September 17, 2009

Balloux in Heredity on Mitochondrial Phylogeography

Francois Balloux has some scathing criticism on mitochondrial phylogeography as it is currently practiced (doi: 10.1038/hdy.2009.122). I recommend reading the whole thing. The beginning:
Let us assume I gave a seminar. I would tell the audience about my latest results on the population history of the pigmy shrew. My findings would be based on a stretch of DNA comprising several metabolic genes, showing no signs of genetic recombination. Armed with sequences from a large number of individuals sampled over a broad geographical area, I would make some inference on the colonization routes and times. To make life easier, I would restrict my analysis to the mutations I liked best, with nice names having been given to related sequences, rather than relying on dull mathematical quantities. As I reach one of the key conclusions of the lecture, which would go as follows: 'It is obvious from the distribution of haplotypes Amanda, Eugenie* and Hector_2alpha that the Outer Hebrides were colonised about 50,000 years ago, this was followed by considerable population fluctuations, a bottleneck during the last Ice Age, a swift recovery and a dramatic recent expansion over the last 200 years and...'. Imagine that, at that climactic stage I was interrupted by someone in the audience. The impertinent would say, 'Sir, can I just ask you whether this confidence in your conclusions may not be misplaced; your analysis is based on a single genetic marker, which comprises genes with a central role in metabolism and is thus likely to have been affected by natural selection'. An awkward silence may ensue, as I would find it difficult to dismiss this criticism easily.
and the end:
Despite mitochondrial sequence variation covarying with climate in humans (Balloux et al., 2009), there are better ways to measure temperature. And, I would argue there are also better genetic markers than mtDNA to infer past population history. I fully appreciate that mtDNA has given us some of the most fundamental results on human evolution at a time when using mtDNA was the only realistic option at hand. I do not question the value of mtDNA in forensics and pedigree reconstruction. It is also likely to remain a valuable tool for inference at a localized geographical scale, particularly when testing specific hypotheses rather than making quantitative inferences on the age or size of the populations studied. It is convenient to type and analyse, and its use in humans raises no serious ethical or societal issue. But all these qualities do not counterbalance the fact that a single locus likely to be under selection is inappropriate for population inference at large geographical scales (or over long periods of time in the context of ancient DNA analysis). We have reached an era in which publicly available data sets of large numbers of complete human genomes are a tangible prospect, and I believe it is now time to move on from the excessive reliance on uniparental markers. Exploiting these new resources of autosomal variation will present significant challenges, but it will not help overcoming them if a large fraction of the community of human population biologists persists in sticking to mtDNA as the marker of choice.
The utility of mtDNA for studying modern populations is indeed limited now that we can study hundreds of thousands of markers per individual. However, it is still a very useful marker for ancient DNA, both because it is often the only game in town because of the relative ease with which it can be typed due to its large copy count, and also because it has proven itself to be capable of generating interesting results, as in the recently discovered discontinuity between Paleolithic and Neolithic Central Europeans, studying the mtDNA diversity of Neandertals compared to humans, or detecting sex-biased gene flow in relatively recently admixed populations.

See some of my previous criticisms on facile correlations between mtDNA time depth and archaeological-historical correlations:
UPDATE:

John Hawks also comments at length on the paper. An excerpt:
So what can we do? Fortunately we have lots of options. We can test the proposed demographic hypotheses against the historical record. When we make observations that show that people 1000 years ago had very different frequencies of common haplotypes, well, we know it was selection. There hasn't been any genetically significant bottleneck in the last 1000 years! When we see small Neolithic population samples dominated by haplotypes that are very rare today, again, no historically possible bottleneck could have caused that.
I am fundamentally in agreement that bottlenecks, so often invoked in the mtDNA literature, are really a non-issue. Consider why this is the case: every mtDNA paper normally takes a random sample of a few tens or hundreds of people from a population that usually numbers in the thousands or millions. The assumption is that such a small random sample generally preserves -within confidence limits- the haplogroup frequencies in the population. But a bottleneck is exactly such a random sample. You can't, at the same time, use a sample of 100 people to infer haplogroup frequencies, and claim that a bottleneck that reduces the population to a 100 people will radically shift haplogroup frequencies. And, of course, there is absolutely no evidence that any major post-Neolithic human population, save for the Andaman Islanders, the Samaritans, or various such populations ever underwent a bottleneck anywhere near that severity.

However, I am in disagreement that a change of haplotype frequencies across 1,000 years is evidence of selection. A different explanation is that of migration, the introduction of a new population element.

Sometimes, migration is easy to infer. For example, we can be fairly certain that modern Europeans are different from Paleolithic Europeans because of Neolithic and post-Neolithic migration into Europe, because there is an introduction of new haplotypes that were absent in the Paleolithic population. One possible explanation is that instead of "absent" we should say "possibly present at very low frequencies". But, once we see that these haplotypes were present on the early Neolithic migrants, it doesn't take much to put 2+2 together and infer that migration is a likelier explanation.

The same process of migration could be inferred for the Neolithic populations of the Lake Baikal district, where a postulated hiatus in occupation, followed by recolonization by immigrants, proposed on archaeological grounds, coincides with the discovery of a sharp difference between pre- and post-hiatus populations in mtDNA haplotype frequencies. Similarly, the absence of Mongoloid mtDNA before the 7th c. BC in Central Asian samples, followed by its introduction after it, can be parsimoniously explained by admixture, since that admixture is evident also in anthropological and autosomal studies.

In other cases, selection may be a more plausible possibility. For example, the reduction in the frequency of haplogroup I in Denmark since the Viking and Iron Age, or changes of frequency in haplogroups in England since the 11th c. AD, such as the reduction of U5a1 and the increase in H may in fact be due to selection. H was present -although not very frequent- in Neolithic farmers from Central Europe, Corded Ware people from Eulau, and its very high present-day frequency in Europeans (roughly 50%) as there is no plausible source or mechanism that would have brought large numbers of it in Europe.

In conclusion, both migration and selection may help explain shifts in haplotype frequencies over time. As we plug in the holes in our knowledge of the mtDNA distribution across space and time, we will be able to decide between the two.

54 comments:

The Other Jim said...

Just that re-occurring comment. Is decoupling your core metabolic pathway (al a Mishmar et al.) actually a "a sensible biological explanation" for heat generation?

John Hawks said...

Thanks for your link and comment -- that was a bit of sloppiness on my part, since some frequency shifts might indeed reflect migration or elite dominance. On the other hand, those hypotheses are among the ones that may be tested historically.

I turn to selection sooner, because it takes an awful lot of migration to make any substantial difference in gene frequency, particularly in the large populations of the last couple of millennia. A very slight amount of selection can cause the same change. But your point is well taken.

--John

Dienekes said...

I turn to selection sooner, because it takes an awful lot of migration to make any substantial difference in gene frequency, particularly in the large populations of the last couple of millennia.

This is generally true for large geographical entities, but such migration-induced gene pool shifts are almost certain at "crossroads" populations (e.g., Ethiopia or Central Asia) in-between genetically differentiated populations (Caucasoids vs. Sub-Saharan Africans and Caucasoids vs. Mongoloids respectively). It is also quite likely in populations that were once culturally/politically distinct but were assimilated into much larger neighboring entities (e.g., Etruscans in Latin-speaking Italy or Phoenicians in the Aramaic and then Arabic-speaking Levant).

Jean said...

...large populations of the last couple of millennia

Here I would agree on the whole, but I feel that prehistoric migration has been seriously underestimated. I say this in more words at Peopling of Europe. There is a supplementary page on migration, which mentions the Central Asian crossroads.

argiedude said...

