April 24, 2014

More ancient Scandinavians (Skoglund, Malmström et al. 2014)

A new paper has just appeared in Science which adds new ancient Swedish hunter-gatherer samples, as well as a new Gökhem2 Swedish farmer. Much lower quality data from the same archaeological sites were studied in 2012 by much the same team, but the new study has sequenced several new Pitted Ware individuals from Ajvide, as well as a  Mesolithic Swede.

The Swedish hunter-gatherers appear to be similar to those of Lazaridis et al. (2013) in that their ancestry is a mixture of both European hunter-gatherers like LaBrana1 and ~15% of something related to MA1, which seems quite close to the 19% of ANE ancestry for the older Motala hunter-gatherer also from Sweden. The finding of Y-haplogroup I2a1 also parallels the Motala hunter-gatherers, so everything seems quite consistent with the Mesolithic Swedes being genetically very close to the Pitted Ware Neolithic ones. However, there is one difference in that the new hunter-gatherers were ancestral for SLC24A5 while the Motala one was derived (this is the "skin lightening" allele that was curiously missing in both Iberia and Luxembourg hunter-gatherers).

The authors also find that the Iceman and Gökhem2 are a mixture of Basal Eurasians and something related to hunter-gatherers. A interesting new detail is that the Swedish farmer had more of the hunter-gatherer ancestry than the Iceman (the estimated difference in their non-Basal Eurasian ancestry is 77.2-56=21.2%) which seems reasonably close to the 16% difference in the related "Atlantic_Baltic" ancestry for the previous lower-quality Gok4 farmer and the Iceman I estimated in 2012.

Finally, the authors also study the genetic diversity of the Swedish hunter-gatherers:
The Scandinavian Neolithic hunter-gatherer group had significantly lower conditional nucleotide diversity (0.181±0.0015) compared to the Scandinavian Neolithic farmer group (0.201±0.0038, Figs 3A and S9). While the specific properties of ancient DNA may still affect comparisons with sequence data from modern-day individuals, the conditional nucleotide diversity in the hunter-gatherers was also lower than in any modern-day European and a Chinese population (22) analyzed using the same approach as for the ancient groups.
It is not easy to estimate nucleotide diversity with low coverage data (because you can't tell whether a sample is heterozygous in some position if you only have a handful of reads covering it), but the authors cleverly use the fact that they have multiple individuals from the hunter-gatherer population to estimate this. The low diversity in hunter-gatherers also parallels the finding of low genetic diversity in the Luxembourgeois Mesolithic hunter-gatherer, so it does seem that hunter-gatherers in Europe were a very low diversity population, which seems reasonable for people engaging in foraging (which does not allow for growth to large population numbers) and having ancestors who endured the Ice Age in Europe.

The last few months have been extremely generous in new ancient DNA studies and I hope that more stuff is coming this year as in 2013.

UPDATE: Also important (from the Independent):
“We see clear evidence that people from hunter-gatherer groups were incorporated into farming groups as they expanded across Europe. This might be clues towards something that happens also when agriculture spread to other parts of the world,” Dr Skoglund said.

Science DOI: 10.1126/science.1253448

Genomic Diversity and Admixture Differs for Stone-Age Scandinavian Foragers and Farmers

Pontus Skoglund, Helena Malmström et al.

Prehistoric population structure associated with the transition to an agricultural lifestyle in Europe remains contentious. Population-genomic data from eleven Scandinavian Stone-Age human remains suggest that hunter-gatherers had lower genetic diversity than farmers. Despite their close geographical proximity, the genetic differentiation between the two Stone-Age groups was greater than that observed among extant European populations. Additionally, the Scandinavian Neolithic farmers exhibited a greater degree of hunter-gatherer-related admixture than that of the Tyrolean Iceman, who also originated from a farming context. In contrast, Scandinavian hunter-gatherers displayed no significant evidence of introgression from farmers. Our findings suggest that Stone-Age foraging groups were historically in low numbers, likely due to oscillating living conditions or restricted carrying-capacity, and that they were partially incorporated into expanding farming groups.

Link

121 comments:

Krefter said...

I am wondering what I2a1-P37.2 subclade did the hunter gatherers belong too? Is there any Y DNA from the farmers? Also what hair and eye color did the farmers and hunter gatherers likely have?

