search this blog

Friday, March 18, 2016

New insights into human adaptation and population structure thanks to ancient genomes


Open access at Nature Communications:

Abstract: The influence of positive selection sweeps in human evolution is increasingly debated, although our ability to detect them is hampered by inherent uncertainties in the timing of past events. Ancient genomes provide snapshots of allele frequencies in the past and can help address this question. We combine modern and ancient genomic data in a simple statistic (DAnc) to time allele frequency changes, and investigate the role of drift and adaptation in population differentiation. Only 30% of the most strongly differentiated alleles between Africans and Eurasians changed in frequency during the colonization of Eurasia, but in Europe these alleles are enriched in genic and putatively functional alleles to an extent only compatible with local adaptation. Adaptive alleles—especially those associated with pigmentation—are mostly of hunter-gatherer origin, although lactose persistence arose in a haplotype present in farmers. These results provide evidence for a role of local adaptation in human population differentiation.


Key, F. M. et al. Human adaptation and population differentiation in the light of ancient genomes. Nat. Commun. 7:10775 doi: 10.1038/ncomms10775 (2016).

See also...

Lactase persistence and ancient DNA

54 comments:

rozenfag said...

Received
16 June 2015
Accepted
18 January 2016
Published
18 March 2016

Why it was kept three months after being accepted? They waited for the other paper(Extracting Neandertal and Denisovan DNA from the genomes of present-day Melanesians )?

Matt said...

I kind of had the following thoughts on this:

IIUC, the things this paper does is estimate that:

- At the most differentiated (from African) alleles Eurasians are enriched in genic SNPs .

(Makes sense; it's relatively difficult for an allele to drift its way to very high degrees of differentiation)

- Of these, modern day Europeans are relatively more enriched at genic SNPs compared to non-genic SNPs, compared to East Asians, particularly due to weaker effects of drift in modern day Europeans

(Makes sense; but note doesn't necessarily mean Europeans have more absolute numbers of highly differentiated genic SNPs from Africans, and in fact they probably have fewer based on previous work.

See also the figure S24 which shows that on a PCA based on genic SNPs, EAs as far away from Africans than Europeans are, so about as differentiated.)

- (I think) They claim Europeans being particularly more enriched at genic SNPs compared to non-genic SNPs compared to East Asians is also likely to be due to more differing selection forces in Europe (I'm not so sure about this).

- "In the genic European tail, Stuttgart has 10.4% fewer of the alleles that are highly differentiated from Africans in present day Europeans, compared to Loschbour"

- "Hunter gatherers contributed disproportionally to the highly differentiated alleles within genes in Europe but, intriguingly, not outside of genes"

What they don't do, really (although the probably can't), is estimate whether:

- WHG has an enrichment of genic SNPs relative to non-genic SNPs, compared to EEF

That's pretty important. It's particularly important due to the fact that WHG seems to have had really low population sizes compared to EEF, and to have had a much higher effect of drift (see - http://www.nature.com/ncomms/2015/151116/ncomms9912/images_article/ncomms9912-f3.jpg for instance), then lower contribution to Europeans.
Possible WHG had more highly differentiated SNPs both among genic and non-genic SNPs, and the genic SNPs "introgressed" out of proportion to the WHG population contribution.

I think they probably couldn't test this though. They did try to simulate out whether differences in NE (effective population size) could have contributed to the differences in contribution to Europeans highly derived SNPs, but I'm not too convinced by this indirect evidence.

Matt said...

One thing I do question about the finding is:

"No ancestral genome carries rs16891982’s derived allele, but Loschbour carries the haplotype that, in present-day populations, is linked to the derived allele"

So does this contradict previous findings by Mathieson and Gamba that WHG did not have the rs16891982 variant on SLC45A2 (e.g. http://www.nature.com/ncomms/2014/141021/ncomms6257/images/ncomms6257-f3.jpg) or is more like WHG likely did not have the variant, but some related HG did?

And one thing particularly they do seem to be quite wrong about was "No ancestral genome carries rs16891982’s derived allele" yet Mathieson already showed that a good 40% of the Anatolian_Neolithic already did http://biorxiv.org/content/biorxiv/early/2015/10/10/016477.full.pdf (and lower frequencies in the Iberia EN with higher WHG ancestry), by directly sampling the SNP. The AN also had some degree of the derived OCA2 variant.