Your belief is filled with contradictions, but you think you're demic diffusion theory has been proven. It doesn't make any sense that N1a would be mass-replaced by yet another mass migration; it's one thing for hunter-gatherers to be replaced by a mass movement of people, but then that's it. There isn't going to be yet another wave of even more numerous farmers coming in. So we have the problem that N1a was 25% of the farmers but today it's only 0,25%. So you think it's because of natural selection...

Then why is N1a also just 0,25% of the Middle East? Did evolution also affect them? Did the Middle East also suffer a 100-fold reduction of its original haplogroups? And why didn't this evolution occur already between 10,000 and 7,000 BC, before they left the Middle East?

The mtdna of Europe and the Middle East includes half a dozen major haplogroups. If evolution preferred some haplogroup so as to lead to a 100-fold reduction of the original mtdna, then it would almost certainly consist of a single haplogroup, not U, H, T, J, K, etc. The mtdna distribution throughout Europe is incredibly homogenous. Why aren't there differential frequencies of certain haplogroups in the cold north or dry southwest? Different climates, different adaptive advantageous.

Today U5 and U4 make up 10% to 15% of the mtdna of any part of Europe. These are the supposed aboriginal haplogroups. So N1a, which made up 25% of the farmers, has been reduced to virtually nothing, while the aboriginal U4/U5 are still 10%-15% of the modern mtdna? Did they have an advantage over N1a? What a coincidence that would be: apparently the only 2 haplogroups present in the aborigines happened to have an advantage (or evolved one later) that allowed them to survive the evolutionary onslaught that reduced N1a to practically extinction.

Europe's mtdna is virtually identical in every corner of the continent. Under your belief of wildly fluctuating haplogroup frequencies, let me restate, extremely wildly fluctuating, we should expect to find today, obviously, extremely wildly fluctuating haplogroup frequencies. Of course, we don't, it's the exact opposite, every corner of Europe is amazingly homogenous.

The N1a study didn't find a single U4 or U5 in 24 samples. And the recent study found that even the latest group of aborigines were still mostly U4 and U5. So apparently these 2 groups were overlapping everywhere in Europe for 3000 years and yet managed to maintaint themselves completely separate all that time?

The evolutionary advantages that you're talking about would require pretty noticeable advantages. And again, the variety of climates of Europe and their extremely homogenous mtdna composition strongly argues against mtdna haplogroups having any sort of advantages over each other. Which leaves you with no explanation for these bizarre ancient mtdna results. These results don't support your theory. They don't support genetic continuity, either, but you're claiming the former.

Dienekes said...

My guess is that European DNA is mostly of Neolithic origin, and that some lineages which were present in the early Neolithic colonization have been reduced in frequency since that time, due either to climatic adaptation (some places in Europe are much colder than Anatolia and the Near East) or to adaptation to the new diet.

Dienekes said...

Then why is N1a also just 0,25% of the Middle East? Did evolution also affect them? Did the Middle East also suffer a 100-fold reduction of its original haplogroups? And why didn't this evolution occur already between 10,000 and 7,000 BC, before they left the Middle East?

First of all, we don't know where N1a originated in Central Europe. There is no a priori reason to think it came from the very early farmers in the Fertile Crescent. It could have come from Central Anatolia, or the Balkans.

Selection can easily explain its disappearance; the LBK people were at the beginning of the Neolithic in Europe or very close to it. If N1a was indeed at a disadvantage, then it was only beginning to be reduced in frequency.

Dienekes said...

If evolution preferred some haplogroup so as to lead to a 100-fold reduction of the original mtdna, then it would almost certainly consist of a single haplogroup, not U, H, T, J, K, etc.

That is not so. Many different haplogroups may have a similar level of fitness. Some of them (I suspect H to be one such) may have a slight fitness advantage, and some of them (I suspect N1a to be one such) may have a slight fitness disadvantage.

So, what we see today is a basically Neolithic gene pool, where some U- Paleolithic lineages have introgressed, and where some lineages have shifted in frequency either upwards or downwards.

Dienekes said...

And again, the variety of climates of Europe and their extremely homogenous mtdna composition strongly argues against mtdna haplogroups having any sort of advantages over each other.

It is primarily the new Neolithic lifestyle, and later the emergence of civilization, that has driven evolution in Europeans, not climate, except perhaps in the northernmost regions. Homogeneity does not argue against natural selection, since natural selection proceeded in parallel across Europe.

argiedude said...

You didn't really explain anything, you just reiterated what you believe.

"Homogeneity does not argue against natural selection, since natural selection proceeded in parallel across Europe."

You THEORIZE that this happened. You state a lot of things as fact that are speculation.

You want to believe in the demic diffusion theory, I've noticed this for quite a while, I don't know why, and now the results of this study seem to, at first glance, give you the right. But actually, these results are very strange and don't fit with any reasonable explanation.

You think Europe's modern mtdna composition evolved in the last 5000 years. Ok, then why is Egypt's mtdna so similar to Europe (or the Middle East)? Excluding the L lineages (15%), the remaining West Eurasian lineages in Egyptians are close to the rest of Caucasians. Their biggest haplogroup is H, at 22%, J is 7%, K is 4%, T is 28%, and U is 9% (half of it U5). The remaining 30% are N, pre-HV, M1, I, and X, similar to the Middle East. How did Egypt arrive at such a similar mtdna composition as other Caucasians? Did they also evolve into this combination? Is there some ideal combination of haplogroups that populations tend to evolve towards?

Why isn't India's mtdna similar to the Middle East? All it required was a few samples to make it across into India and then the fight-for-the-fittest begins. If H can manage to muscle its way into becoming 20%+ in places as different as Norway and Egypt, then surely it can do the same in India. Or perhaps India's M has an advantage over the Middle East's H? But then why didn't India's M cross over into the Middle East and become just as dominant there as it was in India?

Why is there such a rigid division between black Africa and the Caucasian world? If you believe mtdna haplogroups have different evolutionary advantages then why does black Africa have almost 100% L and the Caucasian world almost 0% L? Are their adaptive advantages equal? Not one of the Eurasian haplogroups managed to find an "ecological" niche in black Africa? Not one mtdna L managed to find its own niche in the Caucasian world? If you're going to believe that haplogroups can compete with each other, then you have to take things to their ultimate consequence. Apply this same reasoning to East Asia, also. H, K, U... all absent (and vice versa in West Eurasia). All it takes is one lineage, then, if it's better adapted, it will become the dominant lineage in due time. Evolution is all you have to explain away the bizarre ancient mtdna results of these 2 studies.

Dienekes said...

"Homogeneity does not argue against natural selection, since natural selection proceeded in parallel across Europe."

You THEORIZE that this happened. You state a lot of things as fact that are speculation.


There is plenty of anthropological evidence for lock-step processes of adaptation across Europe. During the early Neolithic, there was reduction in stature, gracilization, possible decreases in health, exposure to similar pathogens, etc. And, even more recently, the rapid brachycephalization in the Middle Ages, followed by rapid debrachycephalization more recently.

So, there is plenty of evidence for similar selection pressures across Europe. There is also some room for region-specific adaptations, e.g., lactase persistence in northern Europe.

You want to believe in the demic diffusion theory

Both craniometry and genetics support it. Do you have a better explanation why LBK had an Fst greater than 0.15 with Paleolithic Central Europeans if they were not migrants?

You think Europe's modern mtdna composition evolved in the last 5000 years. Ok, then why is Egypt's mtdna so similar to Europe (or the Middle East)?

I don't see why it shouldn't be, as Egyptians are basically Caucasoids.

Why isn't India's mtdna similar to the Middle East? All it required was a few samples to make it across into India and then the fight-for-the-fittest begins.