Fanty said...

"A interesting new detail is that the Swedish hunter-gatherer had more of the hunter-gatherer ancestry than the Iceman "

The Swedish FARMER you mean?

jackson_montgomery_devoni said...

Were any new mtDNA results reported for the newly hunter-gatherer and farmer remains?

Dienekes said...

>> The Swedish FARMER you mean?

Indeed.

Grey said...

Does I2 HGs with derived SLC24A5 genes imply female transmission of those genes?

If so I wonder if it came via the megalith farmers of the Atlantic coast of the west, the LBK farmers to the south or the forest to the east?

#

"so it does seem that hunter-gatherers in Europe were a very low diversity population"

It seems to me if there's a population B derived from a subset of a population A then B should always be lower diversity than A simply from the fact they're a subset. So populations at the end of a migration chain should be less diverse.

Also, if populations are adapting to different environmental conditions along that migration chain then they should be getting even less diverse due to selection pressure and bottle necks.

The thing about that - if correct - is that if the relative age of two clades of y DNA is being measured solely by their diversity then it could be wrong as population B could be exactly the same age as A but much less diverse because B went through the migratory winnowing process.

Chad said...

Gokhem has 21% Basal ancestry correct? That would make them about 50% EEF like myself, if that is the case. Very interesting!

German Dziebel said...

The most interesting graph is Fig. S7 in Suppl Mat, which shows Amerindians and Australians as basal to all non-Africans. The authors have this to say:

"To test the robustness of our main conclusions we repeated the fitting procedure described above and included genomes from a modern-day French and a modern-day Sardinian individual, repeating the analysis 10 times and picking the run with the highest log-likelihood. This analysis suggested mixed ancestry in these modern-day groups (represented by single individuals) similar to the farming-associated ancient individuals from Europe (Figure S7). We found that in this fitted model, the lineage leading to Neolithic farmers and modern-day Sardinians and France no longer formed the “most basal
lineage”...".

Here go with the wind Basal Eurasians and Ancient Northeast Asian.

Table S13 also suggests that Ajvide58 shares greater affinity with Karitiana than with MA-1 or Australian.

Shaikorth said...

Fig. S7 also shows Sub-Saharan Dinka admixture to Sardinians and both Neolithic farmers.

eurologist said...

The admixture between early farmers and local Mesolithic people is interesting, since it is the inverse of modern admixture results of dominating expansions (y-DNA by the dominating, expanding population, a lot of mtDNA from natives, often complete autosomal domination; e.g., much of the US until recently, Australia, New Zealand -- but note still at times up to 30-60% native autosomal DNA in groups growing to a majority, e.g., Mexicans in the SW US).

Conversely, the outcome here was up to ~30% autochthonous y-DNA ("I," especially in N and C Europe, but also in Sardinia), even after additional metal-ages invasions, now, and something like 60 - 80% autosomal HG DNS, then, and presumably ~30% - 50% autosomal now, even after additional invasions, in much (but not all) of Europe.

I think part of this can be explained by the fact that the Neolithic people were smaller, weaker, and had less weapon usage/ training; i.e., in contrast to modern expansions, they had no "military" advantage. As I have often mentioned, this means that there necessarily was "something in it" for the HGs. The agriculturalists were matriarchally organized, and evidently the elder women did not allow any contact with native girls and had no use for them, or did not believe they could be trained or trusted to work on the farms (lack of native mtDNA with very few exceptions, and those mostly early-on in the expansion in and around the Balkans, likely due to a deficiency of agriculturalists females, also as shown in the lack of agriculturalists' y- and mt-DNA in HGs). Conversely, it appears that native men (a lot taller and stronger*) had it easier to be allowed sexual contact or marriage.

A comment on the difference between ancient European and Siberian autosomal DNA: Both are closely related at a deep time level; more deeply than any other widely-distributed extant NW Eurasian population. Given the deep inferred time frame, it is highly likely that the connections go back to the Aurignacian and/ or Gravettian. One possibility is that the distinction is mostly early Gravettian (~35 - 30 kya), with a split between a W and E (Siberian --> Beringian) population, and likely Aurignacian admixture in most of the European population (I still believe that SE Europe was affectd differently by the Gravettian and then maintained ~25,000 y of difference in the "Epi-Gravettian" and thereafter).