If Reich lab have sampled AN, who may or may not have actually been admixed with WHG, and they actually have the SNPs, it seems a bit pointless to go "Oh, well EEF (based on Stuttgart) didn't have these haplotype", when the AN have the SNP.

At the same time, "But the European tail includes a large number of alleles in the lactase enhancer region and the LP haplotype (chr2:135859371-136740900) that are exclusively present in Stuttgart (65% of Stuttgart specific targets; Fig. 6c). Thus the haplotype that is today associated with LP in Europe originated most likely in this genetic background, which we detect only in the Stuttgart farmer, although this individual itself did not carry the LP allele. "

Yet it seems like the European LP variant is not necessarily likely to have come from an EEF like population (rather a Yamnaya / CHG like one).

This paper was kind of written on 16 June 2015, but then the first version of the Mathieson paper was published online 14 Mar 2015, and they actually reference it, already showing SLC45A2 in the EN, so kind of don't quite get that...

Romulus said...

"No ancestral genome carries rs16891982’s derived allele"

What about Stora Forvar? Reich lab created this myth of dark skinned hunter gatherers which has turned out to be totally false. Looks like 19th century anthropologists are proven to be SPOT ON once again thanks to ancient DNA.

The Stora Förvar 11 remains were found on Stora Karlsö, a small island off the west coast of the larger Swedish island of Gotland, in the Baltic Sea. The remains were dated to 7,500 to 7,250 years ago, and they were found in a late Mesolithic context.

Stora Förvar 11 belonged to mitochondrial haplogroup U5a1. He had two copies of the depigmentation allele in the gene SLC45A2, and since Motala 12 had at least one copy of the depigmentation allele in the gene SLC24A5, we now know that both of the two major Caucasoid depigmentation mutations were present in the hunter-gatherers of Mesolithic Europe.

capra internetensis said...

"Ancestral genomes" in this study were only Loschbour, Ust' Ishim, and Stuttgart. They weren't talking about all ancient genomes.

Krefter said...

@Matt,

The frequencies of SLC24A5 and SLC45A2 in pre-3000 BC times doesn't matter. They tell nothing about natural selection. They had been popular for 1,000s and 1,000s of years before the Neolithic(except in WHG). Their high frequency today have nothing to do with farming or Ice age climate or whatever.

The key is the Late Neolithic/Bronze age Europe. SLC45A2 more than tripled in frequency between 3000 and 2500 BC. That's pretty significant.

Davidski said...

Matt,

The current thinking seems to be that the LP allele originated in early European farmers, but shot up in frequency during the Bronze Age in populations of mixed farmer/steppe origin.

That's because none of the pure steppe samples carry the LP allele. It first shows up in a Bell Beaker sample, and then on the steppe in Srubnaya samples.

So it looks like the steppe people moved into Central Europe, picked up the allele, and then spread it around Europe and parts of Asia, as far as India.

Rob said...

But it was already spread
All we're seeing is an increase in frequency
It's is premature to attribute the spread of LP to steppe groups in Europe, when it already existed there

Karl_K said...

@"Rob"

"But it was already spread
All we're seeing is an increase in frequency
It's is premature to attribute the spread of LP to steppe groups in Europe, when it already existed there"

But the frequency went up after the steppe people moved in, as the proportion of farmer ancestry went down. The selection was due to the old allele being associated with a new culture. In Europe, the people with the least steppe ancestry tend to have the least LP, and in India, the people with the most steppe ancestry have the most LP.

Karl_K said...

@Matt

"or is more like WHG likely did not have the variant, but some related HG did?"

Exactly. The mutation arose on this haplotype background. That means it most likely was a related HG.

So the rs16891982 variant on SLC45A arose in WHG probably.

Rob said...

Ok thanks "Karl"

Karl_K said...

@Krefter

"The frequencies of SLC24A5 and SLC45A2 in pre-3000 BC times doesn't matter."

Dude. You gotta be a little more careful with your words (and not just the grammer and typos).

I could just as easily say that none of this matters.

It is of great interest to many people to know in which population a genetic variant actually came to exist. This is part of it's history. This is why there is interest in Neanderthal and Denisovan introgression.

FrankN said...

The statistical baseline, i.e. their simulated "no specific selection" case against which they measure the significance of actual findings, is problematic. The scenario places UI as basal to West Eurasian HGs and EEF farmers alike, highly questionable when looking at UI's yDNA K, against WHG I and EEF G2. Alberto has made some good points on this.
They also model current Europeans as 50% HG, 50% EEF. Whatever their understanding of HG (i.e. whether just WHG, or also EHG/CHG), that ratio is obviously off the mark. A more realistic demographic modelling could have affected the statistical significance of their findings.