Obviously the native population of India consists of an M-dominated population into which Western Eurasian elements were added. In order for the Western Eurasian elements to increase in frequency, it would mean that they alone had a selective advantage, while none of the M- subclades did, which is absurd.

Why is there such a rigid division between black Africa and the Caucasian world? If you believe mtdna haplogroups have different evolutionary advantages then why does black Africa have almost 100% L and the Caucasian world almost 0% L? Are their adaptive advantages equal? Not one of the Eurasian haplogroups managed to find an "ecological" niche in black Africa?

There is such a rigid division between black Africa and the Cauasian world because until very recently there has been no intermixture between black Africans and Caucasoids. Intermediate populations have mixed gene pools, in which selection may very well be proceeding in favor of either Caucasoid or Negroid mtDNA types.

Apply this same reasoning to East Asia, also. H, K, U... all absent (and vice versa in West Eurasia). All it takes is one lineage, then, if it's better adapted, it will become the dominant lineage in due time.

You answered your own question: there was no intermixture between Caucasoids and East Asians until fairly late in history, and this occurred in Central Asia and Siberia where the two met.

pconroy said...

I have a theory, involving selection, which could partly explain the phenomenon of increasing frequencies of some mtDNA haplotypes at the expense of others.

We know that there is good synergy between mtDNA H and Y-DNA R1b or R1a, and that children with this combination are more fertile on average.

What if for each mtDNA type there is a Y-DNA type that is most fertile - presumably based on their co-evolution over time in some geographic location. Then you have male mediated gene flow into a population - let's say Steppe warriors - and they carry R1a. Initially they slay the native farmer Y-DNA carriers, and mate with their mtDNA N1c carrying women. Over time though, those few women carrying mtDNA H will become preponderant, due to greater fertility.

In such a scenario, male mediated gene flow, would lead to an initial increase in the immigrant Y-DNA, and a later increase in mtDNA similar to the immigrants.

To test this, we could check if mtDNA H is under positive selection in Latin America, over the last 500 years, due to the inflow of Y-DNA R1b from Iberia.

Comments anyone?

German Dziebel said...

Dienekes: "I am fundamentally in agreement that bottlenecks, so often invoked in the mtDNA literature, are really a non-issue. Consider why this is the case: every mtDNA paper normally takes a random sample of a few tens or hundreds of people from a population that usually numbers in the thousands or millions. The assumption is that such a small random sample generally preserves -within confidence limits- the haplogroup frequencies in the population. But a bottleneck is exactly such a random sample. You can't, at the same time, use a sample of 100 people to infer haplogroup frequencies, and claim that a bottleneck that reduces the population to a 100 people will radically shift haplogroup frequencies."

Nice observation from the field of the social studies of science. Similarly, the first mtDNA studies used Afroamericans as representatives of Sub-Saharan Africa but not America. However, taken at face value, America, since 1492, has accrued the greatest levels of allele diversity in the world. The reason we're not taking this fact as a sign of the great antiquity of man in America is because we know that much of that diversity is derived from Africa, Asia and Europe in recent times. But the blindness of science to the effects of its own activities on the theory it's trying to prove conceals a possibility of the origin of African diversity outside of Africa and its transfer to Africa in several migratory waves from 45,000 BP onward.

Maju said...

I don't get this:

Sometimes, migration is easy to infer. For example, we can be fairly certain that modern Europeans are different from Paleolithic Europeans because of Neolithic and post-Neolithic migration into Europe, because there is an introduction of new haplotypes that were absent in the Paleolithic population. One possible explanation is that instead of "absent" we should say "possibly present at very low frequencies". But, once we see that these haplotypes were present on the early Neolithic migrants, it doesn't take much to put 2+2 together and infer that migration is a likelier explanation.

You have just mentioned a few days ago one of those studies that show that Neolithic Central Europeans (not representative of all Neolithic Europeans anyhow) had different mtDNA than modern Europeans (N1a and stuff like that). So what are you trying to say here?

Also:

In other cases, selection may be a more plausible possibility. For example, the reduction in the frequency of haplogroup I in Denmark since the Viking and Iron Age, or changes of frequency in haplogroups in England since the 11th c. AD, such as the reduction of U5a1 and the increase in H may in fact be due to selection.

Well, it's a possibility but isn't it possible that we're comparing apples and oranges because of too limited sampling? Are those aDNA samples being compared with their modern nearby residents or with much larger but essentially different samples.

We know from Portugal that mtDNA H was absolutely dominant in the Epipaleolithic and also in the Neolithic. This seems also to be the case in Taforalt (UP Morocco), with the additional circumstance that the haplotypes found are in concordance with modern inhabitants of the region (not the wider region of North Africa but specifically the local districts of the Rif). And then you have that CRS haplotype of Gravettian Italy, that should be H*, H1 or at most U, all common haplogroups today in the region and in Europe in general.

So we do have mtDNA H popping up in mtDNA in the Southern Europe and North Africa for before and after the Neolithic. In fact, AFAIK, it also pops up in the samples of the North just that at lower frequencies than today.

If mtDNA H (which in fact are many sublineages) has been under positive selection why has it remained the same, apparently, in Portugal or Morocco? It should have grown to become nearly universal and nope. If anything it may have decreased a bit.

So I don't know how to interpret all this seemingly contradictory data but it sounds to me to that study on Basque Medieval aDNA that, by comparing different localities, decided in their conclusions that the lineages had changed, when in fact, if they would have compared with local modern peoples and not with some random Basque sample from a different region, as they did, they would have noticed striking continuity (of course with some differences as should be the case with any two random samples).

Maju said...

To test this, we could check if mtDNA H is under positive selection in Latin America, over the last 500 years, due to the inflow of Y-DNA R1b from Iberia.

I can tell you from now this is not the case. Most LA mtDNA is Amerindian because very few women migrated from Europe to Latin America. This regardless that autosomal genetics and X-DNA may be in many cases mostly European.

If there would be any truth in your "synergy" hypothesis it would mean shared migration of peoples from where these haplogroups were common in prehistorical times (namely the south) to where they were apparently not (namely the north). This of course doesn't seem to fit with any prehistory I know since Magdalenian (unless you want to force-feed a mass migration with Dolmenic Megalithism or something like that - would not be consistent either).

terryt said...

"I am fundamentally in agreement that bottlenecks, so often invoked in the mtDNA literature, are really a non-issue".

I have noticed that the words 'bottleneck', 'drift', and 'founder effect' usually rise up when the evidence doesn't actually fit a particular theory. Of course the theory itself is never re-examined.

"If evolution preferred some haplogroup so as to lead to a 100-fold reduction of the original mtdna, then it would almost certainly consist of a single haplogroup, not U, H, T, J, K, etc".

The selection involved need not be in the actual mtDNA. It could arise from some cultural advantage possessed by a particular group. In which case several haplogrouops could quite easily share this cultural advantage, and this would lead to their eventual dominance in the population. Nothing to do with the metabolic role of the mtDNA.

argiedude said...

Superb post, Maju.

Dienekes: In other cases, selection may be a more plausible possibility. For example, the reduction in the frequency of haplogroup I in Denmark since the Viking and Iron Age, or changes of frequency in haplogroups in England since the 11th c. AD, such as the reduction of U5a1 and the increase in H may in fact be due to selection.

Maju: Well, it's a possibility but isn't it possible that we're comparing apples and oranges because of too limited sampling?

You're right. The Danish study consisted of 2 samples, I believe. In any modern population, 10% or more of the haplogroups belong to rare and exotic lineages, such as, in the case of Europe, I, L, M, N, X, etc. In a small sample size it's very possible one of these rarer types will be detected.