* It is well known that early Near-Eastern farmers lost stature and health quickly, due to their limited and carbohydrate-rich diet. Some of that also became genetically selected for, perhaps due to annual or other semi-periodic famine. Conversely, N European HGs were not that dissimilar from extant Scandinavians or Baltic people - both quite a bit taller, and with more muscle mass.

Tobus said...

@German:
Table S13 also suggests that Ajvide58 shares greater affinity with Karitiana than with MA-1 or Australian

Only relative to La Brana, not in absolute terms. (I assume you are comparing the three D(Yoruba, Anzick/MA-1/Australian, Aj58, LB) scores, and I assume you mean Anzick, not Karitiana since Karitiana don't feature in S13) These results mean Aj58 is a a fair bit more like Anzick than LB is, a fair bit more (but slighly less than with Anzick) like MA-1 than LB is and not significantly any different to Australians than LB is - it doesn't mean Aj58 is more like Anzick than he is like MA-1 or Australians, you'd need something like D(Yoruba, Aj58, MA-1/Australian, Anzick) to be able to tell that using a D-stat.

Some of the other data in the paper suggests it is probably true in regards to Australians, but unlikely to be true in regards to MA-1 - the f3 stats in S4 and the and admixture graph in S6A both put Aj58 solidly in Europe and since MA-1 is midway between Europe and Anzick we'd expect Aj58 to be closer to MA-1.

mooreisbetter said...

What is MOST fascinating by all of these Neolithic discoveries is that Haplogroup I1a (P37.2) and likely M26 represent the first spread of FARMERS. Speculation had been rife before that it had been a Mesolithic haplogroup, right?

These discoveries prove a number of scientists right: Farming did spread in Europe through a DEMIC diffusion. As this paper notes, they mixed with the hunter gatherers along the way, and thus farming spread with both cultural and genetic mixing.

The model seems to be that truly Paleo Europeans were Haplogroup C (La Brana) and perhaps F* or others. Farmers spread I2a, Herders G2.

The interesting thing is where the Demic theories were wrong...unless...

The Demic theories (Cavalli Sforza, others) thought farming spread with Indo Europeans (R1a) out of Anatolia, then again, out of the Southeast and East.

This appears to be wrong.

The Indo European connection appears to be wrong.

Unless...

I had read a theory a number of years ago that was quickly shot down, which was that Western IE languages spread with Hg I.

If you look at HG I, modern P37.2 is the cradle of Slavic languages; modern I1 is the cradle of Germanic languages; M26 appears in sufficient numbers in Celtic homelands.

It's certainly intriguing.

I personally think it is more likely that I2a correlates with a pre-IE language group that we now call "proto-Iberian." This would include Basque, Vasconic, Aquitainian, and the Nuragic language of Sardinia.

It would also include whatever languages were spoken in pockets of Scandinavia, Italy and the Balkans before indoeuropeanization.

A self-prescribed ethnonym is the least likely to change. For example, the Greeks have called themselves Hellenes for millenia. Little could make them change that.

If you look at the current aggregate studies for distribution of M26

http://snplogic.blogspot.com/2014/03/corrected-study-confirms-even-more-i.html

you see places where the ethnonyms are:

(o) Skadi
(eu) Skara
Oski (Osci)

etc.

and likely similar place names in Sardinia:

http://en.wikipedia.org/wiki/Paleo-Sardinian_language

I know that's a lot of theories in one post. Hope it is digestable.

Unknown said...

German,

Table S13 also suggests that Ajvide58 shares greater affinity with Karitiana than with MA-1 or Australian.

If that statistic bears out across a number of samples, it could be explained by the branch of ANE directly ancestral to Ajvide being more similar to the branch of ANE that admixed into the Americas than is the ANE branch represented by MA-1.

Of course, you could also argue that MA-1 represents a population that splintered off in the Amerindian ---> European process and therefore has weaker affinity to Ajvide than a truly ancestral group (i.e. ancient Amerindians).

The most interesting graph is Fig. S7 in Suppl Mat, which shows Amerindians and Australians as basal to all non-Africans.

It really just shows ENA as basal, rather than Australians and Amerindians specifically, and it still shows Amerindians as admixed. The interesting difference is that SSA admixture assumes the role of Basal Eurasian admixture in all the other similar figures in this and other papers. But I can't see anything (either in ADMIXTURE or formal statistics) suggesting the farmers were SSA admixed.