Pretty disappointing is they only focusing on the usual stuff (skin/ eye colour, lactase persistence). I generally would like to see less discussion of skin colour, which is obviously prone to racist (mis-)interpretation, but of little help in understanding prehistory. Much more interesting would be information related to metabolism (->nutrition habits) and immunology. Recent research has identified positive selection, with geographically distinct patterns (Africa vs. Europe vs. E.Asia) on immune-system related genes such as Toll-like receptors (TLR- they identify bacteria/virus-specific DNA/RNA, and activate corresponding immune response; some, e.g. TLR7, are coded on the X Chromosome). That research isn't referenced at all.

I did a random scan of some of the genes for which they established genic enhancement across all 4 European reference groups (not that I have any expertise in genetics, I just googled what comes up as recent research), and found:

- CELF6: Related to brain (mis-)function, including autism and abnormal serotonine levels
http://www.ncbi.nlm.nih.gov/pubmed/23407934

- SLC4A4: Regulates sodium bicarbonate transport and thereby transmission of neural signals. Autosomal disorder is linked to brain retardation and ocular abnormalities
http://www.uniprot.org/uniprot/Q9Y6R1

-APBA2 Involved in neuronal signal transduction. Malfunction is linked to autism and Alzheimer.

-SLC10Manganese and zinc resorption. Malfunction is linked to Parkinson and Alzheimer. The genic European tail includes a number of other genes related to liver/kidney functions, e.g. AQP6 (Aquaporin) and SLC12A1 (Sodium-Potassium-Chloride resorption). Possibly nutrition-related, deserves further analysis.

- EP400: Target of Adenoviruses, which may have mild (cold-like) to deadly effects on humans, and are otherwise especially prominent with canines ("canine hepatitis"), and bovines including equids. Clearly a gene worthwhile further analysis in relation to the domestication history of dogs, horses and cattle; a Loeschbaur v.s Stuttgart (vs. Yamnaya) comparison would have been quite informative.
http://www.uniprot.org/uniprot/Q96L91
https://en.wikipedia.org/wiki/Adenoviridae

- IFI16: Interferon regulation. Key role in antiviral response, especially to Herpesviriridae, Lentivirus (HIV), and papillomavirus (causes warts, which may evolve into cancer). Also linked to inflammatory reaction and autoimmunity.
http://www.ncbi.nlm.nih.gov/gene/3428
http://www.ncbi.nlm.nih.gov/pubmed/25742143/

- AVEN: Cell-death activator, poorly understood, but a/o linked to Leukemia, possibly cancer defence in general, and also infertility.
http://www.ncbi.nlm.nih.gov/gene/57099
-MLLT3: Mixed-line Leukemia, related to AVEN above
-TPRG1: Tumor proteine p63 Regulated
-WWOX: Tumor-suppression gene.
http://www.ncbi.nlm.nih.gov/pubmed/18437686

Speculatively, pigmentation change could have resulted in a need to stimulate cancer defence, especially as concerns Melanoma.

Karl_K said...

No problemo "Rob"

But for future reference, I prefer "Karl_K" , to further distinguish myself. Someday, someone named just "Karl" will come in and say a bunch of stupid stuff. I remember that this already happened to M#rnie_K.

FrankN said...

I don't understand this: There are two teams from the same institution - Leipzig MPI for Evolutionary Anthropology - yet one team doesn't take notice of the other teams' findings.

This recent publication, also from Leipzig, is exactly providing what I would be looking for:
http://www.sciencedirect.com/science/article/pii/S0002929715004863

"Here we document a cluster of three Toll-like receptors (TLR6-TLR1-TLR10) in modern humans that carries three distinct archaic haplotypes, indicating repeated introgression from archaic humans. Two of these haplotypes are most similar to the Neandertal genome, and the third haplotype is most similar to the Denisovan genome. The Toll-like receptors are key components of innate immunity and provide an important first line of immune defense against bacteria, fungi, and parasites. (..)