Maju: We know from Portugal that mtDNA H was absolutely dominant in the Epipaleolithic and also in the Neolithic. This seems also to be the case in Taforalt (UP Morocco), with the additional circumstance that the haplotypes found are in concordance with modern inhabitants of the region (not the wider region of North Africa but specifically the local districts of the Rif).

Thanks very much for pointing this out. I remembered that Morocco study (I thought it was Tarofalt) and I've been trying unsuccesfully to find info on it. Actually, I still can't, even with the name corrected, there are only 4 results in PubMed. I recall there were something like 24 samples, and the notable thing about them was the total lack of haplogroup L, which was interpreted by everyone as proof that mtdna L in North Africa was of historic origin. The recent mtdna results from the ancient inhabitants of the Canary Islands seem to indicate L was possibly present in North Africa from before, as you've also pointed out when the study was released.

But back to H. I would also add that there was a recent study that compared mtdna H in Iberians, North Africans, and Middle Easterners, and found that North African and Iberian mtdna H are much closer to each other than either is to the Middle East. It was very obvious from just glancing at the results table in the study.

I wasn't aware of the ancient mtdna results from Portugal. That's very interesting, to say the least. Can you tell me what study that was?

Maju: In fact, AFAIK, it also pops up in the samples of the North just that at lower frequencies than today.

I also wasn't aware of this. Can you tell what study this is?

Maju: If mtDNA H (which in fact are many sublineages) has been under positive selection why has it remained the same, apparently, in Portugal or Morocco? It should have grown to become nearly universal and nope. If anything it may have decreased a bit.

Excellent observation.

PS: Dienekes, or anyone really, have you taken a look at the FST calculator I included in the comments in "Y chromosome and mtDNA of goats in North Africa"? I think I did a good job in demonstrating that small sample sizes are indeed inadequate to establish relationships between very closely related peoples.

Maju said...

Hey, argiedude: I made a reference post on those two papers and others on mtDNA H (a total of six) at this post at my blog. All links can be found there, though Kèfi's paper is in french (a link to a post at Mathilda's where she patiently translates the essentials is also in that post).

At Chandler's paper on Portuguese aDNA, 5/9 Epipaleolithic samples (from the size of the circles at graph 9, I could not find a table) are H, 2/5 are U, the other two are exotic. The apportions remain largely the same in Early Neolithic populations, just that the exotic haplogroups have vanished, H may have increased its apportion and V appears for the first time, suggesting some level of migration from a more easternly location (probably within Iberia). Neither K, J, T, I, X or W are found in either sample, suggesting that all those lineages arrived later to Portugal and maybe elsewhere in Iberia.

Maju: In fact, AFAIK, it also pops up in the samples of the North just that at lower frequencies than today.

I also wasn't aware of this. Can you tell what study this is?


From Kèfi 2005:

Parmi les populations actuelles du Nord de l'Afrique
étudiées à ce jour, la population berbère du Nord du Maroc
présente des similitudes avec la population ancienne de
Taforalt. En effet, les berbères du Nord du Maroc possèdent
le taux le plus faible d'haplogroupe de type subsaharien
(3,2%). De plus, tous les haplogroupes observés chez
Taforalt sont retrouvés chez cette population actuelle, même
l'haplogroupe J/T. Ce dernier, rarement représenté dans les
populations actuelles, est absent dans les populations nord
africaines à l'exception de la population berbère du Nord
du Maroc (1,6%).
Ces observations conduisent à penser à une continuité
génétique
entre la population ibéromaurusienne de
Taforalt et la population actuelle occupant la même aire
géographique: la population berbère du Nord du Maroc
.

For a translation of the essentials of this paper, read Mathilda's post.

Maju said...

Oops, I think I gave the wrong answer to your question.

When I write AFAIK means "from memory". I have lost my huge collection of bookmarks when my old computer collapsed some months ago but I do recall that N1a only represented some 25% of the Central European Neolithic aDNA, the rest was just "normal", what means lots of H and some U too (but can't recall the details).

Dienekes said...

You're right. The Danish study consisted of 2 samples, I believe. In any modern population, 10% or more of the haplogroups belong to rare and exotic lineages, such as, in the case of Europe, I, L, M, N, X, etc. In a small sample size it's very possible one of these rarer types will be detected.

They didn't find 2 samples of "rare haplogroups", but 2 samples out of 10 of haplogroup I, which occurs at a frequency of less than 2% in the modern population. The difference is significant at the 0.05 level.

The Other Jim said...

The selection involved need not be in the actual mtDNA. It could arise from some cultural advantage possessed by a particular group. In which case several haplogrouops could quite easily share this cultural advantage, and this would lead to their eventual dominance in the population. Nothing to do with the metabolic role of the mtDNA.

Just re-posting terryt's comment, because I feel this is a very important one.

There is no functional evidence that the human mtDNA polymorphisms themselves have biochemical phentoype. I only know of one study trying to address this, and the results did not fit the climate-selection hypothesis. (See Amo and Brand BiochemJ (2007) 404:345; http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pubmedid=17355224).

There are many unusual shifts in haplogroup frequencies in human mtDNA. But these mtDNA may very well have hitchhiked along with other factors, including terryt's cultural innovation idea. The climate hypothesis has some testable prediction on the biochemistry of the mitochondria involved. So far, those predictions do not hold.

In the end, this theory may be correct, and would be very interesting if it is found to be true. But "we" are accepting this hypothesis a bit prematurely.

Anonymous said...

The H mtDNA supposedly found in some skeletal remains of around 28 ky of age is dubious. In fact I consider it to be the result of either fraudulent reporting, not uncommon in science, or contamination by any number of personnel who have handled the remains since being uncovered in some Italian cave. I don't believe as easily as some of you. No matter what aseptic techniques used by researchers, assuming no fraud, it would be impossible to remove prior contamination. Ancient dna studies are dubious at best. Mitochondrial dna H is not that old so others have found, let alone the CRS version of H. The Paleolithic continuation of European population is a difficult one to shake. It is like a religious dogma. I am obviously a heretic.

I am mtDNA V. It is considered along with H1 and H3 as having a European genesis as far as its distribution and age estimation. Big deal. They could have arisen in North Africa where H1, H3 and V are not uncommon. In fact V is more common in some Berber groups than in Europeans. Of course some might suppose it to be the sign of admixture of Europeans with Berbers or their ancestors. The point being that age and frequency prove nothing other than how common a haplogroup is, and its supposed age. Mitochondria are essentially bacteria, and have bacterial dna, and infested eukaryotic cells many millions of years ago. As this infestation is critical to life being the main way ATP is produced in Eukaryotic cells including of humans, not much change can occur in its dna structure as it would effect its functioning and cause devastating diseases. It is likely that mutations do not occur at a high rate in mtDNA, even the junk dna used to find haplogroups. Y chromosome dna is not so critical, or necessitous. It is a good thing your mtDNA mutations are of no real consequence.

Mitochrondrial dna haplogroup U is very ancient but its age does not make it European in origin. It probably arose in the Middle East and seeded Europe with various migrations of humans over a long time span.

When you find a haplogroup in high frequency, something you cannot select for as it is unseen, unknown in the bearers, it is totally due to a form of founder effect and selection of mates for serendipitous reasons other than the bearing of the haplogroup. For example, mtDNA H women could have had wider hips and appeared more fertile than other women just by chance and nothing to do with being mtDNA H. Think of all those Venuses carved by ancient Europeans. Those fat women could have been mtDNA H. Since mtDNA V is a minority haplogroup in Europeans, the women may have been skinny and less fertile looking. I am joking of course. But that hypothesis is as good as any I have seen presented here.

Maju said...

They didn't find 2 samples of "rare haplogroups", but 2 samples out of 10 of haplogroup I, which occurs at a frequency of less than 2% in the modern population. The difference is significant at the 0.05 level.