Unknown said...

The interesting difference is that SSA admixture assumes the role of Basal Eurasian admixture in all the other similar figures in this and other papers.

Actually, there are no admixtures edges from Dinka to Gokhem2 or the French, so it doesn't completely assume the usual function of Basal Eurasian admixture.

Shaikorth,

Fig. S7 also shows Sub-Saharan Dinka admixture to Sardinians and both Neolithic farmers.

No, only to Sardinians and Otzi.

Unknown said...

@Tobus:
“These results mean Aj58 is a a fair bit more like Anzick than LB is, a fair bit more (but slighly less than with Anzick) like MA-1 than LB is and not significantly any different to Australians than LB is”

This fits with a model of the Solutrean Hypothesis: Let’s assume there’s a population ancestral to both MA1, and the MA1-like components in Ajvide58 and Anzick1. For sake of clarification, let’s call this ancestral population *P-ANE*. And let’s call MA1 *R-ANE*, and call the MA1-like components in Ajvide58 and Anzick1 *Q-ANE*. And let’s say that by the time LaBrana1’s population got to Spain, the descendants of the Solutreans had either crossed the Atlantic into the Americas, or were floating in and around Scandinavia. Here’s a plot of where the closest Y-chromosomes (Q-L804) to the major Native American clade Q-M3 are located: http://www.semargl.me/en/dna/ydna/haplotypes/maps/409/

AWood said...

It's clearly stated the hunter gatherer was I2a, and also stated that the farmer groups lived in relative (mostly, not complete) isolation from the adjacent hunter gatherers.

I'm not certain how anyone can interpret I2a being a farming lineage. I'm certain Sardinia is throwing you off. Although the island was settled in the Neolithic period, there is no reason to assume hunter gatherer weren't incorporated into the settlement.

Krefter 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

Romulus the I2a L233+ Proto Balto-Slav, layer of Corded Ware Women said...

It seems strange that we find I2a1 again and again in Mesolithic Sweden given that I1 is the dominant Y dna haplogroup in the area now. I would love to hear Dienkes theory on how that came to be.

My guess would be that Proto I1 had migrated to Northern Scandinavia and there was a secondary migration of I2a1 groups following in to southern Scandinavia. These were dispersed with the coming LGM. After the LGM I1 expanded southward out of a Nothern Scandinavian Refuge, partially mixing with some Haplogroup N in Finland. It's a conservative guess but at least it isn't some moronic invasion genocide theory.

It's a shame we don't know the mtDNA haplogroup of this individual.

Slumbery said...

mooreisbetter :

"What is MOST fascinating by all of these Neolithic discoveries is that Haplogroup I1a (P37.2) and likely M26 represent the first spread of FARMERS. Speculation had been rife before that it had been a Mesolithic haplogroup, right?"

Almost all of the identified Mesolithic and Neolithic HG Y-haplogrup is some kind of I (with the single exception of La Brana 1), while it is more rare among Neolithic farmers, so we know for sure that it did not spread with farmers first.

This does not mean that all the YHg I found in Neolithic farmers come from the local HG population of course, since the South-Eastern extent of YHg I in the Mesolithic is unknown.

"The model seems to be that truly Paleo Europeans were Haplogroup C (La Brana) and perhaps F* or others. Farmers spread I2a, Herders G2."

Is there such thing in the Neolithic Europe as distinct farmer and herder populations? I doubt it. Also G2 was found in very early classic farmers, so even if such a divided population did exist, G2 still cannot be bound particularly to herders.

Grey said...

@mooreisbetter

"What is MOST fascinating by all of these Neolithic discoveries is that Haplogroup I1a (P37.2) and likely M26 represent the first spread of FARMERS."

Alternatively it shows a HG to farmer transition in northern Europe that tallies with the y DNA I population expansion around the time of Funnelbeaker.

HG to farmer transitions are unusual which implies there was some specific reason why this transition occurred in the north but not elsewhere.

I would suggest the standard neolithic package didn't work that far north but slash & burn farming (like the Forest Finns until recently) did work and that form of farming is easier for HGs to adapt to.