Of the three putatively introgressed core haplotypes, III and IV are most similar to the Altai Neandertal genome (..) Core haplotype VII is most similar to the Denisovan sequence. (..)
Core haplotype III is present in all non-African populations, and we also observe it in two chromosomes from two Northwest Gambian individuals (excluding the potentially admixed African Americans). (..) Core haplotype IV is restricted to specific Asian groups, and core haplotype VII is present in only two South Asian individuals [STU, ITU]. (..)
We note that two early modern humans (dated to ∼7,000–8,000 years before present) from Europe (Stuttgart and Loschbour) also carry TLR haplotypes similar to Neandertal core haplotype III and an early modern human from Asia (Ust’-Ishim, ∼45,000 years) carries a haplotype that is most similar to core haplotype V—the major non-introgressed haplotype in modern humans. (..)

The frequency of the shared SNPs common to Neandertal-like core haplotypes III and IV also varies significantly between populations even within continents. In Europe, a North-South gradient is apparent with significantly high population differentiation between southern Europeans (Toscani and Iberians, with frequencies of 39.3% and 38.3%, respectively) and all other European groups (Finnish, British, and CEPH, with frequencies between 14.8% and 26.4%) (p value < 0.05) (..) In Asia, the most Eastern populations (Japanese and Han Chinese, frequency 53.4% and 53.6%) show high differentiation from other Asian populations (frequency 21.7%–41.9%; p value < 0.05; Tables S5D and S5E). In addition, the 12 SNPs defining Asian-specific core haplotype IV show high population differentiation when Dai and Vietnamese (with frequencies 9.4% and 9.9%) are compared to other Asian populations (frequencies between 0% and 4.7%).
(..)
These archaic alleles lead to significantly increased expression of TLR6, TLR1, and TLR10 in white blood cells, and in present-day people are associated with reduced Helicobacter pylori seroprevalence and increased susceptibility to allergies."


To understand the significance: The TLR1/6/10 cluster has been shown to display strong adaptive selection to the Plague. TLR10 plays an important role in mobilising response to influenza, probably also other viral infections, and modulates inflammatory response. TLR6 acts on Mycoplasma, bacteria immune to most antibiotics, which are involved in respiratory, sexual and rheumatic diseases.
A more detailed assessment appears to be included in the following paper - paywalled, but maybe someone with access can report key results:
http://www.sciencedirect.com/science/article/pii/S0002929715004851

Karl_K said...

@FrankN

"I generally would like to see less discussion of skin colour, which is obviously prone to racist (mis-)interpretation, but of little help in understanding prehistory."

Yet these genes have the highest rates of selection across all modern human genomes. So to imply that they are meaningless is incorrect.

Probably a lot of people in the Bronze Age were selected for racism, as they started to accumulate very compact, inheritable wealth. Metals and livestock changed the game, which changed the genetics.

FrankN said...

Karl_K: I understand your point. An analysis of the relation between skin colour and social status (grave gifts), and changes over prehistoric periods, would surely be interesting. Unfortunately, I am not aware of any aDNA study that ever interpreted pigmentation results accordingly.

Rob said...

Frank N
Can you check your email please ? I need ur ref advice

batman said...

Or the LP-alleles spread from its place of origin, where its major population still resides, via the waterways to the Caspian and the Caucasian steppes - where the wide, open fields were still unpopulated.

Bringing cattle along the LP-adapted communities could spread and multiply in to areas otherwise where foragers and herders had å hard time producing enough staple-food to survive the the winter. Parallel to Europe, where the large plains remains uninhabitted before the bovine LP-adapted populations (R1) start multiplying - inbetween the older societies of herders, hunters and fishermen (GHIJ/NO).

Remarkably, the lp-adaption seem to be based within an arctically adapted population, with high and narrow noses and a de-pigmented exterior.

It's dim up north - and short Seasons to produce and collect the Harvest needed to survive the long, dark and freezing-cold winters. Around the fireplace...

batman said...

https://royalsociety.org/news/2011/Dim-up-North/

Matt said...

Karl_K Exactly. The mutation arose on this haplotype background. That means it most likely was a related HG.

So the rs16891982 variant on SLC45A arose in WHG probably.


So I guess under this connection most likely on a WHG related population who contributed to Anatolia_Neolithic and Scandinavian_HG (who both have the variant) and not to Loschbour-Bichon-Hungary_HG-Iberia_HG, who don't seem to.

Although, because I don't know much about this area, how do we actually know that the rs16891982 variant arose on that haplotype background? Statistical association in modern Europeans (who have around 1.00 at this variant, seems like it would make it difficult to make a statistical association), or some other means?

Karl_K said...

@Matt

"how do we actually know that the rs16891982 variant arose on that haplotype background?"