They were surely relatives, right? And anyhow, while I may be very rare it is a typical haplogroup of Central/Northern Europe if anywhere. I'd really like to know how the sample compares with the local modern populations rather than the whole Danish one.

But in any case it is fully within the probability of any random sample to magnify some rare haplogroups, while others may not show up at all. Also rare haplos could more easily decrease in number because of normal drift, even among agriculturalists. I think this apparent divergence is being hyped within pre-concieved ideas.

On the other hand:

There is no functional evidence that the human mtDNA polymorphisms themselves have biochemical phenotype.

AFAIK, many modernly extant sub-haplogroups (within all haplogroups) are thought to cause (more or less mild) diseases and hence be deleterious. Now, the vanishing of such "unfit" clades may take centuries or even many millennia, specially if the handicap they cause is not so important or manifests only at advanced age.

But there is indeed no strong evidence on the larger, long-lived haplogroups, at most weak statistical correlations that may in fact indicate something else, something related to the overall genetic ancestry and not particular lineages as such.

...

Ponto said:

The H mtDNA supposedly found in some skeletal remains of around 28 ky of age is dubious. In fact I consider it to be the result of either fraudulent reporting, not uncommon in science, or contamination by any number of personnel who have handled the remains since being uncovered in some Italian cave.

Have you read the paper? It clearly states that the sequence does not match any of the few people that was in contact with the remains and also that the test was reproduced independently with the same results. It's as scientific as it can be.

You are just in denial mode.

Now, you could argue reasonably that the CRS haplotype does not mean H but U or something else that does not exist anymore. But you do have other Paleolithic samples (Portugal, Morocco) in which mtDNA H is clearly dominant, so it is virtually a fact that mtDNA H has been in SW Europe and NW Africa since at least 12,000 years ago and that it was already dominant back then. The evidence also supports some U in Epipaleolithic Portugal, including specifically U5, and some JT(xJ,T) in Morocco (this rare lineage is only found modernly in North Africa and Italy).

argiedude said...

They didn't find 2 samples of "rare haplogroups", but 2 samples out of 10 of haplogroup I, which occurs at a frequency of less than 2% in the modern population. The difference is significant at the 0.05 level.

If you take 10 random samples from a population that has 5 rare haplogroups, each with a frequency of 2%, there's a 1 in 8 chance one of the five rare haplogroups will appear twice. You can replicate this in excel, using the RAND() function.

Dienekes, how could you not have been aware of the ancient mtdna studies mentioned by Maju? Isn't ancient dna one of your specialties? You should've immediately recalled these previous results when this new study came out, and then tried to interpret what it all means as a whole.

PS: Thanks very much, Maju.

Dienekes said...

Dienekes, how could you not have been aware of the ancient mtdna studies mentioned by Maju?

Maju's "ancient mtDNA studies" on Europe boil down to a questionable 2005 Iberian conference presentation that saw no followup, to his own unfounded inference that a contested 28,000-yo Cro-Magnon belonged to haplogroup H, to the irrelevant North African data, and to his false statement that "In fact, AFAIK, it also pops up in the samples of the North just that at lower frequencies than today." as no pre-farming samples from the North have ever been found to belong to haplogroup H.

The evidence for southern Europe is just not there, and while haplogroup H may have been present in it before the advent of farming, it either (i) came to it late, after the last glaciation, or (ii) it was relatively scarce, otherwise its complete absence from northern European Paleolithic inhabitants becomes difficult to explain.

The Other Jim said...

Maju said...

AFAIK, many modernly extant sub-haplogroups (within all haplogroups) are thought to cause (more or less mild) diseases and hence be deleterious.

You are correct that that opinion is in the literature. Again, I feel it lacks any substantial evidence to support it, and their are equally critical studies of mtDNA and complex disease phenotypes (diabetes, etc). The jury should still out on that hypothesis as well.

In non-human species, such as insects (larger Ne) fixation of mildly deleterious mtDNA mutations is rampant (pubmed search Rand D[author] as a starting point). Human Ne is thought to be smaller, thereby increasing the probability of fixation of mildly deleterious alleles. Even if there is some "Russell's teapot"-level fitness difference, the power of drift has been stacked against purifying selection for these alleles.

Maju said...

Maju's "ancient mtDNA studies" on Europe boil down to a questionable 2005 Iberian conference presentation that saw no followup, to his own unfounded inference that a contested 28,000-yo Cro-Magnon belonged to haplogroup H, to the irrelevant North African data, and to his false statement that "In fact, AFAIK, it also pops up in the samples of the North just that at lower frequencies than today." as no pre-farming samples from the North have ever been found to belong to haplogroup H.

No falsehood at all!!! From Wolfang Haak et al, 2005:

From a total of 57 LBK/AVK individuals analyzed, 24 individuals (42%) revealed reproducibly successful amplifications of all four primer pairs from at least two independent extractions usually sampled from different parts of the skeleton. Eighteen of the sequences belonged to typical western Eurasian mtDNA branches; there were seven H or V sequences, five T sequences, four K sequences, one J sequence, and one U3 sequence (table S1).

This is the study that found the 25% N1a (6/24 individuals). Assuming that all HV found is H (what is just logical), the so much hyped N1a is still less than H and only slightly more than T or K. In fact we should say that Neolithic Central Europeans (LBK from what is now Germany, notably the Upper Elbe) appear to have been roughly 1/4 H, 1/4 N1a, 1/4 T and 1/4 K in their mtDNA (with a less important presence of J).

Plus the study of Chandler, Sykes and Zilhao is quite relevant, you like it or not. And the Moroccan mtDNA is not irrelevant either, specially as we know now that North African H is with all likelihood derived from Iberian one and also because it estabilishes a clear case (another clear case) of continuity between the UP and now.

With all that, suspecting that the Paglicci CRS sequence is in fact H is only logical, though in that case we can't be certain enough (too small sample, etc.)

I'd even add that the Haak study can be somewhat criticised because most of the sample comes from a very small region of East Germany and in fact most of the N1a is found precisely there (in the Rhin area it was much less than 25% and in the Elbe was more). So while it does appear that LBK people had more N1a than would be normal today, the apportion for the whole culture may well be much lower and this phenomenon be a at least partly a local founder in East Germany - an ill-populated region before Indoeuropean times.

The evidence for southern Europe is just not there, and while haplogroup H may have been present in it before the advent of farming, it either (i) came to it late, after the last glaciation, or (ii) it was relatively scarce, otherwise its complete absence from northern European Paleolithic inhabitants becomes difficult to explain.

Well, there is no such "complete absence" in fact, the southern evidence is there for those who choose not to ignore it and I know of no studies on Northern European (or more properly Central European, as the North was a mere ice desert) Paleolithic aDNA anyhow.

Dienekes said...

No falsehood at all!!! From Wolfang Haak et al, 2005:


These are Neolithic, not Paleolithic Europeans, and as the more recent study makes clear, they were not indigenous to the region.

Whether or not they came from the Balkans or outside Europe remains to be seen. Ancient DNA studies may one day determine this.

As I have argued above, the latter scenario is more likely, as a long stay in the Balkans during Paleolithic times is scarcely compatible with zero spillage of mtDNA H into northern Europe. After all, the Balkans was one of the areas from which Europe was repopulated after the last glaciation, and if H was already established as a major lineage there, already, the absence of H in the north is hard to explain.

Indeed, an origin of the LBK mtDNA gene pool from West Asia appears to be the position of Haak as well, although ancient DNA studies are needed to confirm it:

"Based on extended
dataset (44 individual mtDNAs from different
sites of early LBK culture) it was found that this
population is genetically similar to present day
populations of Northern Mesopotamia, southern
Caucasus and eastern Anatolia."


as we know now that North African H is with all likelihood derived from Iberian one

I guess you think that the H in Central European farmers who moved into a territory inhabited by foragers completely lacking in H came from Portugal...

and I know of no studies on Northern European (or more properly Central European, as the North was a mere ice desert) Paleolithic aDNA anyhow.

