search this blog

Loading...

Friday, April 25, 2014

Low genomic diversity among ancient Swedish foragers (+ no East Asian admix for La Brana1 and MA1)


Here's an Admixture graph from a paper published in Science today, which focuses on ancient genomes from Mesolithic and Neolithic Sweden.


I think it sums up very succinctly some of stuff we've discussed at length on these blogs in recent months. Note that it's the Neolithic and post-Neolithic European farmers, Gökhem2 and Oetzi the Iceman, respectively, who appear to be mixtures of two very distinct Eurasian clades; one shifted towards Sub-Saharan Africa, and the other basically identical to that of European hunter-gatherers.

On the other hand, hunter-gatherers La Brana1 and MA1, from Mesolithic Iberia and Upper Paleolithic Siberia, respectively, are each derived entirely from two closely related North Eurasian hunter-gatherer clades, while Ajvide58, a hunter-gather from early Neolithic Gotland, appears to be a mixture between these clades. Note also that Amerindians Anzick-1 and Saqqaq are significantly East Eurasian (pink clade), but none of the ancient North Eurasians show this influence.

Moreover, one of the main findings of this study is that the Swedish hunter-gatherers had unusually low conditional nucleotide, or genomic, diversity (0.18122). Much lower than the ancient Swedish farmers (0.20134), modern Europeans (from 0.20754 among the British to 0.21269 among Spaniards) and even modern Han Chinese (0.19442), who are known to be amongst the least genetically diverse human groups. Refer to Table S16. in the freely available supplementary material PDF for more details (see here). This quote explaining the results is from the main article:

The distinct features of the two Neolithic Scandinavian groups; non-symmetric gene-flow into farmers, low level of diversity among hunter-gatherers and strong differentiation between groups have important implications for our understanding of the demographic histories of these groups. The greater diversity in the farmer population may have been influenced by gene flow from hunter-gatherers. However, the low level of genetic diversity in Neolithic hunter-gatherers likely has a demographic explanation, similar to the Iberian Mesolithic individual (8). Although we cannot exclude that this low diversity is a feature restricted to the Gotland island hunter-gatherer population, we note that this may be due to the fact that their ancestors resided in ice-free refugia in Europe during the Last Glacial Maximum (LGM), potentially causing population bottlenecks. Climatic changes and occasional population crashes, likely affected the population sizes of hunter-gatherers (24, 25).

The PCA from the paper underlines how extremely closely related La Brana1 was to the hunter-gatherers from up north. That's despite the fact that he was from Iberia and his Y-chromosome belonged to haplogroup C6, while all of the successfully tested Y-chromosomes from among the Northern European hunter-gatherers to haplogroup I.


However, just like Motala3, Motala12 and Loschbour from the Lazaridis et al. preprint, Ajvide58 belonged to sub-haplogroup I2a1, which is presently much less common across Northwestern Europe, especially in Sweden, than I1. It actually peaks today in the Balkans and Eastern Europe, and in fact the Loschbour I2a1b sequence is most closely related to that of a Russian from the HGDP dataset, sampled in the Kargopol district in the northwest of the country. It's still a mystery how I1, and also R1b, became so prevalent in Northwestern Europe after the Mesolithic.

By the way, it warms my heart to see the following two figures in the supplementary material, in which the authors note that the sharp genetic differentiation between Ajvide58 and Gökhem2 is reflected in their varying affinities to modern northwestern and southwestern Eurasians. That's because two years ago I wrote a blog entry based on an ADMIXTURE run arguing that modern West Eurasians descend from two main ancient stocks: the Northwest Eurasians and Southwest Eurasians (see here). Looking back at that effort, I certainly rambled on and didn't quite hit the nail on the head, but the general idea wasn't too far off from what we're now finding out about our deep ancestry thanks to ancient DNA and more sophisticated analysis methods.


Interestingly, the Northern European sample with the lowest affinity to Ajvide58 are the Saami from Norrbotten, Sweden. This is most likely due to relatively recent Siberian ancestry among the Saami, which is essentially East Asian admixture and, as per above, not found among the Neolithic and pre-Neolithic North Eurasian hunter-gatherers or European farmers.

I haven't had a chance yet to look in detail at the inferred pigmentation traits of the ancient Swedes. But Gökhem2 carried the derived alleles for SLC24A5 and SLC45A2, which means she probably had fair skin. On the other hand, Ajvide58 was ancestral at these loci, meaning he was probably dark skinned, much like La Brana1. However, as far as I can make out right now, tables S9 and S10 in the supplementary material suggest that both Gökhem2 and Ajvide58 might have been blue eyed and fair haired.

Citation...

Skoglund, Malmstrom et al., Genomic Diversity and Admixture Differs for Stone-Age Scandinavian Foragers and Farmers, Published Online April 24 2014, Science DOI: 10.1126/science.1253448.

See also...

Ancient human genomes suggest (more than) three ancestral populations for present-day Europeans

Another look at the Lazaridis et al. ancient genomes preprint

Mesolithic genome from Spain reveals markers for blue eyes, dark skin and Y-haplogroup C6

125 comments:

Shaikorth said...

Once the AJV58 sample comes available, you should definitely check its IBS against Han Chinese and compare it to that of Mordovians or Kargopol Russians to see whether that admixture tree fully shows East Eurasian affinities. The distances between "admixture components" indicate there is further investigation to be done.

Also, AJV58 shares as much drift with Saami, Mari and Chuvash as with Spanish, Bulgarians and Tuscans. No modern North European does, this in itself is actually indicative of easterness.

Davidski said...

Well, AJV58 lacks the Basal Eurasian admixture that modern North Europeans have, that's why he appears more eastern and shows lower affinity to Southern Europeans.

Seinundzeit said...

Shaikorth,

For what it's worth, ancient European hunter gatherers actually belong to the same "big" clade as East Eurasians. You can see that in the tree graph, I think MA1 and the European hunter gatherers are technically closer to the indigenous Australian genome than they are to the early European farmers (in terms of lineage).

From the supplements:
"In other words, early European hunter-gatherers are consistent with forming a clade with East Asian groups and the Australian groups to the exclusion
of a population lineage that contributed genes to the early farmers." (p. 29)

That probably explains why it's so difficult to separate ANE, WHG/SHG, and ENA. I think all three are just too close to each other, on a broad scale, especialy when one compares them to Basal Eurasians.

Shaikorth said...

Yeah that explains a lot. AJV58 shares as much drift with Nganassans who are basically the most extreme "modern" Siberians as with Saudis. IBS comparisons might tell us more.

I still don't understand why they projected the ancient samples onto their PCA again. Sure, some are low coverage but AJV58, GOK, La Braña and Iceman surely have enough SNP's shared to make a proper PCA with them?

Shaikorth said...

Looking at table S13, AJV58 looks to have have greater affinity to Anzick-1 (ENA/ANE mix like Karitiana) than to MA-1 (pure ANE). The history of movements through Siberia between the Upper Paleolithic and Neolithic doesn't seem settled yet. We really need that Mesolithic East European DNA.

Davidski said...

I'm starting to think more and more that ANE was ancestral to WHG. In other words, WHG was just a drifted, more recent version of ANE. In fact, maybe ANE was already extinct during the Mesolithic?

The later ANE-like signal that sweeped into much of Europe from the east during the Copper Age (?) might have been from a sister clade to WHG, which is now at much higher levels across Europe, especially in the east, than what the genome of MA1 lets us estimate.

barakobama said...

I have started a thread, but still need editing.

http://www.eupedia.com/forum/editpost.php?do=updatepost&p=430313

Surprisingly Sf11 a 7,500-7,250BP Mesolithic hunter gatherer from the Balktic sea had derived alleles in SNP rs16891982.

Colin Welling said...

no East Asian admix for La Brana1 nor MA1

Correction, MA1 is Asian. MA1 lived deep in Asia, and differentiated itself there, regardless of how close he was to other Asians.

WHG differentiated itself in Europe during and after the ice age.

ANE and WHG are somewhat related to each other but less so than EEF is to WHG (as a result of admixture of course).



Colin Welling said...

By the way, it warms my heart to see the following two figures in the supplementary material, in which the authors note that the sharp genetic differentiation between Ajvide58 and Gökhem2 is reflected in their varying affinities to modern Northwestern and Southwestern Eurasians.

haha, good call

Shaikorth said...

WHG/SHG, ANE and ENA are all closely related Eurasian clades. This closeness is seen in AJV58 having more affinity with Anzick-1 (who shows as a mix of "ordinary" East Eurasian and MA-1 in their admixture tree) than to MA-1.

EEF clearly is closer to SSA than these though.

BTW Davidski, there might be another explanation for affinities of the Saamis. They are extremely drifted isolate group which might reduce their shared drift signals and show them as less SHG than they are. This hypothesis is again easily testable in Finestructure or IBS sharing if someone has Saami samples (Pålsen and Verenich at least IIRC), if they show up as distant as Chuvash or Mari with those methods as well, the shared drift obviously got it right.

I'm basing the idea of isolates getting reduced shared drift signals on the Pickrell and Reich paper, where Orcadians (obvious isolate) have weaker Karitiana-Sardinian two-way admixture signal than French despite obviously being more ANE/WHG-like. http://biorxiv.org/content/biorxiv/early/2014/03/21/003517.full.pdf

Grey said...

@Colin Welling

"Correction, MA1 is Asian. MA1 lived deep in Asia."

ANE isn't eastern or western it's northern - stretching across a vast area from Europe in the west to Siberia and beyond in the east.

That's what the genetics say and there's a perfectly plausible explanation for why that might have been: mammoth.

#

@Davidski

"I'm starting to think more and more that ANE was ancestral to WHG. In other words, WHG was just a drifted, more recent version of ANE."

That might fit a continental (ANE) vs maritime (WHG) split.

About Time said...

I'm still eager to know whether Neanderthalized genes were part of the picture for ANE. Cold and big game hunting adaptations, like close coordination with group, ability to endure physical hardship, store fat (intermittent calorie intake), and periods of inactivity (maybe including neurological mode we'd today call "depression" or "meditative state").

In other words, evolutionary changes that made ANE (+ derived WHG?) more like bears or really wolf like pack hunters.

DarthVadent2 said...

If ANE is ancestral to WHG, what ancestral clusters do you think are ancestral to Caucasus, Gedrosia, Mediterranean and Southwest Asian? I think Gedrosia itself is composed 50/50 of ANE and Basal Eurasian, similar to Caucasus, except I think that Caucasus is more strongly defined by Basal Eurasian than ANE.

Also, are Ötzi and the Gök4 farmer more hunter-gatherer admixed than Stuttgart is?

barakobama said...

HOLY SHIT!!!!!!!!!!!!!!!!!!!

Did anyone see how much drift Sami from Norrbotten, Sweden share with hunter gatherer Ajv58, as much as fellow Sardinians do with Gok4. This could partly be because of their east Asian ancestry(No basal Eurasian), but it defintley is also because they are more Mesolithic than anyother modern Europeans.

Seinundzeit said...

DarthVadent2,

This is a very interesting question. Comparing the papers, Stuttgart and Ötzi seem identical, despite using very different methods. Stuttgart is supposed to be around 44% Basal Eurasian, and Ötzi is supposed to be around 44% Basal Eurasian. The methods are different, but I think we are justified in saying that both are fairly similar. Also, I just checked the Lazaridis et al. world-wide PCA plot, Ötzi and Stuttgart are very close. Ötzi is slightly closer to South Asians (and Africans), but the difference is very minor.