Because chromosome crossovers are very uncommon on a small scale level.

So, for quite a few, the SNPs to each side will not be mixed up randomly by mating. This is classical linkage, but at a very fine scale.

So it everything stays the same on both sides for hundreds of SNPs, then the mutation in the middle must have arisen in that context.

After tens of thousands of years, this will no longer be the case, because of all the random crossovers.

UNLESS! The new allele had a selective sweep that brought all the nearby SNPs with it. Then it will take much longer to erase its history.

Tobus said...

@Karl_K: Where are you getting this Stora Förvar 11 information?

The stuff he put in italics is a quote from Genetikers blog.

ZI Alt said...

@Karl_K

Skoglund et al. 2014 - Genomic Diversity and Admixture Differs for Stone-Age Scandinavian Foragers and Farmers

Stora Förvar 11, in Table S6 is reported as derived at rs16891982.

Matt said...

@ Karl: Ah, my thinking was that the haplotype to variant association could've been a wrong one, if the haplotype was at high frequency, then the variant tended to cross over onto it, due to strong selection on that variant. But that shouldn't happen.

Btw, slight adjustment to my understanding earlier:

- they find that the absolute % enrichment in highly differentiated derived variants at genic sites is only slightly higher in Europeans than Asians, and in fact the number of derived allele variants with fixed differences is higher in the CHB than GBR - 3845 vs 1330.

- where the Europeans are particularly enriched in derived allele variants at genic sites is in the European tail of DAnc (Africa, Europe, Ust′-Ishim).

DAnc(Africa, Europe, Ust′-Ishim) is a statistic which provides a statistical score that measures whether a derived variant is highly frequent in Europeans and low frequency in both Ust'-Ishim and Africans (and variants scoring this way are called as in the "European tail" of the distribution) or, whether a derived variant is highly frequent in Africans and low frequency in both Europeans and Africans (these are in the "African tail").

While even the DAnc(Africa, Europe, Ust′-Ishim) still gives a much lower number of highly differentiated genic variants than DAnc(Africa, Asia, Ust′-Ishim) - e.g. CHB 809 vs GBR 354 - the enrichment in genic variants compared to non-genic is relatively higher in Europeans.

So now I understand it, this is stronger than I thought, as evidence that selection has been relatively more important in change in genic variants in Europeans, relative to East Asians, since Ust Ishim.

I still don't quite understand though:

a) why they didn't seem to run DAnc(Africa, Stuttgart, Ust′-Ishim) and DAnc(Africa, Loschbour, Ust′-Ishim). is this because they need multiple genomes in P2 for it to make sense and be comparable? instead there's kind of a back-checking for the presence of DAnc(Africa, GBR, Ust′-Ishim) in the ancient samples.

b) not sure how they accounted for Basal Eurasian properly in Stuttgart, which itself would seem to lower the enrichment of genic:non-genic differences vs Ust Ishim. they have simulations, hard to be sure whether their simulations have assumptions which make sense. (Like, clearly, the WHG related ancestors of EEF and WHG did not split 60,000 years ago and then never combine after that, and they both clearly didn't have continuous population growth following the same trajectory for 57,000 of those years, given Pinhasi's talk of very low population sizes as low as 30 in the Ice Age).

Grey said...

Interesting that both cases may be of genes developing in one population but expanding more after crossing into a second.

The LP allele developing among ppl who made a lot of cheese but didn't drink much raw milk then expanding dramatically after spreading to a pop who for some reason needed to drink a lot of milk is easy enough to imagine.

If the skin color genes were initially northern HG but spread dramatically into Europe with farmers from the south-east then originally I was thinking that implied the ancestors of the farmers started out adjacent to the northern HGs somewhere then moved south and then from there into Europe so for example Altai -> Mid-East -> Europe.

However following the LP example that wouldn't be necessary if those northern HG genes just gradually drifted south in low frequency where for some reason they gave a particular advantage to proto-farmers.

Tobus said...

@Grey:
If the skin color genes were initially northern HG but spread dramatically into Europe with farmers from the south-east


Only SLC24A5 spread dramatically into Europe with the farmers, who where pretty much already fixed for it. The SLC45A2 allele seems to have increased in Europe only after the Steppe incursion... that's according to the Mathieson "8000 years" paper anyway.

Krefter said...

