agree to disagree. I think that the passage is good and illustrates how aDNA and archaeology can converge.

Kepler said...

Maju, how can you be so sure about Neardenthal populations even "as recently" as 30000 years ago?
Genetic diversity? What about pockets of population we have no traces of?

Maju said...

Can you rephrase that, Kepler? I don't really understand what exactly you're asking or what I said about "30,000 years ago"... Sorry.

Kepler said...

Maju,

If I understand it correctly, you and others think we can already know what the population of Neardenthalers
was based on very small archaeological evidence, perhaps also based on the first genetic tests on skeletons.

I was reading some books by Israel Finkelstein, who specializes on the broad Palestine-Israel region (he is secular, by the way, and not one with a political agenda, unlike some of his colleagues in Jerusalem and the USA). He writes extensively on how they try to calculate populations for the 1500-600 BC period and he owns up those (and his) are still very premature assumptions.

As you know there are quite a lot of sites from that period in Palestine.

If they are so unsure about real population sizes in there, how come others want to be more sure about population sizes for the whole of Europe BEFORE the Ice age came to an end?
Whether based on archaological sites or genes, I doubt we have any real idea just yet.

Maju said...

I can only say that two totally different studies have provided very similar estimates for Neanderthal population in Europe (based on genetic diversity) and on H. sapiens population in UP Europe (based on archaeological data).

You read them and get your own conclusions but I think that such a coincidence is at least remarkable. It can be a coincidence but it's not be a systematic error certainly. And I don't know of any other such estimates, much less ones that contradict them.

Kepler said...

Could you please give me the links to the genetic study on that?

Between current Britan and Norway there was a lot of land that once upon a time was populated by humans...probably before them by Neardenthals. All that is under water. Are we sure that area had the same density?

2 studies seem like too little.
Were they really independent?

Maju said...

Neanderthal mtDNA paper: here (paywall, commented here and here).

Paper on estimates of population of UP humans in Europe here (PDF, freely accessible).

"Were they really independent?"

Yes. They are not just from different authors and disciplines but the conclusion of Briggs was that Neanderthals had smaller numbers than H. sapiens, what doesn't really stand up on light of Bocquet-Appel's older paper, which was surely ignored by Briggs.