Right, on the one hand you "know of no studies" of the North, but at the same time "it also pops up in the samples of the North".

Maju said...

Admittedly I did not pay too much attention to Bramanti's paper because it is behind a paywall. Only now I have looked at the supplementary material, which seems to include all the relevant info.

Those samples are Subneolithic (i.e. contemporary with the Neolithic ones) and certainly do suggest a sharp dividing mtDNA line between Northern mainland European last hunter-gatherers and contemporary Neolithic tribes of the same area (only Danubian Neolithic - there's no comparison with Dniepr-Don Neolithic even if many of the Subneolithic samples come from Eastern Europe).

You don't find that extreme difference in Portugal or Morocco, ratifying the archaeologically-based idea that Cardium Pottery expansion did not imply in most cases population replacement, while Danubian Neolithic did.

It's noticeable that Bramanti's Neolithic data include much less N1a (3/25 = 12%) than Haak's, what is consistent with what I already suspected of accidental oversampling of this lineage in Haak's paper.

You ask (somewhat sarcastically, I feel) if I believe that the mtDNA H, apparently carried by Danubians into mainland Northern Europe has a Portuguese origin. Just in case there's any doubt, my answer is not - though if Megalithism at a later stage had any demic impact, Portugal should be indeed a referential possible origin.

But it is perfectly possible that the mtDNA of Danubians could have been picked in other less northernly areas, like the Balcans (Balcan Neolithic) or the middle Danubian Basin (original formation of LBK as quite distinct of Balcan Neolithic in an area once inhabited by post-Magdalenian hunter-gatherers).

As you say well, further aDNA studies are needed to determine the facts with more accuracy. I'd be particularly interested in Central European (post-)Magdalenians and Eastern European Gravettians and Dniepr-Don remains, as well as in Balcanic and early Danubian Neolithic ones (and let's not forget Denmark). But a fact remains: that in the Western Mediterranean area H is older, probably much older, than Neolithic (and it's found alongside with U, btw).

A reasonable possibility, with the current data, is that this expansion of H into the North was in fact a South to North demic migration within Europe, at least to some extent. That is what archaeology tells us on Balcano-Danubian Neolithic (but does not apply to the other Neolithics). Archaeology does suggest a West Asian (Anatolian specifically) link for the origins of Balcanic Neolithic but this is not as clear cut as the migration northwards in the Balcans and later in Central Europe, which show high levels of cultural homogeneity at the beginning (excepting somewhat the Balcanic/Danubian cultural divide).

For me, from a purely archaeological viewpoint, the genetic pool of Danubians should have been original mostly from Hungary/Austria/Moravia, with whatever apportion coming from further south (the Balcans) and maybe ultimately from Anatolia. This, I insist, does not apply for the Mediterranean Neolithic nor Dniepr-Don. Nor should apply to the Atlantic areas either (Denmark, Britain, Western France, ancient Basque/Aquitanian Country), which were never directly affected by the Balcano-Danubian wave.

eurologist said...

For me, from a purely archaeological viewpoint, the genetic pool of Danubians should have been original mostly from Hungary/Austria/Moravia, with whatever apportion coming from further south (the Balcans) and maybe ultimately from Anatolia.

This is basically what I tried to argue in the other recent thread.

Important cultural differences to the north where in existence already (just) before the advent of agriculture, and the 500 years or so it took this region to develop its own agricultural model and culture, quite distinct from the Balkans, may have skewed their MtDNA pool even further.

Given the fact that several major lineages occur with substantial percentages, and others in small but non-negligible rates, a lot more ancient MtDNA needs to be evaluated before jumping to any conclusions outside of the in this case relatively clear picture archaeology paints.

From a European viewpoint, if most of the origin is middle Danubian, it is local - at least local central European, and not "foreign." That the Northern tribes may have largely lost out in the eventual gene pool is not a surprise, given the very low population density of hunter gatherers the forest could support.

Dienekes said...

You don't find that extreme difference in Portugal or Morocco

According to your own source there is no continuity between the Epipaleolithic and Neolithic in Portugal. Also, the authors' rejection of an extra-European origin for the Portuguese Neolithic is based only on the absence of mtDNA haplogroup J, which is very weak.

But it is perfectly possible that the mtDNA of Danubians could have been picked in other less northernly areas, like the Balcans (Balcan Neolithic) or the middle Danubian Basin

Yes, but as I have pointed out, a European origin of the mtDNA of the Danubians would require that people who lived in Europe during Paleolithic times would be partitioned into two groups (U-dominated north and LBK-like south) with an Fst between them of more than 0.15. It is very hard to imagine why such a sharp divide would persist over a long time in Europe. It's much simple to think that the LBK-like elements arrived in Europe in the Neolithic or just before it. If they had been present before the last glaciation, then they would have participated in the recolonization of northern Europe, and the sharp divide would not be evident at all.

but this is not as clear cut as the migration northwards in the Balcans and later in Central Europe

Craniometrically it is, as the recent study by Pinhasi et al has shown. The LBK shows clear affinities with the Balkans and Anatolia (and Near East) and not with the Mesolithic Europeans.

The Balkan Neolithic shows clear affinities with that of West Asia both archaeologically and anthropologically. If the Central European Neolithic was not derived from SEE immigrants, then this would require something extraordinary: a genetic separation between CE farmers from both CE hunter-gatherers and SEE farmers: LBK would be a true genetic isolate right in the middle of Europe.

Maju said...

According to your own source there is no continuity between the Epipaleolithic and Neolithic in Portugal. Also, the authors' rejection of an extra-European origin for the Portuguese Neolithic is based only on the absence of mtDNA haplogroup J, which is very weak.

Do you only read the conclusions or forge your own opinions on the data? I try to do the second.

The presence of Zilhao among the triad of authors, who is opposed to Paleolithic continuity specifically in Portugal on his own reasons, pre-defined the conclusions, but the data shows clearly a great deal of continuity:

1. H is clearly dominant before and after Neolithic (though it may increase its presence somewhat in the later period)

2. U, including about a half of it being U5 (the other U* could well be U6) keeps the apportions more or less between the two samples.

3. There is no J, T, K, I, W or X in either sample.

So the main differences are some 18% "exotic" L(xR,I,X,W) among Epipaleolithic and some increase of H, accompanied by a small amount of novel V. The differences are small and could even be attributed to mere randomness between any two samples, though they may also be interpreted as sign of some limited demic change too.

The basic structure is mostly H with some U5 and some U(xU5,K) and that persists, not only between the two samples but also when compared with modern nearby populations.

Maju said...

Yes, but as I have pointed out, a European origin of the mtDNA of the Danubians would require that people who lived in Europe during Paleolithic times would be partitioned into two groups (U-dominated north and LBK-like south) with an Fst between them of more than 0.15. It is very hard to imagine why such a sharp divide would persist over a long time in Europe.

I rather agree in this with you: the difference is quite shocking and rises many questions. But the case is that you do have that unexpected difference between Portuguese and Mainland-North European Epipaleolithic peoples, so the issue remains open.

What we have is that:

1. H existed and was dominant among Epipaleolithic Iberians but not among those of the northern rim of mainland Europe, where U5 and U4 were dominant almost to exclusion of anything else, it seems.

2. U5 (and other U) was present among both populations so we can safely agree that U or at least some subclades of it is of Paleolithic origin in Europe.