If David wants to try anything, Ötzi is just as good as Stuttgart. I think some experiments with MA1, La Brana1, and the Iceman, would be very informative.

When it comes to the Gedrosia/Baloch component, I think it's probably the closest West Eurasian component to ANE. I'm sure it's composition involves the highest amount of ANE ancestry found in any West Eurasian ADMIXTURE component/"ancestral population".

To my mind, the main point seems to be that Basal Eurasian ancestry constitutes the main connective tissue between West Eurasian populations. From Ireland to India, there is Basal Eurasian admixture. And in an even bigger sense, "Basal Eurasian" admixture is what makes West Eurasians really different from ENA. ENA, ANE, and WHG/SHG are all purely "Derived Eurasians", and constitute a large clade versus "Basal Eurasians". ANE and WHG/SHG are the western lineage within the "Derived Eurasian" clade. ENA are the eastern lineage within the "Derived Eurasian" clade. Native Americans are mixtures between "Western Derived Eurasians", and "Eastern Derived Eurasians". But living West Eurasian populations are complex mixtures between "Derived" and "Basal" Eurasians, but predominantly "Derived Eurasian", and mostly "Western Derived Eurasian". This also explains why living Europeans constitute a clade with low-caste or tribal South Indians, to the exclusion of Bedouins/Palestinians/Jordanians. It's because most Europeans are predominantly "Western Derived Eurasian", which is extremely closely related to "Eastern Derived Eurasian" (aka, ENA).

Fanty said...

Question is, if ANE and WHG are something related to "East Asian", why is the main WHG male lineage "I" (today (all "I" clades) mainly in Northern and Eastern Europe), a brotherclade of "J" (today mainly in West Asia).

The Y-DNA suggests a deep anchestry relation of WHG with near eastern populations at least on the male lineage (well with the exception of a single "C".)

Daniel Szelkey said...

Eurasian clades; one shifted towards Sub-Saharan Africa, and the other basically identical to that of European hunter-gatherers.

The first thing I have to ask is where does y-dna haplogroup R fit?

The non-IE peoples of the caucasus such as the Georgians on one side and the Chechans on another side are not closely related to each other (I am not talking about Azeris) .However, they have several things in common. First the Caucasus is probably the least (or one of the least) africanized part of western Eurasia. 1000 Chechnans no E1b1b, no L or A, or B, or any type of E. Second the Caucasus has a large paleolithic European contribution on the materal side, and the analysis's should show paleolithic European blood in the Caucasus.

The last point I would like to make is that Otzi must have had a significant amount of Paleolithic blood. This is because he is related to the Sardinians, who show strong paleolithic signals on both the maternal and paternal side, and appear perhaps to be the remnants of a population that was both paleolithic and neolithic European, while other europeans don't seem at least uniparentally including on the maternal side, closely related to either the Neolithic or Mesolithic Europeans.
Something is wrong with these autosomial analysis's.

Daniel Szelkey said...

If basal Eurasian is associated with the caucasus Basal Eurasian is not african, and does not drift towards africa. Story over.

Shaikorth said...

I and J-dominated populations had probably been long separated in the Mesolithic. This can easily explain the autosomal difference, as it does in the case of populations dominated by haplogroup P's derivatives in different parts of Eurasia.

Fanty said...

"In other words, evolutionary changes that made ANE (+ derived WHG?) more like bears or really wolf like pack hunters."

Well, at least they are anatomically different. Mesolithic European hunter gatherers are (sometimes extremely) taller (up to 30cm) than neolithic farmers (they even had larger brains, but that doesnt neccessary mean more intelligence. After all the largest brains in the modern world have Siberians)and have more robust skeletons.

Some of these things are usualy claimed to be connected to climate.

Its usualy said, so more cold, so more robust the skeletons. So more hot, so more filigrane the skeleton. The brain size seems to increase with cold and to shrink with heat aswell.

Tallness however is different. In the world wide context its claimed that cold= small+robust and hot = tall+filigrane. But Mesolithic Europeans are tall (about as tall as 21th century Europeans (taller than any European between the Mesolithicum and the 21th century, including any Germanic warriors digged up so far) AND robust.

Even in modern Europe "tallness" is against the world wide concept.
Northern Europeans are taller than southern Europeans, wich is kind of odd (remember that its suposed to be cold=small and hot=tall). And it doesnt completely follow WHG anchestry either.

Tallest Europeans are Scandinavians, Dutch and Yugoslawians. Smallest are Italians (Southern Italians smaller than northern Italians), Greek and Spanish.

Ryan said...

"The later ANE-like signal that sweeped into much of Europe from the east during the Copper Age (?) might have been from a sister clade to WHG, which is now at much higher levels across Europe, especially in the east, than what the genome of MA1 lets us estimate."

If you look at the Anzick-1 paper, their K9, K10 and K11 runs all break down ANE in to two distinct groups. One contributed to all Native Americans, where the second (probably later) group only contributed to populations in the northern part of North America. Both groups showed affinity to Northern Europeans too. The Anzick paper didn't look at Ma-1's DNA, but I'm pretty sure if it did it would have found that the first group to be much more closely related to Ma-1 than the second. I bet you dollars to donuts that Ajvide58 has a close affinity to that second group (which I will refer to as Dene from now on for simplicity's sake). There were actually at least two other migration events as identified by Reich's group but for simplicity's sake I'll just focus on this Dene one.

So, here are a few things I think people should consider:

Dene groups show high levels of mDNA C4b. The most common mDNA groups in early Kurgans were C4a'b'c and C4a.

In North America, Y DNA group C-M217 is only found in these Dene and Dene-related groups, while it is also widely distributed in Siberia and Mongolia.

Karitiana did not receive any inputs from this Dene migration. Any analysis that uses Karitiana as a proxy for all Native Americans will completely miss the potentially much more recent links between Dene groups and Eurasia.

It seems like in general, most papers that relate to ANE do so either from a European standpoint or a North American standpoint. Very few do both together. I feel like that's causing a lot of structure within ANE to be missed.

Davidski - I suspect you're right about there being a deep affinity between ANE and WHG, but just as C-V20 in La Brana means some sort of link to East Asia, haplogroup I in another hunter gatherer remains mean there must have been some sort of early input from the Middle East. I'd suggest that WHG was likely a mixed population itself.

I'd love to see a K13 run that excludes Africa and includes the various Native American samples from Reich's database and the various ancient DNA samples that are now available if you have the chance sometime.

Cheers.

Moi said...

Any plans to do a segment matching with these guys and the Eurogenes participants?

Evon

Ryan said...

This figured from Reich et al 2012 gives a clearer picture of what I'm talking about:

http://www.nature.com/nature/journal/v488/n7411/fig_tab/nature11258_F2.html

The group that contributed to both the "green" (Dene) and "red" (Inuit/Aleutian) populations is the one I'm talking about.

If I'm correct, then a part of that "grey" wedge from Ajvide58 should be related to the population that contributed to Dene and Inuit peoples.

If I'm right, that gray wedge should be closest to Khanty and Yukaghir people I think?

I would think the Khanty might be a good proxy for how they may have looked. Fair hair, fair eyes, but more East Asian features.

http://s186.photobucket.com/user/zemelmete/media/hantti/QjpKP3uFwwY_zpsaf3503be.jpg.html

Colin Welling said...

ANE isn't eastern or western it's northern - stretching across a vast area from Europe in the west to Siberia and beyond in the east.

I doubt it's even been in Europe since the ice age. Mesolithic WHG groups are very homogeneous and both brana and loschbour lack ANE completely. I don't see how there would be such a strong element of homogeneity from Spain to Sweden yet ANE was somehow strictly limited to eastern Europe since times before the Mesolithic. Additionally you have EEF types who probably have southeast European Mesolithic heritage but still lacking ANE. That suggests Western and Southern Mesolithic Europe lacked ANE.

If we assume that the ancestors of WHG split from MA/ANE before the ice age then where in Europe would the ANE hide during the ice age?

I suspect ANE traveled into Europe, during the Mesolithic, along the far north (Northwest Russia, Baltic States, and Scandinavia) but the bulk of ANE penetration came from the steppes during the late neolithic.

BTW, does anybody know why EEF is "African shifted". Does that mean EEF has SSA admixture or that Basel Eurasian split from Africans later than the rest of the Eurasians.

Colin Welling said...

I doubt it's even been in Europe since the ice age.

clarification, I mean to say that I dought ANE became established in Europe during the ice age.

About Time said...

@Fanty, body size rules are:

http://en.m.wikipedia.org/wiki/Bergmann%27s_rule

Ceteris paribus, colder=bigger body; warmer=smaller body

http://en.m.wikipedia.org/wiki/Allen%27s_Rule

Cold=shorter limbs/appendages (more efficiently preserves body heat).

Good point that tall/lanky North Sea pops (like Dutch, Swedes, English maybe) and tall Yugoslavs are not really northern adapted type. Eskimos have the expected "true cold adapted" body style.

It's been said many times, but the archetypal North Sea type has maybe more of a Mediterranean pedigree. I keep saying, maybe with TRB/Funnelbeaker roots. In other words, a northern readaptated EEF subtype.

Even today, I'm not sure that blond/blue pigmentation doesn't derive from localized EEF populations that got stuck (separated?) near the North Sea.

barakobama said...

23andme may have just solved the mystery of the occurrence of brown skin in my family: It's descended of Mesolithic Europeans!!!

Most notable my brown skinned uncle does not have the Ala111Thr and Phe374Leu mutations(like Loschbour and La Brana-1) which are fixated in modern Europeans and are suppose to have the biggest lighting effect on modern European skin.

http://www.theapricity.com/forum/showthread.php?123503-23andme-reveals-I-have-relatives-with-Mesolithic-European-decended-dark-skin

Maju said...

"When the Scandinavian hunter-gatherers are combined with La Brana1, the figure goes up to just 0.19744"...

That's more than modern Chinese (your figure). I find quite notable that quasi-Paleolithic Europeans had more genetic diversity than a whole 20% of modern Humankind.

So when you say that "the Swedish hunter-gatherers had extremely low conditional nucleotide, or genomic, diversity (0.18122)", I can't but raise my eyebrows and wonder what else can you expect for a marginal very late subarctic hunter-gatherer population from a small Baltic island?

On another note, I strongly recommend everyone to take a look at fig. S7 and compare it with fig. S6 and fig. 2b (same graph with just color makeup on it). When the authors include two modern European populations and allow for two more admixture axis, the results change very dramatically. This affects two main aspects:

1. Ancient farmers return to the West Eurasian abode but one of them (Ötzi) shows Dinka admixture, what explains the "Basal Eurasian" thing at least approximatively.

2. Modern French do not relate with any of the ancient populations studied and appear basal to them (although with heavy proto-Sardinian admixture). This peculiar result may happen because Lochsbour was not computed here (French appear to be particularly direct descendants from a Lochsbour-like population judging on IBD).

All that emphasizes the many nuances implicit in these approximative tools and the limited ancient samples that serve as proxy for the actual ancient ancestry but must necessarily represent only a small fraction of the actual ancient diversity.

Maju said...

@Fanty:

"Northern Europeans are taller than southern Europeans"...

That's not really precise at all. Brits for example are rather short, especially the women, while Greeks are rather tall, especially the women again. Greek women are clearly much taller than British women on average.