All you'd need is for people to see runaway pigs and cattle happily grazing on the new growth that springs up after a forest fire.

There may even be recorded instances from around the world during the European colonial period of forest HG populations switching to slash & burn farming in a similar way.

.

"Speculation had been rife before that it had been a Mesolithic haplogroup, right?"

Not speculation. Mesolithic finds had that DNA.

.

@eurologist

"The admixture between early farmers and local Mesolithic people is interesting, since it is the inverse of modern admixture results of dominating expansions (y-DNA by the dominating, expanding population"

Yes.

"I think part of this can be explained by the fact that the Neolithic people were smaller, weaker, and had less weapon usage/ training; i.e., in contrast to modern expansions, they had no "military" advantage."

The size / health advantage existed further south as well though.

I'd say the military advantage the first farmers had was numbers - relative population density, so I'd suggest they *lost* their numbers advantage further north because the neolithic package couldn't so completely out-compete the locals in a northern forest ecozone.

Instead I think a slash & burn version of farming was developed in the far north (hence the lack of settlements) which the indigenous HGs were much better adapted for than the regular form of farming.

Hence the population expansion of native I which allowed them to survive better in the north than most other places.

ren said...

@German: That's a good catch about how Anzick has greater affinity to Ajvide58 than Mal'ta does. The simplest model to explain that is that both Ajvide58 and Mal'ta are admixtures of an Anzick-like ancient Siberian population.

This is also the simplest model to explain the original Raghavan and Lazaridis data without a need to postulate additional Basal Eurasian and Ancestral North Eurasian hypothetical populations. Now I see people making it even more convoluted with ANE population A and ANE population B.

Unknown said...

Narcissus,

Now I see people making it even more convoluted with ANE population A and ANE population B.

What non-convoluted explanations do you have for the many, many facts that don't remotely fit your theory?

Tobus said...

@Narcissus:That's a good catch about how Anzick has greater affinity to Ajvide58 than Mal'ta does.

The data doesn't show this - it's a misreading of the D-stat results. Individual D-stat runs count different alleles sets (check the loci column for the ones in question!) and the results aren't guaranteed to be directly proportional or otherwise relative to each other.

To determine which of Anzick or MA-1 is closer to Aj58 we'd need a D-stat using both MA-1 and Anzick as a pair and Aj58 and an outgroup as the other pair, which we don't have. The PCA plot in the paper proper, the f3 in S4 and the trees in S6/7 all indicate that MA-1 is closer.

Rob said...

@ Romulus

"It seems strange that we find I2a1 again and again in Mesolithic Sweden given that I1 is the dominant Y dna haplogroup in the area now. I would love to hear Dienkes theory on how that came to be. "

Cannot be sure, however, I2a1 actually *is* present in the north (Scandinavia) and NW (British Isles) at low but nevertheless important frequencies. I2a must have dispersed throughout Europe with the Mesolithic. Its derivatives survived only in pockets, pockets, eg Sardinia (which it was almost 'fixated') , northwestern Europe, and somewhere in the Carpathians, where it much later flourished with the Slavic expansions. Why it declined -inter alia- cannot be sure, but usual factors such as drift, admixture and simple population 'die-out' are all possible.

Anc contrary to what Eurologist has said, the notion that Neolithic people were peacful little Mediterranean is demonstrably incorrect. In fact, there is evidence of endemic warfare in the Neolithic, which only became pacified during the Bronze Age - exactly the opposite to what outdated explanations have long assumed

terryt said...

"Now I see people making it even more convoluted with ANE population A and ANE population B".

I believe the one thing of which we can be sure: the historic genetic threads in any region are very convoluted. Is there any particular reason why they should not be? Unless you believe humans spread in huge numbers from some single Garden of Eden.

Unknown said...

Tobus,

The data doesn't show this

Agreed. I can't believe I missed the fact they weren't talking about what the data does show: that there's a larger difference between LB-1 and Aj58 in their relation to Anzick than in their relation to MA-1.

I agree, this is strange if both populations are related to Anzick via MA-1's population. But if we consider that Aj58 may have received admixture from an ANE source closer -- even just slightly closer -- to that which admixed into Anzick's population, I think that would explain the Z score. I don't even agree with Narcissus that this is particularly convoluted. We already see reasonably large differentiation between MA-1 and AG, and these are the only two ANE samples we currently have (although it must be mentioned that AG's genome is lower quality), and we know that it was a Q population that likely admixed into the Americas, which MA-1 didn't have and possibly his entire population lacked. We also know that Q exists at low frequencies in Europe, and this may have been introduced by the same population that admixed into Swedish hunter gatherers.