@Tobus,

So far all, EHG and CHG have SLC45A2. It isn't only from EEF, I'd say it's from Steppe and later intrusions form West Asia. IMO, it's important to remeber SLC45A2 rose in frequency because of natural selection not because of Steppe ancestry.

Romulus said...

SLC45A2 did not originate in West Asia amongst ANE, neither Afonta Gora nor Malta boy was positive for it. The oldest samples found containing it are all found in Scandinavia (Stora Forvar, Motala), exactly where it has its peak distribution today. Obviously to anyone with 2 brain cells it originated in Europe. It doesn't matter if 2 EHG samples have it because they contain significant WHG admixture which other highly ANE populations without the allele do not.

Let me put it in plain english, ANE R1 people, lightened up, by mixing with WHG. They originally looked something like a Native (South) American. Paleolithic Europeans are the source of the light skin, light eyed, European Phenotype.

Krefter said...

@Romulus,

Like I said before, the key is Late Neolithic Europe. Forget about the Paleolithic. Irish are mostly a straight up mixture of Neolithic_Anatolia and Yamnaya. Neolithic_Anatolia and Yamnaya were both wogs. How do you explain modern Irish, then? The only explinaton is natural selection. And we actually have it documented, there was a huge rise in frequency of Pale-alleles in Late Neolithic Europe.

Tobus said...

@Romulus: Let me put it in plain english, ANE R1 people, lightened up, by mixing with WHG.

That's simply not correct - none of the three WHG samples have any of the known light-skin variants. The ancient European pops with you are talking about with the derived SLC45A2 are SHG, not WHG, and these pops didn't directly contribute to the modern European genome, not even in Scandinavia.

The primary skin-lightening allele is SLC24A5 which was fixed in Anatolian farmers (and elsewhere) before they arrived in Europe, and the frequencies in Early European farmers are slightly lower than the Anatolian levels, indicating WHG mixture *lowered* the frequency, not increased it. The secondary allele in SLC45A2 probably has a North-East European origin (earliest samples are SHG and EHG) but its frequency in Europe was minimal before the Yamnaya incursion, suggesting that modern Europeans owe their high levels of it to Steppe, not WHG, ancestry.

As Krefter correctly points out, selection for SLC45A2 in modern populations didn't happen until the Bronze Age, so the high frequencies in Mesolithic Scandinavia are a red (pickled?) herring.

Romulus said...

@Tobus

The secondary allele in SLC45A2 probably has a North-East European origin (earliest samples are SHG and EHG) but its frequency in Europe was minimal before the Yamnaya incursion, suggesting that modern Europeans owe their high levels of it to Steppe, not WHG, ancestry.

SLC45A2 was also present in TWELVE of the Anatolian Neolithic Samples with approximately ZERO ANE ancestry so explain how it has anything BUT an WHG origin.

Tobus said...

@Romulus:..so explain how it has anything BUT an WHG origin.

...because it doesn't occur in any WHG sample. How can you seriously suggest a WHG origin for something that not one single WHG sample has?

Romulus said...

shg is whg.

Romulus said...

look at a pca, any pca, first visit to the site tobus?

Romulus said...

every single sample which has tested positive for slc45a2 has had substantial whg admix. you have to have a borderline iq not to be able to figure this out

epoch2013 said...

@FrankN

"I generally would like to see less discussion of skin colour, which is obviously prone to racist (mis-)interpretation, but of little help in understanding prehistory."

First, I stopped using the word "racism" or measuring myself to anti-racist moral standard a long time ago. It is a meaningless and empty word. I don't care the tiniest bit if something I say is either considered racism or the opposite. I tend to piss of both blacktivists and stormfronters on a equal scale on other forums and that is the last bit I would like to say about it.

However, Europeans tend to be quite divers in looks (eye color, hair color) but not in skin color. That makes it interesting enough, I would say.

That being said you have a point: The most interesting part of La Brana that was wildly overlooked in my humble view are the genes associated with resistance to pathogens. Similarly it is quite interesting that among the Neanderthal genes that were selected for quite a number are involved in metabolism.

epoch2013 said...

@FrankN

https://westhunt.wordpress.com/2015/07/31/tlrs-pamps-and-alley-oop/

Romulus said...

@epoch

The only trait that can make one human superior to another is encephalization but there can be exceptionally encephalized members of any race so "racism" is truly stupid. Its tragic that the massive importance of brain size is so abudantly clear in the record of human evolution and yet most of us are intentionally shielded from this fact due to its inflamatory nature.

Amanda S said...