See: http://forwhattheywereweare.blogspot.com/2012/08/the-height-of-europeans-and-myth-of.html

The tendency is more real among men than among women but still the French men are shorter than Spaniards for example. Spaniards are quite average for both genders, while Portuguese are indeed in the short range.

Finns are also rather short (men average, women short).

The tallest of all are the Dutch and the area of greatest height is around Germany, with Slovenes, Lithuanians and Scandinavians being also among the highest for both genders.

Davidski said...

Maju, the French appear basal in that graph because they have significant MA1-related ancestry (aka. ANE admixture). But their WHG ancestry pushes them down the tree, away from MA1, and also the Sardinian admixture edge contains some WHG and WHG-like stuff.

It's just another way of showing the same thing. But the other graph looks more sensible to me, because the modern samples (with their recent drift) are absent and thus don't screw up the results.

And it doesn't matter if the Han Chinese make up 5%, 20% or 99% of the world population. They obviously descend from a handful of people who managed to expand very rapidly not that long ago, which is why they have such unusually low genomic diversity.

Maju said...

It could be, David, but why does no admixture axis evidence such a notable MA1 ancestry, as happens with Ajv58, or why doesn't the French show up as derived from the Braña-Ajv58 clade either? The admixture axes in that graph are as low as c. 10% or less, so the admixture from the represented clades in the French (other than the c. 25% proto-Sardinian one) must be very low as such. Otherwise it would show up.

Anima Mortel said...

Text: "However, as far as I can make out right now, tables S9 and S10 in the supplementary material suggest that both Gökhem2 and Ajvide58 might have been blue eyed and fair haired."

At S9 and S10 Skoglund used an approach akin to that of ref (85) which had an accuracy of 44% for Hair and 36% for eyes. In other words a low chance of being correct about these pigmentations.

Skoglund should have used the correct SNPs which have a far higher accuracy like the ones used by Lazaridis. And Gok2 was also derived for SLC45A2. The hunter gatherers were both ancestral (dark skin) for both SLC45A2 and SLC24A5. Oetzi the other farmer in was also derived for both SLC45A2 and SLC24A5 as Gok2 testetd in this paper and Keller et al. 2012.

It remains a mystery why the hunter gatherers were very dark skinned but the farmers were all light skinned.

Davidski said...

I've added the info about SLC45A2 to the text.

By the way, the dark skin pigmentation of the hunter-gatherers was definitely a surprise, but perhaps not a mystery. European hunter-gatherers probably ate a lot of vitamin D-rich fish, meat and mushrooms, so they didn't need to rely on the sun as much as the farmers for this vitamin, right?

Lank said...

"Question is, if ANE and WHG are something related to "East Asian", why is the main WHG male lineage "I" (today (all "I" clades) mainly in Northern and Eastern Europe), a brotherclade of "J" (today mainly in West Asia).

The Y-DNA suggests a deep anchestry relation of WHG with near eastern populations at least on the male lineage (well with the exception of a single "C".) "

According to the model presented originally by Lazaridis et al., Near Easterners are a mixture of "Basal Eurasians", and a West Eurasian population related to WHG and ANE. The revised version of that paper even suggested that WHG and the West Eurasian component in the Near East form a clade, and ANE forms a parallel clade. This new paper by Skoglund simplifies the model and refers to all of the blue in European farmers as ancestry "related to hunter-gatherers", even though a significant portion of this component in the farmers actually arrived from the Near East, along with Basal Eurasian.

Davidski said...

Yes indeed, they delved into that issue a little more in the revised manuscript as a result of one of my comments at bioRxiv. I have it in writing. ;)

About Time said...

So in Fig S7, two things:

Sardinians and Iceman but not Gokhem2 have Dinka admixture.

The line from the Denisova clade to Australia shares some but not all of the Denisova drift. Has implications for when or where admixture took place: it must have been prior to full differentiation of Denisova, closer in time or space to the common modern/Denisovan root populations.

Anima Mortel said...

"By the way, the dark skin pigmentation of the hunter-gatherers was definitely a surprise, but perhaps not a mystery."

Yes a surprise not a mystery. Especially after Motola12 was derived SLC24A5 in Apr. Lazaridis and now a hunter gatherer much later than Motola12 is again as dark as Loschbour and La Brana. Surprised and puzzled.

At least there is a stable continuity with the farmers Stuttgart, Gok2 and Oetzi although Stuttgart was ancestral to SLC45A2 (but derived SLC24A5) but Stuttgart was also long before Oetzi and Gok2.

"European hunter-gatherers probably ate a lot of vitamin D-rich fish, meat and mushrooms, so they didn't need to rely on the sun as much as the farmers for this vitamin, right?"

I do not know but probably a reason.

A true mystery however are the PCA plots from Skoglund. Not only do they considerably differ towards other papers like Lazaridis and Raghavan but what exactly is the reason why Oetzi is closest to the Tuscans and MA1 (Mal'ta) is closest to the Finns in this plot?

Davidski said...

Projection bias. See here...

http://polishgenes.blogspot.com.au/2013/10/ancient-dna-from-prehistoric-bulgaria.html

Scroll towards the end.

Shaikorth said...

Gokheim doesn't get Dinka because its "basal eurasianness" or whatever is low enough (thanks to hunter-gatherer admixture) to fit in a derived position of an admixture tree without it.

Sardinians and French in S7 don't look admixed with ancient genomes due to their more recent shared drift so instead French fit via Sardinian admixture (and possibly Dinka admixture that comes with it) and Sardinians fit via Dinka admixture.

Anima Mortel, it's because, as the supplementary information says, all of the ancient genomes are projected into the PCA, just like they are in the West Eurasian PCA from Lazaridis et al. This affects their position considerably. To give an example of how much distortion it causes, the Saami who are clearly removed from others in the Skoglund PCA would cluster between Orcadians and Russians if they're projected, as done here:
http://www.nature.com/ejhg/journal/v19/n3/fig_tab/ejhg2010179f3.html#figure-title

spagetiMeatball said...

Aren't they missing out on a lot of data by not performing any tests on existing middle eastern populations and remains? I feel like they're guessing around with the whole cryptic "basal eurasian" thing when the answer might just be a few more sequenced genomes away. They've been concentrating on bones from central/western europe though.

About Time said...

@spaghettiMeatball, Egypt, Jordan, Israel, Lebanon are gold mines of ancient remains that were carefully preserved in situ and preserved by a dry climate. Of those, Israel is well excavated and funded, with lots of qualified scientists to do genome extractions from PPNA right up through the Iron Age.

No excuses for not getting in there and publishing some data. Truly. Let's see what the bones of the ancestors have to say.

Anima Mortel said...

@ Davidski

Thank you. That explains such a result.

@ Shaikorth

I understand that part.
But Lazaridis has a different position of the Ancient samples in his PCA plots. And Fig. S5. in Skoglund shows a different closeness than his PCA plot.

As for the Admixture graph.
As far as i understood is the lines and especially their colors only indicate the migration weight with yellow being close to 0.
And the K20 admixture analysis in Lazaridis did not show an African admixture in Stuttgart or Sardinians.

Davidski said...

The position of an individual on a PCA plot, whether projected or not, depends on the choice of the samples in the analysis.

Also, ADMIXTURE is often unable to detect minor ancient ancestry in samples that have experienced a lot of genetic drift since the admixture event. So there's still a chance that the so called Basal Eurasian signal is just some sort of Northeast African ancestry.

Anima Mortel said...

"The position of an individual on a PCA plot, whether projected or not, depends on the choice of the samples in the analysis."

But that still would not explain why Oetzi is now closest to the Tuscans or most of all MA1 is closest to modern Finns. The Skoglund PCA plots do not seem very reliable. Which was the same case before and his PCA plots in his 2012 paper about Gok4 which were greatly altered in Lazaridis.

Davidski said...

The PCA from this latest Skoglund paper, which I posted above, includes the Saami and Mari. These two groups have significant East Eurasian admixture, and the Saami also appear to be heavily drifted. The Lazaridis et al. PCA lacks such groups, and this might be enough to produce slightly different results in PC2 for the Sardinians and/or Oetzi.

Certainly, the presence of the Sami appears to be the key to MA1's unusual position on this plot. Their drift seems to push him west.

Grey said...

@me
"ANE isn't eastern or western it's northern - stretching across a vast area from Europe in the west to Siberia and beyond in the east."

@Colin Welling

"I doubt it's even been in Europe since the ice age. Mesolithic WHG groups are very homogeneous and both brana and loschbour lack ANE completely. ... That suggests Western and Southern Mesolithic Europe lacked ANE."

ANE extending from Siberia to Europe doesn't necessarily mean all of Europe if the reason for the extent of their range was following particular herd animals. The range would follow the range of the herds.

So the range wouldn't extend to Iberia, Italy or along the Atlantic coast.

.

"If we assume that the ancestors of WHG split from MA/ANE before the ice age then where in Europe would the ANE hide during the ice age?"

They wouldn't be hiding anywhere they'd be in the interior parts of Europe that were part of the mammoth steppe.

.

"I suspect ANE traveled into Europe, during the Mesolithic, along the far north (Northwest Russia, Baltic States, and Scandinavia) but the bulk of ANE penetration came from the steppes during the late neolithic."

I agree with that mostly except I think the ANE followed the herds they hunted back and forth across the mammoth steppe from Europe to Siberia hence why they are a distinct genetic population.

There weren't separate European mammoth hunters, central Asian mammoth hunters or NE Asian mammoth hunters (and eventually north American mammoth hunters). They were basically the same population moving back and forth with the herds.

.

"BTW, does anybody know why EEF is "African shifted". Does that mean EEF has SSA admixture or that Basel Eurasian split from Africans later than the rest of the Eurasians."

Because "Basal Eurasian" isn't basal. It's a WHG-like population mixed with an Afroasiatic population.

"Basal Eurasian" is what WHG came from or the layer before that.

Either (imo)

ASE
->(+SSA)->Afroasiatic
-> WHG
-> ANE
-> ENA

or

ASE
->(+SSA)->Afroasiatic
->ANE->WHG
->ENA

or

ASE
->(+SSA)->Afroasiatic
->WHG->ANE
->ENA

or something like that.

.

"as C-V20 in La Brana means some sort of link to East Asia"

It doesn't necessarily mean a direct link though. If you have a population in region A and some of them move west to region B and some move east to region C then the common link is region A.

.

@Anima Mortel

"It remains a mystery why the hunter gatherers were very dark skinned but the farmers were all light skinned."

Both ANE and WHG are particularly large and heavy boned with no sign of rickets which implies they got plenty of Vitamin D - seemingly more than the farmers.

Living beside the coast and eating a lot of seafood might provide that for some but what about the ANE inland?

The physical size and robustness is a hint that they were already depigmented when the farmers arrived via MC1R and / or IRF4.

Shaikorth said...

Davidski, Saami or Mari are unlikely to be the reason for projected MA-1's "western" position. In the Lazaridis PCA of Europe from the updated article, he is projected right next to Mordovians and North Russians even though there is nothing eastern pushing him there.

The presence of the Saami in the Skoglund PCA seems to have no other effect than pulling Finns to the east of Mordovians and North Russians. Otherwise that PCA actually looks a lot like one of your West Eurasian PCA's.

Unknown said...