3. Danubian Neolithics had an array of lineages of which:
3.1. H would look of southern European origin attending to its presence in Portugal and other aDNA samples (Italy, Morocco).
3.2. K could be West Asian or, as Gioello suggests in the other thread, European (related U8b has been found in Jordan and Italy, while more distantly related U8a is rare but Western European and almost exclusively Basque). Whatever the case, it has a subalpine main distribution, which does suggest it might have spread in association with Danubian Neolithic and associated processes.
3.3. T2 (specifically T2e but this subclade only appears as "proposed" in PhyloTree and is only defined by HVS mutations, so not sure how strictly it should be considered). T2 has an specifically European distribution that has been suggested to be Neolithic and derived from West Asian T since long - though someone could well argue for it to be old in Europe as well.
3.4. N1a. Well, weird lineage but surely of West Asian origin.
3.5. Some J, also typically considered Neolithic of West Asian origin, almost without doubt.

So IMO the Danubian mtDNA could well have two constitutive components almost 50/50: H and K could be ancestrally European and N1a, T2 and J could be West Asian. This is arguable of course but it's my tentative opinion anyhow. The relative aboundance of H and K among Danubians could well be a matter of founder effect and the absence of them further North (among Epipaleolithic Northerners) of founder effects/drift as well.

It seems to suggest that U was for whatever reasons more important in the North and that Neolithic (not in all Europe but certainly in that area of Central Europe) would have carried southern lineages northwards, both European and West Asian. It would imply a "mediterraneization" of Northern Europe.

But the real issue is maybe what happened to those lineages that appear to have been frequent in Danubian Neolithic but are not more (N1a specially but also T2 and even K up to a point). Does it mean that another population with more H in its genetic pool partly replaced these agriculturalists? Or is what we see a matter of mere drift? Or selection if you wish but I see the evidence in favor of that as very weak.

A real possibility is that Danubians were not only culturally and politically "sandwiched" between Western Megalithic and Eastern Idoeuropean cultures but that such cultures also partly replaced their populations. But so far we can only speculate: we'd need much more aDNA to be reasonably sure. And we'd certainly need some aDNA from Neolithic Eastern Europe, where the main post-Danubian wave originated.

Maju said...

It's much simple to think that the LBK-like elements arrived in Europe in the Neolithic or just before it. If they had been present before the last glaciation, then they would have participated in the recolonization of northern Europe, and the sharp divide would not be evident at all.

Or maybe there was no such radical recolonization of Northern Europe (it's clear now that some pockets of populations must have survived the LGM, even maybe influencing the origin of Magdalenian, which is nothing but evolved Aurignacian) or maybe this recolonization was a process with various centers and various founder effects too. Hard to judge, really.

At least Mellars proposes that Aurignacian (that for him includes Bachokirian as precursor and proto-Aurignacian as southern vanguard) was two waves, one by the south (Northern Italy) and another by the north (Central Europe). I can't really judge but if this would be true it could explain some things.

argiedude said...

Maju, I don't understand why you think the Paglicci sample could be H. The study says:

+7,025 AluI, 00073A, 11719G, and 12308A. Therefore, this sequence belongs to either haplogroups HV or pre-HV

H is defined by -7,025 AluI. HV is defined by +7,025 AluI.

I don't understand the study's conclusion about the other sample:

Paglicci-12 shows the motifs 00073G, 10873C, 10238T, and AACC between nucleotide positions 10397 and 10400, which allows the classification of this sequence into the macrohaplogroup N , containing haplogroups W, X, I, N1a, N1b, N1c, and N*

10873 defines macrogroup N, but 10238 restricts it specifically to just N1, which includes N1a, N1b, N1c, and I, but not W or X. Maybe the tree chart I'm looking at is incorrect, or maybe there's been a rearrangement of the tree since this study was published?

They then say the single mutation 16223 in this sample suggests a classification in haplogroup N*. But in Richards (2000), which listed thousands of haplotypes, it seems clear that 16223 is very common in all the haplogroups listed: W, X, I, N1a, N1b, N1c, and N*. So 16223 doesn't support N* specifically, it just reenforces the conclusion from the coding region that the sample belongs to the group {W, X, I, N1a, N1b, N1c, N*}. Also, the only samples in Richards' that had exclusively the mutation 16223 were an I and several L3*/N*/W, but no N* specifically, so I don't see how they reached the conclusion it was specifically N*. [16223 is also very common in A, B, C, D, L1, L2, L3, M, and extremely rare in all other haplogroups]

Dienekes, I read the Iberian study/presentation about Portuguese ancient mtdna and I don't see why it's controversial. Can you explain that? You only said that you hadn't heard anything of it since, which I agree is a little odd, but then again, I've been waiting for the publication of a huge 2006 study that tested almost 1000 y-dna and mtdna samples from a crucial country, Mali, and it has never resurfaced. I think some, but not all, of the y-dna samples only, have been put up in smgf, but no publication.

If this Paglicci sample is effectively N1 (it could well be I), then this would take the count of N1 samples in ancient dna studies up to 5, very remarkable. Here's the list, correct me if I'm wrong:

N1 from Paglicci cave.
N1a from Haak study of Neolithic Europe.
N1a from 1000 year old Hungary site.
N1a from 2500 year old Scythian skeleton in Altai region of Russia.
N1 from 4500 year old Eulau site in Germany.

Maju said...

Are we talking of the same study, Argiedude? I searched for the several keywords of the paragraphs you quote in Caramelli 2008 and could not find any one of them.

They insist that the sequence is a CRS haplotype and, while I feel a little unsure looking at the detail of the mutations, it seems that the only consistenly replicated difference to the CRS HVS is a C>T mutation at locus 16294 (table S1). Unsure of what that may mean.

CRS is H2a2 but the same HVS motif is often found in H1 and may also exist in other H and U, for what I've read.

Maju said...

Ops, it's 16274 and only appears in about 50% of the clones (27/56). This is normal and may be due to DNA damage or whatever.

argiedude said...

Maju, here's where I read about the Paglicci study:

http://www.pnas.org/content/100/11/6593.full.pdf

...........................

Another question. Did the ancient mtdna studies of Tuscans and Taforalt perform 2 separate extractions of dna from different sections of each skeleton and tested each independently?

Dienekes said...

Or maybe there was no such radical recolonization of Northern Europe

Distinct gene pools are maintained by either geographical or cultural barriers, or by recent immigrants that have not had enough time to be assimilated by/assimilate the locals.

What possible barrier could have been in Europe for tens of thousands of years, so that an LBK-like gene pool would dominate the south and a U-dominated gene pool would dominate the north? You are talking about maintaining Eurasian-level genetic contrasts across a thousand kilometers.

Maju said...

Argiedude: that's another study that was famous in its day because it demonstrated that Neanderthal mtDNA was very distant from both modern and Paleolithic H. sapiens one. I had even forgotten which the sequences were in that one but for what I read now they were also extremely close to modern European types with one being also CRS and the other a veriation of it with a single nucleotide mutation.

However after thinking for a moment on all that, you must be right that Paglicci 12 was some sort of N1 and that Paglicci 25 was surely HV and not yet H. It is possible that this also applies to Paglicci 23, that is the one I was talking about - can't say.

Maju said...

Dienekes: beats me a bit, admittedly. But the case is that we do have now epi-Magdalenian in one place with 60% H and epi-Magdalenian in another with 100% U(xK).

Why? Maybe because the genesis of both Magdalenians was somewhat different in ways we can't percieve clearly only on cultural elements. Sometimes different ethnicities share most of their material culture, for instance in the early Middle Ages it's virtually impossible discern Basque from Frankish or Visigothic: they all used the same kind of gear and stuff.