Hey, Krefter. Just be careful with the Aussie use of the word "wog". It's an awful word in the UK, very taboo, and it refers to a different group of people. In Australia it's kind of jokey rude, like, Pom and refers to Mediterraneans on the whole.

Tobus said...

@Romulus:
shg is whg

.. in which peer-reviewed published paper is that? SHG and WHG are sometimes combined into just HG, but I don't recall ever seeing Motala included as a WHG sample - probably because SHG contains ANE that WHG doesn't!

look at a pca, any pca

Well, presumably not *any* PCA - it would have to be one with both SHG and WHG, like the one from Lazaridis 2014. Take another look - you can clearly see that SHG and WHG are NOT the same populations, SHG has a truckload more ANE for instance.

every single sample which has tested positive for slc45a2 has had substantial whg admix.

.. but heaps of samples *with* WHG admix *haven't* been positive for it - indeed, the pure WHG samples don't have it at all! I think what you are seeing is confirmation bias.

epoch2013 said...

@FrankN

Come to think of it, Haak et al stated that the shift from MN to LN needed an extra shot of WHG. We know that WHG lived along EEF for a long while. This interests me as it may indicate that whatever happened WHG was better prepared for it than EEF. What if, as some have suggested, the demise of EEF was partly due to the coming of the plague with the Indo-Europeans? WHG being better prepared could be more disease resistant, or even having fixed genes for that only part of EEF would have. We at least now have some clue that whatever set the WHG apart was actively selected for.

Even if hypothetical: It would be an interesting exception to the perceived rule that farmers wipe out HG's by bringing diseases.

Krefter said...

@Romulus,

Me and Tobus are referring to West European-WHG and you're referring to everything WHG-related. EHG had WHG-related ancestry, but they weren't West European WHGs like Loschbour. Ones like Loschbour, did not have Light-skin mutations. You're right all the 6000 BC West Eurasians who had Light-skin mutations, shared lots of WHG-related ancestry. And you're right that could be the ultimate source.

@epoch2013,

Maybe population-size is the reason EEF replaced most WHG genes. Contact between EEF/WHG and EEF/WHG/Steppe in the British isles, resulted in almost complete population replacement. Contact between EEF/WHG and EEF/WHG/Steppe in Iberia and Italy, resulted in lots of replacement(30-50% in most parts), but there was lots of survival of EEF/WHG.

Differences in population-size of native EEF/WHG in the British isles and Italy/Iberia, could be why there were differnt levels of population replacement.

epoch2013 said...

@Krefter
@FrankN

The Netherlands and North-Germany have excellent excavation records. I think FrankN already presented some sites where FB sites were far more numerous than the successor CWC sites.

I have somewhere a paper on CWC sites in the Netherlands. It's a large paper but worth downloading and reading. I'll try and find it and post a link. From what I read sites from the fringes of the FB culture survived far better than main FB sites. I have a hunch these fringe sites are the local successors of Vlaardingen and Swifterband sites, local Dutch spawns of Ertebolla like cultures which almost certainly were WHG continuation cultures living alongside LBK.

Grey said...

@Tobus

"Only SLC24A5 spread dramatically into Europe with the farmers, who where pretty much already fixed for it."

Yes, I should have specified.

Grey said...

@Tobus

"As Krefter correctly points out, selection for SLC45A2 in modern populations didn't happen until the Bronze Age, so the high frequencies in Mesolithic Scandinavia are a red (pickled?) herring."

It may be a red herring in terms of proximate source but I disagree it's one in terms of ultimate source.

The more southerly WHG may not have had those genes but the mixed northern edge (SHG / EHG ) etc did have them so it seems likely the original source was in the far north (as expected).

The *proximate* source in modern Europeans being mostly farmers and Yamnaya would make sense given they rolled over WHG while the SHG / EHG were only along the northern border.

However that still begs a question *if* the ultimate source was in the north then how did the Anatolian farmers get them. Either the proto farmers came from the north originally bringing those genes with them or they drifted south by random chance and then spread dramatically cos farming selected for them for some reason.

If we ever get enough genomes from the early farmers it will be interesting to see when the frequency of those alleles shot up (if that is what happened).

(If the alleles were originally northern and their selection was connected to farming then you'd expect the frequency to be low everywhere and then shoot up dramatically in the farming regions.)

Tobus said...