Interestingly, Y-DNA A1b1b2b-M13, which reaches its peak frequency among the Dinka, has been found in a sample of Sardinians (Semino et al 2000).

Maju said...

@Grey: "ANE" is Ma1 (and nothing else). Ma1 is a genetically and culturally West Eurasian (not quite the same as "European" or "West Asian" but actually a third subclade of the wider one), which is most directly related to the origin of Native Americans and, variably, of many Siberian populations. Ma1 is NOT related with mainline East Asians south of Siberia. Whatever genetic influence Ma1 and proto-Amerindian populations one held on those parts of East Asia has now been almost totally erased.

So Ma1 is "Eastern" in West Eurasia (the NE branch of around Altai in fact) but not "Eastern" in Eurasia in general.

In Europe its inferred genetic influence ("ANE") is clearly associated to Eastern Europe by origin. We have no idea right now of when WHG-like and ANE-like peoples began mixing in Eastern Europe. But judging on the fact that Eastern European Upper Paleolithic diverged from the rest of Europe definitely precisely in Gravettian times and that the cultural context of Ma1 is Gravettian, it's perfectly possible that the ANE component has been flowing to Eastern Europe since that time.

The fact that Epipaleolithic Scandinavians (Motala) show some ANE admixture by that time strongly suggests that the ANE component had already arrived to Northern Europe already in the Late UP or at most Epipaleolithic (Hamburgian-Ahrensburgian or whatever other cultural context is Motala associated to). It seems quite likely that all the North Sea area may have been affected by this Late UP/Epipaleolithic Nordic expansion of ANE genetics by that time, although this tentative conclusion heavily relies on my guess that Motala is Ahrensburgian, as is the case with other Scandinavian sites (but also Scottish, NW English, Low German and Dutch). This explains the NE tendency in, say, Irish, without need to resort to massive Viking or Celtic demic replacements in the islands.

So not everything ANE needs to be directly Indoeuropean and, in any case, ANE is not "Eastern" as in "Chinese" but as in a NE geography within West Eurasia. True East Asian genetics should be measured independently (but clearly have a much more restricted distribution in Europe, mostly Uralic peoples).

truth said...

@Anima

That's not Otzi clustering with Tuscans, it's an Iberian farmer

Maju said...

It's Ötzi. When discussing the samples used, the specifically say: "the Chalcolithic (~5,300 BP) Tyrolean Iceman (9)"

Anima Mortel said...


If Gok2 has no Dinka admixture and is 77% hunter gatherer ancestry than why does he cluster next to the Spanish who had the highest amount of African admixture (Lazaridis K20) and the lowest amount of WHG?

And if Gok2 is overwhelmingly WHG (77% hunter gatherer) than why is he nowhere near the other hunter gatherers. The K4 in the Skoglund 2012 paper about Gok4 suggested nothing like 77% hunter gatherer.

The way is till see it because its written as such in the paper is that the arrows (lines) in the Admixture graph only indicate the migration weight with yellow being close to 0. And the other results in that paper suggest that Gok2 just like Gok4 (in the 2012 K4 admixture analysis) is not really that much hunter gatherer.

Not sure if Skoglund wanted to proof a point to Lazaridis but much is not conclusive and even contradicts its own results and data.

Anima Mortel said...

"The physical size and robustness is a hint that they were already depigmented when the farmers arrived via MC1R and / or IRF4."

The 6 SNP's of MC1R tested by Loschbour and Motola12 were the same as by Stuttgart and none suggested the lightening mutations. IRF4 / rs12203592 is associated with eye color and freckling. But if an individual is missing the 2 skin lightening alleles (rs1426654 A and rs16891982 G) i am not sure how that freckling would look like. Eye color was blue by the hunter gatherers but hair color was also dark.

Davidski said...

Ahh, here's what I should've quoted in my comment above about the different outcomes from the two TreeMix analyses:

"Modern-day non-Africans are complicated to fit in the same framework since the genotypes from these are of much higher qualities than the haploid single sequence-read genotypes from the ancient data, and thus there is a risk of data quality affecting the inferences (for example by error-driven attraction between ancient genomes)."

Page 29 of the supplement PDF.

Maju said...

Anima: regardless of NE African admixture ("Dinka"), which may still exist in Gök2, she and Ötzi share a wider common EEF origin, as should be obvious by them forming a clade in all trees, what implies the same sort of West Asian + Balcanic HG original combo, as discussed in Lazaridis' paper.

The blue color in the graphs is not WHG as in Lazaridis 2013 but something else: probably just meaning "other (or core) West Eurasian", including WHGs but also other European and West Asian ancestry, none of which is discerned in the present study.

The ADMIXTURE analysis you mention is probably accounting for other sub-components, which are not accounted for in Skoglund & Malström's graph. Anyhow it's not necessarily more accurate either.

barakobama said...

"The 6 SNP's of MC1R tested by Loschbour and Motola12 were the same as by Stuttgart and none suggested the lightening mutations. IRF4 / rs12203592 is associated with eye color and freckling. But if an individual is missing the 2 skin lightening alleles (rs1426654 A and rs16891982 G) i am not sure how that freckling would look like. Eye color was blue by the hunter gatherers but hair color was also dark."

Motala12 had A/A alleles in rs1426654, A/A alleles in rs28777, and was not tested for but probably had G/G alleles in rs16891982 because in modern Europeans if an individual has G/C or G/G alleles in rs16891982 they will almost always have A/C or A/A alleles in rs28777. sf11 a 7,500-7,250 year old hunter gatherer from the Baltic sea had A/A alleles in rs16891982, so that mutations existed in Mesolithic Europe.

If you go by the 8-plex system Motala12 is light skinned, but if he did not have A/A alleles in rs16891982 he is at least light to medium skinned. Do you really think skin color varied in small Mesolithic European tribes from deep brown to snow white? You should not assume they had a certain skin color based on mutations that are not European-specific and can't explain skin variation in west Eurasia.

It seems no one noticed that Olalde 2014 was hesitant to say La Brana-1 had dark skin, because he has no close modern relatives who are missing have the same alleles in pigmentation-related SNPs as he does and these SNPs may not be able to explain all modern skin color variation.

At GEDmatch i have found some northern Europeans who may be good proxies of what skin color many Mesolithic Europeans had because of their alleles in many pigmentation-associated SNPs(including mutation in gene IRF4 and blue eyes). I think i was wrong about my uncle because at GEDmatch there were plenty of Latinos with the same alleles as him, so the dark skin is probably from his Native American-African ancestry.

Anima Mortel said...

@ Maju

Maybe it would have been better if Skoglund created a chart and gave the figures in plain numbers rather than a Graph with a color code. And explaining what actually is meant by the hunter gatherer ancestry would have been helpful as well.

@ barakobama

That is what i have written in the entry of (April 26, 2014 at 5:48 AM) that Motola12 is derived SLC24A5 and yes also StoraForvar11 is derived SLC45A2 which makes Ajvede58 being ancestral to both (SLC45A2 and SLC24A5) the more surprising. I expected after Motola12 that the so called Scandinavian hunter gatherers had a different pigmentation than the so called Western hunter gatherers. But Ajvide58 who lived long after Motola12 and StoraForvar11 is just like LaBrana and Loschbour in that respect.

"sf11 had A/A alleles in rs16891982, so that mutations existed in Mesolithic Europe."

I think you mean G/G in rs16891982.

jackson_montgomery_devoni said...

I may be wrong but it seems rather likely that the ancestry that the Neolithic farmers have that is related to the WHG component/clade may be connected to Y-DNA haplogroup J. Since Y-DNA halogroup I is the main Y-chromosome haplogroup of the WHG population and I is most closely related to J then it would make sense that the autosomal component/clade among the farmers that is similar to the WHG component come from probably a Near Eastern population high in Y-DNA haplogroup J.

Grey said...

@Maju
"So Ma1 is "Eastern" in West Eurasia (the NE branch of around Altai in fact) but not "Eastern" in Eurasia in general."

Yes, the center of gravity is eastern (relative to Europe).

I'm saying the *range* is very wide and the western edge of the range extends well into Europe so - if ANE were originally connected to mammoth - then there's no mystery why ANE genes would be so far west along the north European plain. There were mammoth as far west as Britain for a time.

.

Thinking more on "Basal Eurasian" and the possibility of two farmer waves.

Sumer and Akkad tells of an Afroasiatic population taking over the first farmer population in the fertile crescent so what about:

ASE (somewhere around India)
->
WHG/SHG (Europe through to Fertile Crescent)

1) The segment of WHG around the fertile crescent develop into the first farmers

2) The segment of WHG nearest Africa mix with SSA to become Afroasiatics

So now you have
WHG (Europe)
WHG farmers (fertile crescent)
Afroasiatics (WHG + SSA)(Arabia)

first wave of farmers spread in various directions

Afroasiatics conquer first farmers in the fertile crescent creating "EEF" i.e.

"EEF" = WHG farmers + Afroasiatics

then this Afroasiatic admixed "EEF" spread in various directions as a second wave?

Maju said...

@Grey: as you ask me, here is my two cents.

1. ANE:

MA1 is from deep into Siberia, although of West Eurasian origin (i.e. product of Western "Aurignacoid" colonization of Altai plus of whatever Gravettian influences across the steppe).

While ANE admixture is apparent in some Scandinavians very early on (Motala Epipaleolithic of unknown affinities), this is not pure ANE at all but just some lesser admixture, almost certainly mediated by Eastern European HGs somehow. Possibly Eastern European HGs were somewhat higher or at most similar in that admixture but they were not either "pure ANE" even remotely. Probably modern Lithuanians approximate relatively well that Eastern European HG pool, although with some EEF/WHG admixture, of course.

2. Basal Eurasian and such:

I would not use the acronyms you do first of all (WHG is not reserved for Western European HGs, ASE seems an unlikely term but also as concept distinct from ANE).

I'd rather think of proto-Westerners (PW) originating near Pakistan and from them:
·PW → ANE (Central Asia)
·PW → Ancient Levantine (with extension to NE Africa)
·PW → Ancient European HGs (splitting into WHG, EEHG and surely also some smaller groups like Italian HGs, Balcanic HGs, etc., twice before the LGM)
·PW core around Iran, retaining greater basal diversity

Add to that probably another branch via Yemen to The Horn of Africa. And also WHG → NW Africa (Oranian, with some backflow to West Iberia).

In the Late UP/Mesolithic, the Afroasiatics spread from around Sudan in mostly Northwards direction. One of their branches reached Palestine where it formed the Harifian (seed of the later Semitic expansion) and surely, somehow, affected the proto-EEFs, which may have coasted by Anatolia to Thessaly, where they mixed with Balcanic HGs (UHG in Lazaridis) to become the EEFs.

I do not see any reason to think of two waves but it's of course possible that both in the Anatolian journey, as later in the Balcans, the resulting subpopulations may have got slightly varied frequencies of each of the components involved in that genesis by means of founder effects.

Other West Asian populations lacking that Afroasiatic genetic element are the relevant ones in other Neolithic expansions, notably to South Asia. Very different languages involved also: Dravidian does not look at all related to Vasconic, nor either one to Semitic.

Maju said...

Erratum: "WHG is not reserved for Western European HGs" should read "WHG is NOW reserved for Western European HGs" (exactly the opposite!)

barakobama said...

I found quite a few northern Europeans who have very similar alleles in SNPs associated with pigmentation as do Mesolithic European samples.