I know it's a different age but the case is that many questions stand on what do cultures (or maybe just technologies?) mean in the Upper Paleolithic. Some differences do exist anyhow since the beginning: Central European Magdalenians did not paint any cave... though arguably this is because there are not many caves over there. On the other hand Adriatic Epigravettians did.

I'm trying to figure out but really don't have any good answer beyond that, sincerely.

Maju said...

Argiedude: just noticed that, according to PhyloTree, mtDNA H is not defined by any mutation at locus 7025 as you said but at 7028 and 2706. In all the R0 (pre-HV) tree page I could not find any 7025 locus defining mutation. In fact I could not find it in all the human mtDNA tree.

So I'm guessing it's either a typo or it could still be H.

Dienekes said...

Dienekes: beats me a bit, admittedly. But the case is that we do have now epi-Magdalenian in one place with 60% H and epi-Magdalenian in another with 100% U(xK).

First, I don't exactly take a paper in Science and an Iberian conference presentation with few details and no apparent followup to carry equivalent weight as evidence.

Second, assuming the Portuguese results are real ancient DNA, they are still from a time frame that postdates the Neolithic colonization of Europe. While it is possible that non-U mtDNA arrived in Europe shortly before rather than during the Neolithic colonization (as I have stated above), its long-term presence in Europe is problematic for the reasons I enumerated.

Maju said...

Second, assuming the Portuguese results are real ancient DNA ...

I'm only going to consider this part. The other you can discuss with Chandler, Sykes and Zilhao.

... they are still from a time frame that postdates the Neolithic colonization of Europe.

For sure not to that part of Europe nor to anywhere outside some Balcanic areas.

Actually you could well say the same about the Epipaleolithic samples of Bramanti, which are of a very late date, contemporary of regional Neolithic and might (hypothetically) be different than earlier local UP as well.

Furthermore, you still have the Taforalt samples (also high in H, which should be of Iberian origin for what we know) of a much earlier date (12,000 BP) and then the two CRS sequences of Paglicci. So it's kind of beating a dead horse, I feel.

While it is possible that non-U mtDNA arrived in Europe shortly before rather than during the Neolithic colonization (as I have stated above), its long-term presence in Europe is problematic for the reasons I enumerated.

I have ennumerated much stronger reasons, I believe.

Maju said...

Plus, if you look at Chandler's paper just for reference (fig. 4), you realize that modernly all shown European populations have that pattern of dominant H and relevant but secondary U5 (and often other U) that we find in both Epipaleolithic and Neolithic Portugal. Instead West Asia ("Middle East") is significatively weaker in H and stronger in other clades like J or the "other" group (including N1), but not in some of the other Danubian lineages: T or K.

So it does look like a good deal of the Danubian gene pool could be of European origins, even if distinct from the gene pool of the northern continental rim back then. It's not really that impossible that Ahrensburgian could imply some sort of founder effect affecting only to that area, I figure.

For me anyhow it is much more intriguing the matter of why the Danubian mtDNA pool got so much diluted upon Chalcolithic or whenever that happened. And why it was diluted in the direction it was: towards a much higher apportion of H in particular. This is a very pressing mystery, IMO.

argiedude said...

Argiedude: just noticed that, according to PhyloTree, mtDNA H is not defined by any mutation at locus 7025 as you said but at 7028 and 2706. In all the R0 (pre-HV) tree page I could not find any 7025 locus defining mutation. In fact I could not find it in all the human mtDNA tree.

Perhaps the problem is because 7025 isn't an SNP but rather an insertion, and maybe these charts don't list insertions/deletions. Thanks for that chart you linked, it's very detailed, up to date and easy to read.

However after thinking for a moment on all that, you must be right that Paglicci 12 was some sort of N1 and that Paglicci 25 was surely HV and not yet H. It is possible that this also applies to Paglicci 23, that is the one I was talking about - can't say.

I thought there were just 2 fossils tested from the Paglicci cave?

Maju said...

I thought there were just 2 fossils tested from the Paglicci cave? -

No. Caramelli 2005 (link above) analyzed another one: Paglicci 23 and proclaimed that it was a clearly uncontaminated CRS haplotype. They did not analyze the coding region though.

Unlike you I was only really aware about Paglicci 23 and had totally forgotten about the other two (Caramelli 2003), so thanks for refreshing my memory on this fact.

argiedude said...

Maju, thank you very, very much for all this info and analysis. :)

argiedude said...

Paglicci 25, the HV fossil, has 4 mutations: 0073A, 11719G, 12308A, +7025 AluI.

0073A and 11719G define haplogroup R0:

www.phylotree.org


12308A is present in the CRS sequence, so I presume it's common to all of R0. So far, we can't place Paglicci 25 specifically into HV, only into R0.

But the last mutation, +7025 AluI, excludes H. But there seems to be a crucial point here. +7025 AluI wasn't gained by HV, it was lost by H. In other words, +7025 is the default state in the mtdna tree, meaning the sample is technically R0(xH), not HV. So while the sample could still very well be HV, it could also just as well be pre-HV/H, and this would actually bolster the argument in favor of the ancient presence and development of H inside Europe.

Here's a link to a rootsweb post showing a Norwegian K with +7025AluI, as an example:

mtdna K +7025AluI

Maju said...

Don't know what to do with that 7025 Alu insertion that is "default" for all the mtDNA tree. If it's truly that default, then it should not be considered an insertion but a deletion at the root of H and nothing else.

So far, apart of the claim by Caramelli in 2003, I have not seen any other evidence that this insertion/deletion is really informative. Sometimes a mutation that once was believed to be informative is found to behave chaotically and be irrelevant for lineage mapping. I wonder if it is the case with this one.

argiedude said...

Don't know what to do with that 7025 Alu insertion that is "default" for all the mtDNA tree. If it's truly that default, then it should not be considered an insertion but a deletion at the root of H and nothing else.

Wouldn't it's status be defined relative to the CRS sample, as per the usual method? I'd personally ditch the current CRS tree and rebuild everything from a root at L0.

Here's a study that tested extensively for 7025AluI:

mtDNA polymorphisms in five French groups (Dubut, 2003)

The supplementary DOC file has lots of 7025AluI data.

Maju said...

Wouldn't it's status be defined relative to the CRS sample, as per the usual method? I'd personally ditch the current CRS tree and rebuild everything from a root at L0.

Well, guess it can be said that way. But strictly speaking it'd be a deletion anyhow.

The supplementary DOC file has lots of 7025AluI data.

Well, not really: it has a lot of -7025AluI mentioned apparently as marker to determine H but is of about the same date, so guess it was usual to do that way back then. I'd like to know why is it not used anymore.

What is clear in this paper too (and I see that repeated in all studies) is that the CRS haplotype in fact means H (and mostly H1 in fact). I have not yet seen a single CRS, being U, HV or R0, even if some authors seem to think it is a possibility and don't like to take risks.

Maju said...

Addendum: While trying to find more on this indel marker, I did find one example of CRS U* in Asturias (Maca-Mayer 2003) but has a mutation at the site 12308 that defines this haplogroup, while Paglicci 25 has this site defined as in CRS. So guess it can't be U* either for this individual.

Maju said...

Ok, I'm through looking for 7025 AluI references: it does seem a lot of people has been using it, instead of the "official" markers for determining haplogroup H, including as recently as Roostalu 2007 (hg H in Caucasus and Near East - btw, worth checking the age estimates in that paper, which are generally quite old: H11 for instance could be from 48,000 BP).

However it may be worth noticing that Pereira dos Santos 2005 (Cape Verdean mtDNA) mentions (p. 428) that some control measures are needed to make sure that there are no errors in testing for such kind of unique markers, which could lead to a high number of false results (can't copy the text, you'll have to browse).