@Grey:
I think you are probably right with regards to SLC45A2 - it originated in the North-East (whether in SHG, EHG or some xHG that we haven't found yet), and spread to AN/EEF in low numbers (~20% frequency), presumably via North-South interaction in the Steppe/Caucasus area or further east. It came to Europe in small numbers via EEF and then increased dramatically (>75%) when the Yamnaya arrived from the Steppe, who IIRC only had it at ~40% themselves.

SLC24A5 is different story however, it is essentially fixed in the North, East and South in all the Meslothic and Neolithic samples we have, and came to Europe in the high 90%s with EEF. It's origin could be anywhere east of Europe and it's probably much older than SLC45A2, and selected for at a different time/place.

An interesting question is why SLC24A5 is fixed in the farmers whereas SLC45A2 stayed at ~20% until the Bronze Age, as both alleles give similar effects to the phenotype supposedly under selection.

Grey said...

@Tobus

Fair enough; I can never keep the labels straight in my head anyway so not a subject I should argue about too much.

"An interesting question is why SLC24A5 is fixed in the farmers whereas SLC45A2 stayed at ~20% until the Bronze Age, as both alleles give similar effects to the phenotype supposedly under selection."

That gets to the crux of why it matters imo (or might matter).

Matt said...

Interesting question raised by Tobus.

The drop in SLC45A2 in Europe during the Early Neolithic, from AN->Iberian_MN looks proportionate to WHG ancestry (around 10% drop from 40% in AN->30% in CEM, proportionate to 10% WHG ancestry, then 10% more between Central European Neolithic->Iberia Neolithic, which seems again consistent with the greater WHG in the pooled Iberian sample).

That's from Mathieson - http://tinyurl.com/hgqvy3f

So it looks close to no selection since entry of the AN population to Europe (instead the variant just drops proportionate to WHG admixture). Why would that happen?

Btw, for EHG, based on Sandra Wilde's results (see ANL - Aneolithic steppe), this looks like - http://i.imgur.com/gTtggio.png. Most likely frequency 20% for EHG, based on four samples.

I'm pretty inclined to trust that, since her results for other alleles on the steppe frequencies tend to be consistent with Mathieson (Lactase, HERC2, SLC24A5) - http://i.imgur.com/Ii9M6DV.png.

So simple CHG+EHG admixture wouldn't seem to explain Yamnaya's frequency of SLC45A2 (a higher frequency reduced). Seems like contra to MN Europe, the steppe did experience selection on this variant, unless the 40% frequency in Yamnaya is just intermediate a higher frequency in their Neolithic ancestors and a lower frequency from EHG?

Tobus said...

@Matt: Most likely frequency 20% for EHG, based on four samples.

That doesn't sound right. Samara is fixed derived at SLC45A2 and Karelia is heterozygous... so EHG should be 75%. Is EHG the "ANL" on that image? If so it looks like they got their ancestral/deriveds mixed up.

Tobus said...

I checked Wilde's paper again - the "ANL" are from Smyadovo in Bulgaria 4500-3000 BC, whereas EHG are from much further north and west in Russia 4700-4000 BC. I don't think these can be considered the same population. Just from their location I'd expect ANL to be much closer to WHG or even CHG than to EHG.

Matt said...

Tobus, thanks for checking and those populations on the Western edge of the Black Sea makes sense given the Western Yamnaya / Ukraine focus of Wilde's study. I would expect them to be more transitional between EHG and WHG, and not EHG given that, so less informative (uninformative?) for what EHG were like.

Grey said...

Matt

"So it looks close to no selection since entry of the AN population to Europe (instead the variant just drops proportionate to WHG admixture). Why would that happen?"

You'd think if it was originally selected for among the farmers and then dipped it would bounce back after dipping.

Might a caste system slow that down? If so maybe it [did / would have] bounced back over time but slow so IE arrived before complete?

Grey said...

Also...

the other thing i wonder about is what if it's not skin but eyes?

I read that in arctic conditions SF use black camo paint around the eyes to stop light reflecting into their eyes causing snow blindness.

What if at the equator there's too much light in the eyes so dark eyes/skin reduces it but as you go further north you need either lighter skin to reflect light into the eye *or* lighter eyes.

It's just a thought but if WHG had light eyes as their solution to the problem then there'd be less selective pressure for the new lighter skin allele (being another region's solution to the same problem).

Just speculating - evidence for it might be if eye problems were a bigger deal than rickets for dark eye/skin ppl in the north.

(I googled for this once briefly but got bored before i found anything.)