Here is one individual who gave me some feedback.

Ancestry: North-west European+2% Native American: rs12203592 T/T(in gene IRF4), rs16891982 C/G(in gene SLC45A2), rs28777 A/C(in gene SLC45A2), rs1426654
A/A(in gene SLC24A5), rs1042602 C/C(in gene TYR), and h1 haplotype for blue eyes.

EEF:46.9

WHG:37.7

ANE:15.3

Loschbour: rs12203592 T/T(in gene IRF4), rs16891982 C/C(in gene SLC45A2), rs28777 A/C(in gene SLC45A2), rs1426654
G/G(in gene SLC24A5), rs1042602 C/C(in gene TYR), and h1 haplotype for blue eyes.

He described himself as light skinned but can tan very well, black haired, and blue eyed. Besides A/A alleles in SNP rs1426654 he is almost a perfect match with Loschbour.

I found that a very large number of people with T/T alleles in SNP rs12203592 have blue eyes, i wonder if that combination is some type of Mesolithic European haplotype.

I am still very undecided on the pigmentation of typical Mesolithic Europeans but my guess at this point is black haired, brown skinned, and blue eyed. Hopefully i will get feed back soon from the other Europeans missing light skin mutations. If i have strong faith in these SNPs then Motala12 should have been very light skinned, but why would skin color vary from brown to white in small family tribes?

Ryan said...

@Grey - Yah, just to be clear, I don't mean the people actually living in East Asia now. Just that there is some common link between Jomon C-M8 and La Brana C-V20.

I guess that shouldn't as surprising as it was initially though. People have been suggesting Central Asia (http://link.springer.com/article/10.1007%2Fs10038-005-0322-0) and Lake Baikal (http://link.springer.com/article/10.1007%2FBF01790088) as the origin of Jomon people at around the time that Ma-1 lived.

Grey said...

@Maju

"While ANE admixture is apparent in some Scandinavians very early on (Motala Epipaleolithic of unknown affinities), this is not pure ANE at all but just some lesser admixture, almost certainly mediated by Eastern European HGs somehow."

Sure. If the center of gravity of ANE is somewhere in central north eurasia but with a very wide range then there would likely be some admixture with the HGs in the west e.g. between western HGs and ANE or western HGs and ANE mixed individuals.

My point is you don't need *tribes* of ANE moving from Siberia to Scotland or Scandinavia to explain ANE admixture in Scotland or Scandinavia because - if ANE were connected to mammoth - their range across north eurasia was huge and extended far west.

I'm not saying you're saying this but some comments I've read seem to suggest some people believe ANE lived in Siberia and then migrated west as a group.

Grey said...

@Ryan

"Just that there is some common link between Jomon C-M8 and La Brana C-V20."

Yup.

Grey said...

@barak

" If i have strong faith in these SNPs then Motala12 should have been very light skinned, but why would skin color vary from brown to white in small family tribes?"

*If* IRF4 and/or MC1R are relevant to skin depigmentation but it is being masked by the near fixation of the SLC genes and one or both are recessive (IIRC MC1R is but I don't recall if IRF4 is) then maybe you need two copies of the gene to get the full effect.

red version of MC1R (without SLC genes)
0 copies = dark skin
1 copy = semi-dark skin
2 copies = light skin

red version of MC1R (with SLC genes)
0 copies = light skin
1 copy = light skin
2 copies = light skin

(just a guess)

Shaikorth said...

To be honest I wouldn't be surprised if proto-Indo-Europeans turn out to be nothing more than another bunch of Neolithic or WHG/SHG-type folks like the ones we've seen.

Populations can be modelled multiple ways in the Lazaridis model. We know that Basques for example can be modeled with or without ANE, and it wouldn't be surprising if for example English could be modeled with reduced (but still present) ANE and increased WHG besides the single fit they published.

Maju said...

@Grey:

What I think most likely is that there was first a colonization of Eastern Europe from Central Europe (twice: Aurignacian first, Gravettian later) and that explains on its own that Eastern European HGs were like WHGs but not quite the same (shorter IBD re. Lochsbour and such). Later these Eastern European HGs would have got whatever interactions with the third population (Ma-1, ANE).

In synthesis:

Early Euro HGs → Eastern Euro HGs (1) + some ANE admixture = Eastern Euro HGs (2).

... "if ANE were connected to mammoth - their range across north eurasia was huge and extended far west".

Then why didn't Lochsbour show any sign of ANE admixture? Nope. I think that you hype ecological conditions and very specially mammoths. Central Europeans for example hunted mammoths but they left no ANE-related legacy.

I rather think that these ancient populations had their "countries" of sorts (cultural regions) and that between these there was just "semi-deserts" (very low population areas) most of the time. ANE's "country" was Altai in essence and EEHG's "country" was the Dniepr-Don area. They must have interacted at various times but they were also distinct all the time.

"... some comments I've read seem to suggest some people believe ANE lived in Siberia and then migrated west as a group".

I think that ANE lived in Central Siberia and probably also Central Asia and interacted loosely with Eastern Europe, notably in the Gravettian period. That should allow for admixture at the levels that we perceive: Russians are quite less ANE than Native Americans, even if these are largely East Asian (i.e. not pure ANE at all).

barakobama said...

Grey, i found some more modern north-west Europeans at GEDmatch missing SLC and TYR light skin mutations, and all so far are typical light skinned Europeans. I found one who has rs16891982 C/C(in gene SLC45A2), rs28777 A/C(in gene SLC45A2) like Loschbour hopefully i will get some feedback. I am even more confident than before that there are unknown causes to European light skin.

Shaikorth said...

The closest population to pure ANE in modern Eurasia are Central Siberian Kets (in both shared drift and IBS), followed by Selkups (in IBS sharing only).

The Siberian admixture in the HGDP Russian sample interestingly seems to be totally comprised of a component specific to Kets and Selkups(in contrast to Chuvash who have "Nganassan" and some "Yakut" too) in tests where such a component appears, like this one from Fedorova et al @ k13.

http://www.biomedcentral.com/1471-2148/13/127/figure/F6?highres=y

Ryan said...

@Shaikorth - "The closest population to pure ANE in modern Eurasia are Central Siberian Kets (in both shared drift and IBS), followed by Selkups (in IBS sharing only)."

Which is notable given that the Kets speak a Dene language.

Grey said...

@barak

" i found some more modern north-west Europeans at GEDmatch missing SLC and TYR light skin mutations, and all so far are typical light skinned Europeans."

That's very cool. I thought there might be some but from what people said about the SLC genes being fixed I thought they'd be like a needle in a haystack - great job.

Grey said...

@Maju

"I think that ANE lived in Central Siberia and probably also Central Asia and interacted loosely with Eastern Europe, notably in the Gravettian period. That should allow for admixture at the levels that we perceive: Russians are quite less ANE than Native Americans, even if these are largely East Asian (i.e. not pure ANE at all)."

Sure, that explains Eastern Europe but the bit that wants explaining is Scotland and Scandinavia.

And that is easily explained by assuming ANE covered the whole extent of the mammoth range which extended far enough west for them to have mixed directly with a more maritime distributed WHG.

http://hanskrause.de/images/HKHPE10/image006.jpg

http://img571.imageshack.us/img571/6525/7dt9.png

There are a lot of maps of the mammoth steppe and no doubt they're not 100% accurate but they all show basically the same thing - a center of gravity in central north Eurasia with a very wide range west to east.

It seems to me admixture between ANE and WHG at the far western end of the range is the simplest explanation (at the moment) for at least some of the ANE present in Scotland and Scandinavia.

Ryan said...

@Grey

Why is ANE in Scandinavia all that surprising? Uralic peoples have been present in the area for a while.

barakobama said...

"That's very cool. I thought there might be some but from what people said about the SLC genes being fixed I thought they'd be like a needle in a haystack - great job."

I am conducting a study at GEDmatch(on hundreds or thousands of people), to see see for myself how much of an effect these SNPs really make on skin color and how accurately hair and eye color can be predicted.

I got some feedback from the north-west Europeans with rs16891982 C/C(in gene SLC45A2) and rs28777 A/C(in gene SLC45A2) like Loschbour.

She says she has 'ivory" skin that tans very quickly and easily, medium brown hair(not black like she should have) and that her mother and brother are very dark skinned like Mexicans. In my email to her i mentioned that Mesolithic Europeans had ancestral dark skin alleles like her and that she is one of the best modern proxies for their skin color i could find at GEDmatch, so she may be exaggerating the darkness. One of the others said they had very olive skin but the pictures showed they were typical light skinned Europeans.

Mesolithic Europeans were most likely dark skinned in my opinion but they may have been light skinned, and Motala12 was most likely light skinned. They were probably mainly dark haired and blue eyed, but strangely both Loschbour and Stuttgart(farmer) had a mutation associated with red hair and 0% of CEU(north-west European Americans) have it.

Maju said...

"Sure, that explains Eastern Europe but the bit that wants explaining is Scotland and Scandinavia".

Indeed. My first thought is that we have to look at Hamburg-Ahrensburg culture, whose ultimate origins are unclear, at least to me. Maybe it originated in the East? If this ancient North Sea population was already carrier of ANE ancestry, it would explain many things.

However Motala, the only Nordic Epipaleolithic people studied at this level to date, is by the moment unclassified, at least I could not find their cultural affinity, unlike several recent findings from Scotland, Norway, etc. which are quickly described as Ahrensburgian. So it's unchartered territory by the moment.

Davidski said...

Ryan,

Present and former Uralic and Turkic areas of Europe are the only regions where ENA is found on the continent at appreciable levels. So ENA, and more specifically post-Mesolithic Siberian admixture, appears to be a genome-wide genetic marker of the early Uralic and Turkic groups. See here...

http://eurogenes.blogspot.com.au/2014/01/another-look-at-lazaridis-et-al-ancient.html

Thus, if Uralic people were present in Scandinavia during the Mesolithic and/or early Neolithic, and interacted in any meaningful way with hunter-gatherers from southern Sweden and Gotland, then we should expect ENA admixture among these hunter-gatherers. But it doesn't appear to be there.

Shaikorth said...

We don't really know whether Ajv58 had ENA from this tree, because the components it represent are themselves eastern-shifted. They'd have to add something like the HGDP Russians instead of Sardinians and French to their tree and see whether admixture from Saqqaq/Anzick appears in them to ascertain that.

In any case, compared to La Braña Ajv's increase in affinity towards mostly ENA Anzick-1 is greater than towards pure ANE MA-1. This is one of the reasons why this matter needs further investigation, like checking its IBS with East Asians and comparing it to modern East Euros etc.

When it comes to post mesolithicness and specific associations of autosomal components to only certain language groups, we'd need DNA from Oleny Ostrov site and from eastern europe to ascertain the date of ENA's appearence and its absence from assumed Indo-European remains.

Davidski said...

The early Indo-Europeans evidently passed on ANE to modern Europeans, but no ENA. This suggest they didn't carry any ENA.

On the other hand, early Uralics and Turks did pass on ENA to their descendents. Even Estonians still have a bit of it.

Shaikorth said...

The ANE in Europeans is not necessarily an Indo-European signature at all though, see my earlier comment about the Lazaridis fits. To ascertain whether the assumed Volga-Ural region Indo-Europeans had no ENA we'd need ancient DNA from post-Mesolithic remains from the region. Mesolithic ENA in Europe would rule out associations between modern languages spoken in the continent and spread of these autosomal components.

From modern populations we can ascertain little, small levels of some componenent can be wiped out to noise levels within a few generations and the genetic outliers are also geographical outliers. Estonians can fit in the Lazaridis three-way model - which makes it a not-so-good indicator of past or present ENA - and who knows what else would fit when swapping Loschbour for AJV58 (which is not symmetrically related to ANE and ENA/ANE mixes compared to WHG's) or Stuttgart for Gökheim.

Davidski said...

It can't be a coincidence that ANE spread across most of Europe at the same time as Indo-European languages (ie. after the early Neolithic but before the Uralic and Turkic expansions carrying ENA). There's no other genetic signal we can attribute to the early Indo-Europeans.

Also, this version of ANE was indeed of a different character than that already present in Scandinavia during the Mesolithic. From Lazaridis et al., page 67.

"K=15 shows the appearance of a component, maximized in the Kalash, that becomes the most predominant signal in Indus Valley and Caucasus populations. It is also prominent in the rest of South Asia, Central Asia, Near East and in diminishing strength in Europe. It is absent in Sardinians, Basques, and all ancient Europeans, although it is present in MA1. This component also does not appear in North and East Africa (except for Egyptians and Tunisians) where other West Eurasian admixture is observed. This is consistent with an MA1 related population having contributed some ancestry to present-day Europeans not accounted for by West Eurasian Hunter Gatherers and Early European Farmers. The presence of this component in the Near East contrasts with its absence in Stuttgart, consistent with the widely shared negative f3(Near East; Stuttgart, MA1) statistics (Table 1) indicating that present-day Near Easterners have been affected by gene flow not present in early Near Eastern migrants into Europe. Interestingly, a study of present-day South Asian populations (18) also revealed the presence of a common ancestral population (labeled “k5” there) between northern parts of South Asia, the Caucasus and Central Asia and Europe (but not Sardinians). The absence of the similar component in ancient Europeans reinforces our idea that present-day Europeans have ancestry from a third ancestral population after the Neolithic transition and that this may be related to the evidence for admixture between Ancestral North Indians (related to the “k5” population) and Ancestral South Indians that took place over the last ~4,000 years (19)."

Shaikorth said...

ANE might not be as widespread in Western Europe as one might deduce at a first glance - in fact it's likely that most Europeans can be modeled with increased WHG and less ANE like how Basques can be modeled without ANE. I'm not going to attribute any autosomal signals to modern European languages unless ancient DNA from Eastern Europe proves otherwise. It would also be nice to see how Northeast European (of which Estonians already fit) populations fit into models using North European, instead of Western European, hunter-gatherer genomes. If they do it removes need for post-Neolithic autosomal contributions *unique to carriers of modern language groups* as well.

The Kalash component appearing at high K values could well just be a signal of their high drift, it's not the first time such a component appears and putting modern higher quality genomes into admixture runs with ancient genomes of varying coverage can have funky effects as we see in East Eurasian type admixture they model into MA-1 and so on. We see other oddities in that admixture run, like for example all hunter-gatherers eventually turning into 100% Loschbour though that is certainly not what they are. The Kalash component changes again at K20, into a component that doesn't appear in Europeans but does appear in MA-1 which, if anything, is only indicative of problems admixture analyses have in these situations.

Grey said...

@Maju

"Indeed. My first thought is that we have to look at Hamburg-Ahrensburg culture, whose ultimate origins are unclear, at least to me. Maybe it originated in the East? If this ancient North Sea population was already carrier of ANE ancestry, it would explain many things."

Sounds reasonable.

Grey said...

@Ryan

"Why is ANE in Scandinavia all that surprising?"

It's only surprising if you assume ANE = Siberia. It's not surprising at all if you don't.

Davidski said...

Well, you can remain sceptical if you like, but it's difficult to argue now that there wasn't a major episode of gene flow from the east into Central Europe during the Copper Age, including Y-hg R1a and mtDNA U2e, for instance, with the spread of archeological cultures associated with the early Indo-Europeans before their remains were ever tested for DNA, like those belonging to the Corded Ware and Unetice cultures. It's very likely that this is also when a fresh wave of ANE moved across the continent. So I'd say that everything fits, and should look even better as more ancient genomes come in from prehistoric Europe, in particular Eastern Europe.

Grey said...

@Davidski

"It can't be a coincidence that ANE spread across most of Europe at the same time as Indo-European languages (ie. after the early Neolithic but before the Uralic and Turkic expansions carrying ENA). There's no other genetic signal we can attribute to the early Indo-Europeans.

Also, this version of ANE was indeed of a different character than that already present in Scandinavia during the Mesolithic."

What sounds most plausible to me is initially a wide spread of related ANE HG groups across northern Eurasia until one of those groups develop a horse culture and spread over a much wider area.

If that is correct then the pattern should be - if R1a is the signal - a number of regionally concentrated branches of R1a with one particular branch which is much more widespread.

Davidski said...

The thing about R1a is that it's not very diverse. Almost all of it falls under R1a-M417, which is about 6,000 years old, and its two main branches, R1a-Z282 and R1a-Z93, dominate Eastern Europe and South Central Asia, and spill out into neighboring areas.

Maju said...

"There's no other genetical signal we can attribute to the early Indo-Europeans".

While totally agreeing with this assessment, I'm also puzzled by it: how can an estimated >25% Indoeuropean (or otherwise ANE-influenced) ancestry be passed around without any obvious association to uniparental genetic markers? Other quite more subtle ancestries such as East Asian in Uralic peoples or NE African in early farmers and their modern descendants are directly associated with Y-DNA lineages N1 and E1b1b1 respectively but in this case what? Nothing specific: Spaniards look ~25% IE (30% more ANE than Basques, so roughly 1/4 of their autosomal ancestry is IE) but their R1a is negligible (~1% overall, <5% in the most signified areas).

This has me quite a bit worried.

Of course one can imagine a model in which successive admixture events diluted the IE Y-DNA input. Or also that we should not just look for R1a but maybe also Nordic variants of I or something like that, which, who knows?, maybe a hidden IE (extra ANE) signature. But still I'd expect ~25% German-like Y-DNA signature (assuming ANE arrived with Celts) and I seriously doubt you can find anything like that in Iberia, judging on the ridiculously low frequencies of R1a alone.

Then again part of that extra ANE (IE genetics) may have arrived from Italy with the Romans. If so, should we consider Italian affinity in Iberian Y-DNA, particularly J2b, an IE marker locally? But then again are Italians ANE enough to leave such a mark with so little Y-DNA input?

Does anyone have a half-decent idea of how could this happen?, how can this make sense?

Grey said...

@Davidski

"The thing about R1a is that it's not very diverse."

I think dna at the end of a migration chain will always be less diverse.

If you have a population A and part of it moves to region B and then part of region B moves to region C and part of region C move to region D then the people at D will be a subset of a subset of a subset of A's diversity.

This is before you take into account bottlenecks caused by people adapting to new environments - especially if the environment was harsher than the one they moved from.

If correct then Bedouin or Tuaraeg might be an example of the same thing (assuming adapting to a desert caused a bottleneck).

Maju said...

@Saikorth: see my previous comment focused on Iberian way-too-high ANE.

Otherwise, even if Scots might have most of their ANE inherited from Ahrensburgian first settlers, it is clear that this is not the case south of Rotterdam in the mainland, and surely not in most of the rest of Britain & Ireland, which has limited to no Ahrensburgian/Maglemosean roots, never mind a clearly strong Neolithic settlement (judging on dramatic diet shift and other archaeological evidence), which in any case would be close to Lochsbour of all WHG subpopulations (origins in NW France).

So, if I'm correct, other Western Europeans also seem to lack uniparental markers (focusing on Y-DNA especially) in enough frequencies to justify their ANE (and hence IE) autosomal scores.

Ryan said...

Davidski - Yah, for some stupid reason I thought people were talking about modern Scots and Scandinavians. :/

"ANE was actually present in Scandinavia during the Mesolithic, because Motala12, the 8,000 year-old hunter-gatherer genome from Sweden, has an ANE ratio of 19%." - From your post about Lazaridis et al.

Why couldn't some amount of ANE have been present early on on the fringes of Europe? I wouldn't think there would be some sort of massive barrier between reindeer and mammoth hunters in Eurasia and red deer hunters in Cantabria once the ice sheets began to retreat.

The Swiderian culture supposedly migrated to the north east as the climate improved too didn't they? Perhaps they had some level of ANE in them, and as the reindeer retreated, so did they.

Either way, your suggestion of two layers of ANE (one being Indoeuropean in origin) makes sense.

You're probably right about it being too early for Uralics that far West. ~5000 BP they were supposed to be busy getting loanwords from Indoeuropean languages, which would place them farther east.

"Well, you can remain sceptical if you like, but it's difficult to argue now that there wasn't a major episode of gene flow from the east into Central Europe during the Copper Age, including Y-hg R1a and mtDNA U2e"

I'd add mtDNA C4a to that. C4a'b'c and C4a were prevalent in early kurgans around ~5000 BP. Around ~4500 C4b was busy expanding from Bristol Bay in Alaska with the Arctic Small Tool Tradition.

Grey said...

@Maju

"Does anyone have a half-decent idea of how could this happen?, how can this make sense?"

My theory is ANE were originally nomadic mammoth hunters who settled down in different regions across north Eurasia after the climate changed.

Assuming
1) HGs have unusually high relative population density in wetlands

and

2) places like the Black Sea were shrunk much smaller in the LGM and the difference in size was wetlands

then I think this initially fairly homogenous nomadic population diverged into large settled "clumps" of ANE around the Black Sea, Caspian Sea, Azov etc. (Baltic as well?)

(like the pottery using sedentary foragers west of the Black Sea.)

So basically lots of recently diverged ANE around the Black and Caspian Seas and any other areas where there were clumps. Then one of these clumps of ANE develop a horse culture and push the others west.

So you get ANE with different Y signatures being pushed west by Indo-European ANE with the R1a signature (or one of the R1a signatures).

(So the alternate explanation for the Scotland / Scandinavia ANE is they were one of the non IE ANE who were pushed west.)

So basically a prequel to the Goths etc getting pushed west in later millenia.

(domino effect)

Shaikorth said...

West Europeans may seem odd...unless ANE is not a specific IE marker and/or Lazaridis fits, which after all are based on only three ancient genomes of varying coverage, overestimate ANE at the expense of WHG as I previously proposed.

I'm not too attached to uniparental markers as indicators of autosomals either.

Regarding present-day Iberians, I wonder whether French + some North African is a better fit than a more even split of Basque + Tuscan. The former would explain higher ANE.

Davidski said...

Well, again, MA1 belonged to Y-hg R, which first appears in ancient DNA from Europe during the Copper Age (R1a at Eulau), at a time when there are major shifts in the mtDNA composition and material cultures in Central Europe.

I'd have to suspend my disbelief to accept that these are all coincidences and not directly linked to the appearance of Indo-Europeans deep in Europe.

Ryan said...

@Maju - Couldn't whoever introduced R1b into Atlantic Europe also have had some amount of ANE affinity?

Shaikorth said...

Here are some of the things that makes me wonder about those fits.

In extended figure 6, there is no discernable difference between Ukrainians and Norwegians in their Mal'ta-Stuttgart, Mal'ta-Loschbour and Loschbour-Stuttgart f4 comparisons.

However the Ukrainian fit is significantly more farmer-shifted than the Norwegian fit. Both get the exact same amount of ANE. Besides issues with software or genome quality I see no reason for that.

French get about the same amount of ANE as Icelandic. The two show no difference between their Loschbour-MA-1 relation, but French are considerably more "Stuttgart" than "MA-1" or "Loschbour" compared to Icelandic and the difference in fit seems to cut EEF its share entirely from WHG instead of ANE. The end result is what I think is an artificially elevated ANE reading for French.

About Time said...

@Maju, the Irish (a conquered people) have a folk saying along the lines of: daddy can be a soldier from anywhere. Maybe he stays, maybe he doesn't. Maybe Junior learnsa few words of Dad's language, maybe a lot, maybe nothing.

But from the cradle - baby grows up speaking his Mother tongue. Maybe Gumbutas had it backwards: IE was the "Mother Language."

Helgenes50 said...

@Shaikorth

It would be necessary, when you say french, to know the sense of French. I am French citizen, knowing I am Normand. But a guy from Marseille or Strasbourg is French like me, on the paper no difference, but in genetics, it is not the same case.

First of all, it would be necessary to break the different French populations
The difference is maybe not so great as in Italia but certainly more than for British.

Davidski said...

I don't think Maju can bear the thought of a late (post-Neolithic) entry of R1b into Western Europe with a horde of MA1's cousins rich in ANE. But R1b surely is associated with ANE if we go back far enough, so perhaps at least some subclades of R1b arrived deep in Europe with largely ANE populations during the Copper Age? That's certainly my bet for R1a, and all you gotta do is change one letter to get from R1a to R1b.

About Time said...

Sorry to double post. What I mean is, maybe we need to be thinking more of mtDNA (moms) as the factor that moved some of this ANE westward at times.

Foreign men (soldiers, tradesmen, etc) move in to a central location, marry local women. Stay for a while (maybe a few generations), then the natives rebel and they go back home, bringing their wives and children.

Effect on the "home" population (the soldiers' homeland) is that Y-DNA doesn't change too much, but new mtDNA hg enter the gene pool.

Lots of examples. Think of the Hyksos, or even the colonial outposts of European powers. People get tired of deteriorating life in the colonies and go home.

Back at home - it is their mother tongues that remain constant. Maybe some of their Y-DNA stays behind in the abandoned colonies.

About Time said...

@Grey: "I think dna at the end of a migration chain will always be less diverse."

"If you have a population A and part of it moves to region B and then part of region B moves to region C and part of region C move to region D then the people at D will be a subset of a subset of a subset of A's diversity."

"This is before you take into account bottlenecks caused by people adapting to new environments - especially if the environment was harsher than the one they moved from."

"If correct then Bedouin or Tuaraeg might be an example of the same thing (assuming adapting to a desert caused a bottleneck)."

Very good point. Something that also happens with languages. Contrast Balkanic languages (very diverse) with uniformity in Scandinavia or India. Both of those were sort of cul-de-sac end points, especially India. I


I think people who moved all the way to India never left for the most part. Changes to dense/humid culture and mountain barriers were too much to "turn back" from.

Those people just pushed deeper into India over time. The "paths of the pitris" were one way there. That's why very old customs stuck around there. Less so for Central Asia.

Shaikorth said...

Helgenes50, the French sample in Lazaridis et al seems to be from one place (Creuse department according to coordinates in supplementary data). The study has an additional Southwest French sample which is different to the one I was talking about.

There could well be differences between the individuals in the French sample anyway - their Belorussian and Finnish samples have some who cluster with Ukrainians, some who cluster with Mordovians and Russians and some who cluster with Estonians and Lithuanians in FINESTRUCTURE and we might see similar behaviour in their French sample if that part of the graph was also zoomed - but my point was mainly about the issues in their fitting system. You can replace "French" with "Croatians" and the essentials stay.

Helgenes50 said...


@Shaikorth

Thank you for these details. Creuse is in the center of France
it's important to know where the samples come

Maju said...

@Ryan: I'm awaiting that R1b pops up any day in Megalithic farmers (so far most sequences are women or have not been sequenced for Y-DNA successfully). Based on autosomal and mtDNA I see clear continuity in the Basque Country since Neolithic and Basques are 90% R1b and clearly neither IE nor ANE-related. So as such R1b, at least the largest "southern" clade S116, should not be related to either intrusion, really. In Iberia at least other types of R1b are negligible at the levels of R1a. That's also the case in Ireland and, essentially, France.

@About Time: what you say is most unlikely. There's very little mtDNA H in Central or Eastern Europe before at least Chalcolithic, it's origin in modern Europeans is almost certainly to the SW, where we do see much of it very early on (but no paleohistorical ANE). Anyhow, IEs were conquerors and 95% of Irish now speak English almost only. Of course they got the language with almost no genetic intrusion but they got it by conquest and their language is not their grandmothers' cradle language anyhow.

@David:

"I don't think Maju can bear the thought of a late (post-Neolithic) entry of R1b into Western Europe with a horde of MA1's cousins rich in ANE".

It's not that I can or cannot bear: it's just that it makes NO SENSE. R1b-S116 did not arrive from Eastern Europe, not even Central Europe and shows no correspondence with ANE frequencies (French and Spanish are more ANE than Basques but have a lot less R1b-S116). Maybe R1b-U106 has some relation with ANE via Doggerland or whatever but S116 certainly not.

So if anything there is a negative correlation of R1b(S116) with ANE. A very clear one and you should be able to see that and not argue for something that just doesn't make any sense at all.

Shaikorth said...

I don't think the Irish would turn out less ANE in the Lazaridis 3-way fits than Ukrainians or Poles. That either makes R1b-S116 irrelevant when it comes to gauging ANE admixture or associated with ANE admixture and perhaps Indo-Europeans too - or not, if the fits exaggarate ANE at the expense of WHG as I suspect.

Locrian said...

David — I am prepared to bet ($10.00? :) ) that even though R1a speakers were IE’s, that IE languages spread BEFORE any significant spread of R1a into Western Europe. So the early bearers of IE languages were not R1a. (I don’t include in this a spread of IE by R1a into Northern India — that case seems plausible.) As Maju notes Britain and Spain (and France) do not have an excess of R1a even now and yet have been IE speakers for around 2500 years. My guess is that IE was spread by a diverse group of Y-haplogroups and that while Europe was still mostly G2 IE languages were in place.

Davidski said...

Locrian,

You're throwing away your money, because you obviously haven't bothered to look at this issue in any great detail.

Here are some points you might find useful:

a) European Neolithic farmers in all likelihood mostly belonged to Y-chromosome haplogroups G, from the Near East, and the indigenous European I2, and did not speak Indo-European languages, because modern European languages carry non-Indo-European substrata associated with farming.

b) You don't need an excess of a Y-chromosome haplogroup to speak the language of the people who spread this haplogroup, because languages can be learned. In any case, R1a is found in Western Europe, including subclades which are specific to the region.

c) Y-DNA R appears to be associated with ANE, because MA1 belonged to R*, and R1a and R1b only appear in ancient Central and Western European DNA from the Copper Age onwards, the same timeframe that ANE seems to have moved into these regions, which is also the early Indo-European timeframe.

d) Even if R1b wasn't strongly associated with ANE in Western and Central Europe during the Copper Age, there is evidence from tooth morphological data from ancient skeletons that the newcomers into Central Europe at the time, like Corded Ware and Unetice groups from the east and Bell Beakers from the west, mixed by swapping women (ie. they practiced female exogamy with each other), which seems to be backed up by the mtDNA found at the Bell Beaker Kromsdorf site, because the haplogroups found there included U2 and U4, which are more typical of Eastern Europe.

Now let's sit back and wait for more ancient genomes to fill in the gaps, because the basic story is already in place.

barakobama said...

I don't have time to really get into this, but Maju has an interesting hypothesis. West European R1b L11 may have nothing to do with Indo Europeans(Basque are great evidence), and instead was brought by another eastern people. There is alot though that that hypothesis can't explain, like the genetic makeup(much higher amounts of ANE and WHG than Neolithic farmers) and close relationship of north-west Europeans and a Germanic-specific subclade.

I have noticed western Europe in admixtures looks mainly like a mixture of Irish and Sardinians. For example, Spanish fit almost perfectly in Eurogenes K13 has half Irish and half Sardinian.

Once genomes are finally sampled from the Bell beaker and Unetice cultures in central Europe i expect Y DNA R1b L11 and for them to be like Irish minus some Sardinian like ancestry. I have not done much research on this, so i could be totally wrong and I'm open for alot of other possibilities.

Daniel Szelkey said...

Perhaps Indo-European languages are extremely old dated to at least 20k and originated somewhere between Turkey and Lake Baikal. Populations with R1b originally spoke IE (CENTUM?), while those with R1a spoke another kind of IE (SATEM?).
Mal'ta did not speak an indo-European language most likely, although this is possible.
1. Ancient dna suggests that Indo-European languages are extremely old.
2. Why wouldnt they be? no haplogroup exists in both India and Ireland in excess of 5% that have a common ancestor within the past 10,000 years although many exist that have a connection within the time of mal'ta
3. Ancient dna does not enlighten people on where IE languages began, however it does prove that it is possible that IE languages began in siberia, and unlikely that they began in Scandinavia, or West Europe the Iberian Refugium hypothesis of IE origins is dead. It seems to me that the places where IE languages could have began include Anatolia, Iran, Central asia, Siberia, or East Europe. A huge area.
(scenarios not likely but not inconsitant with any data)
4. IE languages of West Europe have an origin in anatolia, and have paleolithic roots in Central Eurasia ) (tend to be centum, with exceptions).
5. IE languages of East Europe and South Asia have an origin in Siberia, and have paleolithic roots in Central Eurasia from the same population (tend to be satem with exceptions) .
6. The basques possibly speak a neolithic european language, they have plenty of I2 and U5 compared to others in iberia, and they just happen to be one popualtion that retained their language, they are mostly IE on the male side, and neolithic non IE farmer on the female side, like most others in Europe. However, the basques speak the langauge of their mothers not their fathers.
7. Finno-Ugrian might be the paleolithic European language, with Siberian (non-malta) elements , or the paleolithic European languages may have went extinct.
8. All of European may be mostly neolithic with paleolithic elements however, most of the males come from IE.

Daniel Szelkey said...

WHG and ANE, are and maybe cache all categories, WHG was not a pure population and ANE was probably not a pure population, although they were less diverse then most modern populations. WHG probably descends from a U5-Population (I2?) U4 population (J1-dagestan variant?-) and U2 population (C6?). U is very old, and phylogenetic analysis suggests that it may even be a sister clade to R-mtdna (haplogroup N-UR-U/R)_. U5 is the most divergent branch of U making it farther from U4 then H is from T (MTDNA). To me it appears that 90%+ of the WHG came from 3+ highly divergent elements.

Davidski said...

Indo-European languages aren't older than the final Neolithic. See here...

http://www.youtube.com/watch?v=d864bwyCAoA&list=PLAXoDomeFLX90fTHi0W8lYBtEoZHSBH2i&index=10

Daniel Szelkey said...

You mean Indo-European languages probably aren't older than the final Neolithic. You can not be certain. It is possible but not likely that IE is paleolithic.

Maju said...

We would not be able to discern such a great similitude as we do among IE languages if they'd be so old. We would probably see very limited similitude instead, as may happen with Amerind languages. Notice that language divergence tends to accumulate exponentially rather than linearly and also that IE, having expanded so much, has necessarily changed quite faster than the base rate of language evolution.

Or just accept Starostin's authoritative criterion: we cannot discern with any certainty linguistic relatedness older than ~10 Ka. Afroasiatic for example is in that max. limit and therefore somewhat disputed.

Also there's no real need to relate languages and genes. Neither Irish nor Jamaicans have significant English ancestry, for example. Caribbean Spanish ancestry seems to be essentially Guanche (Canarian Aboriginal, Berber), not really European. Etc. Languages can expand with almost no genetic influence: elite domination is enough (and elites are always a tiny few).

barakobama said...

"Dene groups show high levels of mDNA C4b. The most common mDNA groups in early Kurgans were C4a'b'c and C4a."

No they had typical modern European mtDNA(mixture of farmer and hunter), with a notably high amount of U5a1 and U4(eastern hunter gatherers). You need more evidence than a few ancient Kurgen samples to say those lineages are west Eurasian.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CCgQFjAA&url=http%3A%2F%2Fwww.theapricity.com%2Fforum%2Fshowthread.php%3F118795-List-of-65-ancient-Pontaic-steppe-individuals-and-DNA-results&ei=KXdkU5bnI8mqyATx5oCwAQ&usg=AFQjCNEWI44y5U5UMQYW2kCdi5SaV2Tuhg&bvm=bv.65788261,d.aWw

barakobama said...

'Aren't they missing out on a lot of data by not performing any tests on existing middle eastern populations and remains? I feel like they're guessing around with the whole cryptic "basal eurasian" thing when the answer might just be a few more sequenced genomes away. They've been concentrating on bones from central/western europe though."

Your right, I am sick of guessing how much middle eastern ancestry Stuttgart and modern west Eurasians have. I can't wait for an early near eastern farmer's genome to be published.

barakobama said...

"And if Gok2 is overwhelmingly WHG (77% hunter gatherer) than why is he nowhere near the other hunter gatherers. The K4 in the Skoglund 2012 paper about Gok4 suggested nothing like 77% hunter gatherer."

Early European farmers seem to be majority European hunter gatherer because there were close relatives to European hunter gatherers in west Asia who admixed with basal Eurasians, who then brought farming to Europe and mixed with European hunter gatherers.

barakobama said...

In Table S13 there is alot of interesting information. Sardinians are consistent with having more hunter ancestry and less basal Eurasian ancestry than Otzi. Because Anzick1, MA1, La Brana-1, Saqqaq, and Australian are closer to Sardinians than to Otzi.

Sf11 though is much closer to Otzi/Iceman than to Sardinians, Gok2 is also closer to Otzi than to Sardinians. This suggests Otzi and Gok2's hunter gatherer ancestors were more related to Sf11 than to La Brana-1 and Sardinian's hunter gatherer ancestors. This is consistent with early central European farmer Stuttgart being more related to central European hunter gatherer Loschbour and Swedish hunter gatherer Motala12 than to Spanish hunter gatherer La Brana-1. This is also evidence that Scandinavian hunter gatherers were closely related to central European and Balkan(?) hunter gatherers.

Ajv58 has a much closer relation to Gok2 than any of the other European hunter gatherers do, suggesting that Gok2 had recent PWC-related ancestry. Otzi and Gok2 are closer to Ajv58 than to La Brana-1, while Sardinians are closer to La Brana-1 than to Ajv58. The reason for this is Sardinians hunter gatherer ancestors came mainly from western Europe, while Gok4's hunter gatherer ancestors came mainly from central Europe and Scandinavia, and Otzi's hunter gatherer ancestors came mainly from central Europe.

Gok2 probably had some ANE ancestry because he had Scandinavian hunter gatherer ancestry. Otzi may have also had some ANE ancestry since Anzick is closer to him than to Sardinians.

Notably Ajv58 and Sf11 don't act the same way in D-statistics. Ajv58 is closer to Sardinians than to Otzi, while Sf11 is much closer to Otzi than to Sardinians. So, Scandinavian hunter gatherers were not a uniform population.

Sardinians and Otzi are almost equally closer to Ajv58 than to each other. I don't know how to explain this, except maybe Sardinian's farmer ancestors were more related to Gok2's farmer ancestors than Otzi's farmer ancestors.



barakobama said...

"I've added the info about SLC45A2 to the text.

By the way, the dark skin pigmentation of the hunter-gatherers was definitely a surprise, but perhaps not a mystery. European hunter-gatherers probably ate a lot of vitamin D-rich fish, meat and mushrooms, so they didn't need to rely on the sun as much as the farmers for this vitamin, right?"

Davidski, i think your making a bad assumption saying Mesolithic Europeans had dark skin. You can't expect known light skin mutations to be the only causes of European light skin. I found a northern European at GEDmatch who had rs1426654 A/G, rs16891982 C/G, rs28777 A/C, rs1042602 C/C and he is a typical light skinned European.

I also found a northern European and Arabian who had the same alleles(including rs16891982 C/C) in SNPs associated with skin color, but the Arabian individual is brown skinned and the European individual is a typical light skinned European, who claims to be able to tan nicely.

There are unknown causes to European light skin. The unknown causes could have originated in EEF, WHG, or mixed descendants. Today they are widespread in Europe, so probably very ancient or were selected for all over Europe and became dominate very quickly.

Modern day middle easterns obviously are not dominated by these unknown light skin mutations, so the unknown light skin mutations probably descend from European-specific ancestors. It is reasonable to say Mesolithic European hunter gatherers are the source of the unknown light skin mutations, partly because northern Europeans have lighter skin than very Neolithic-like south-west Europeans. It is also reasonable to say early European farmers are the source, because they are the main source of known light skin mutations in Europe which were selected for in European's ancestors during and maybe after the Neolithic.


barakobama said...

"WHG and ANE, are and maybe cache all categories, WHG was not a pure population and ANE was probably not a pure population, although they were less diverse then most modern populations. WHG probably descends from a U5-Population (I2?) U4 population (J1-dagestan variant?-) and U2 population (C6?). U is very old, and phylogenetic analysis suggests that it may even be a sister clade to R-mtdna (haplogroup N-UR-U/R)_. U5 is the most divergent branch of U making it farther from U4 then H is from T (MTDNA). To me it appears that 90%+ of the WHG came from 3+ highly divergent elements."

What???!!!!!

mtDNA U is descended of mtDNA R, and there are no known intermediates between R and U. U5 is a subclade of U, like U1, U6, and U2'3'4'7'8'9. H is more distant from T, than U5 is from anyother U subclade because H and T's common ancestor is R, while U5 and other U subclades common maternal ancestor is U(subclade of older R). I guess it is possible that U broke off very early from R, while R2'JT and R0 are much younger and therefore may have a more recent common maternal ancestor. If that was true though you should expect R2'JT and R0 to share mutations that U does not share, which is not the case.

Of course WHG and ANE were not pure, but it is very difficult to discover how they were mixed without very ancient Upper Palaeolithic genomes. Your making big assumptions about the origins of WHG's main maternal lineages: U5, U4, and U2. U* and U8 have also been found in Upper Palaeolithic Europe, so stone age European hunter gatherers were fixated for mtDNA U and had an incredible variety of U lineages. This is evidence all their U lineages came from an ancestral population fixated for U, that was around when U was very young.

My main point is WHG's mtDNA is evidence of low diversity, less than practically all modern non-Africans, including isolate populations discovered by westerners in historical times like Onge, Americans, and Papuans.

WHG was also nearly fixated for Y DNA I, which is evidence they were also very undiverse paternally. People make too big of a deal about La Brana-1 having C1a2-V20, that lineage was certainly a small minority in Mesolithic Europe. C1a2-V20 is prove WHG had some diversity in paternal lineages, but the diversity is so ancient it's not very significant and hard to identify in autosomal DNA. F-96 is a subclade of macro-non African haplogroup F, and is exclusive to Europe so probably also descended of WHG.

In my opinion most of WHG's ancestors arrived in Europe over 30,000 years ago, and there is already over 30,000 year old mtDNA evidence the Gravettian(MA1 aka ANE was apart of a related culture) people were the ancestors of WHG, or at least partial ancestors of WHG.

barakobama said...

"The 6 SNP's of MC1R tested by Loschbour and Motola12 were the same as by Stuttgart and none suggested the lightening mutations. IRF4 / rs12203592 is associated with eye color and freckling. But if an individual is missing the 2 skin lightening alleles (rs1426654 A and rs16891982 G) i am not sure how that freckling would look like. Eye color was blue by the hunter gatherers but hair color was also dark."

Actulley Loschbour and Stuttgart had derived C/C alleles in SNP rs1110400(in gene MC1R) which is associated with red hair, and is 0% in CEU. All of the MC1R mutations they were missing are very very very rare in modern Europeans, and i don't see how their hair color could be accurately predicted with the SNPs tested.

C/C and C/G alleles in SNP rs16891982 is said to give Europeans a 7x better chance of having black hair, but most C/G an C/C(only one) Europeans I found at GEDmatch had a brown color(like most Europeans), some even medium brown. Blonde and brown hair are predicted most unaccuratley using SNPs, because Europeans oftenly have a mix of the two. Motala12, Loschbour, La Brana-1, and Stuttgart almost defintley all had dark(probably dark brown or black) hair, but we can't say it's impossible any of them had blonde hair, and for Loschbour and Stuttgart red hair.

barakobama said...

Also Anime, La Brana-1 and Loschbour did have the h1-haplotype which is found in 97% of modern blue eyed people but Motala12 did not. I don't think Motala12 had blue eyes, but instead a mixed color, maybe grey i don't know.

Grey said...

@barak
good stuff

Grey said...

@barak

http://www.dailymail.co.uk/femail/article-1103391/Mixed-race-couple-birth-black-white-twins--second-time.html

Mixed race couple, woman with red hair, two sets of twins, in each case one black, one white with red hair i.e. implies father carrying MC1R also and the kid that gets it from both ma and pa are white with red hair.

About Time said...

@Grey, weird. In the ancient world, there were some myths about this kind of thing. The Irish one is the story if Eber Finn and Eber Donn.

Finn means "white" and Donn means "dark." Sometimes I wonder if it's a story about the same person at different times in their life, but maybe it was a twin story. People don't change: this kind of thing would surely be noticed in the ancient world, especially if the "white and black" twins were royals/elites or whatever at the time.

Today it's just a tabloid headline. How Joycean.

But to be fair--that dad in the Daily Mail looks pretty mixed. If I had to guess at a glance I'd say he was Egyptian.

Also diff topic: people debate whether ANE was part Asian (aside from drift). But what if we reverse the presumption and ANE and EEF were the "unmixed" pops. What would that make WHG (aside from drift)--part African (like paleo-Berber) maybe?

Grey said...

@About Time

"But to be fair--that dad in the Daily Mail looks pretty mixed."

Yes, that's the point. He'd have to be carrying MC1R himself as a recessive.

#

"The Irish one is the story if Eber Finn and Eber Donn."

Interesting and it seems to me that would follow. If IRF4 and / or MC1R had this effect there could be dark, halfway and light people in the same family.

#

"What would that make WHG (aside from drift)--part African (like paleo-Berber) maybe?"

Yes I wonder about that connection.