Kristiina said...

"It's a shame we don't know the mtDNA haplogroup of this individual."
In the supplementary materia which is freely available they say "The hunter-gatherer individuals belonged to haplogroup U
(Ajvide59), U5a1 (the Mesolithic StoraFörvar11), U4d (Ajvide53, Ajvide58, Ajvide70
and Ire8) and V (Ajvide52) (Table 1). These results are consistent with the high
frequencies of U found in hunter-gatherer populations in Scandinavia (5, 7) and in
Europe (6, 10, 56-60). Furthermore, the haplogroups observed among the TRB
individuals, were H (Gökhem4), H1c (Gökhem2), H24 (Gökhem7) and K1e
(Gökhem5). The H and K haplogroups found here are also seen in other Neolithic
farming populations (6, 58, 61-64)."

If I only remember correctly, in Lazaridis paper also U2e was found. Here, farming populations carry H haplogroups and K. In a previous paper, the Scandinavian farmer carried hg T. U4 has been constantly found in the Mesolithic burials of the far north: in Russian Karelia, Central Siberia.

Kristiina said...

"It's a shame we don't know the mtDNA haplogroup of this individual."
In the supplementary materia which is freely available they say "The hunter-gatherer individuals belonged to haplogroup U
(Ajvide59), U5a1 (the Mesolithic StoraFörvar11), U4d (Ajvide53, Ajvide58, Ajvide70
and Ire8) and V (Ajvide52) (Table 1). These results are consistent with the high
frequencies of U found in hunter-gatherer populations in Scandinavia (5, 7) and in
Europe (6, 10, 56-60). Furthermore, the haplogroups observed among the TRB
individuals, were H (Gökhem4), H1c (Gökhem2), H24 (Gökhem7) and K1e
(Gökhem5). The H and K haplogroups found here are also seen in other Neolithic
farming populations (6, 58, 61-64)."

If I only remember correctly, in Lazaridis paper also U2e was found. Here, farming populations carry H haplogroups and K. In a previous paper, the Scandinavian farmer carried hg T. U4 has been constantly found in the Mesolithic burials of the far north: in Russian Karelia, Central Siberia.

Grey said...

A map of european climate zones

http://en.wikipedia.org/wiki/File:Floristic_regions_in_Europe_(english).png

If the boreal zone extended further south into modern northern Germany/Poland when it was colder such that its western edge was adjacent to the Atlantic zone around Denmark and its eastern edge adjacent to the Pontic zone northwest of the Black Sea what do you get?

you get a map that (roughly) mirrors the distribution of:
-cardium
-megalithism
-LBK
-Yamnaya (including the territory of the advanced cultures of the Western Black Sea that were squished)
-Corded Ware (made up of the combined TRB territory in the west and Globular Amphora territory in the east

If correct then neolithic swedes would turn out to be a combination of Atlantic farmers and traders (possibly different sources) and Boreal HGs and (imo) mixed Boreal farmers.

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.

German Dziebel said...

@Hamar Fox

"It really just shows ENA as basal, rather than Australians and Amerindians specifically, and it still shows Amerindians as admixed."

Amerindians as admixed in that chart is just inertia of thinking. There can be no gene flow from a downstream lineage. The branching pattern already explains the affinity between Amerindians and MA-1 nicely.

"Of course, you could also argue that MA-1 represents a population that splintered off in the Amerindian ---> European process and therefore has weaker affinity to Ajvide than a truly ancestral group (i.e. ancient Amerindians)."

Yes.

"The interesting difference is that SSA admixture assumes the role of Basal Eurasian admixture in all the other similar figures in this and other papers. But I can't see anything (either in ADMIXTURE or formal statistics) suggesting the farmers were SSA admixed."

Could it be that extreme derived non-Amerindianness gets interpreted as SSA?

@Tobus

"The data doesn't show this - it's a misreading of the D-stat results. Individual D-stat runs count different alleles sets (check the loci column for the ones in question!) and the results aren't guaranteed to be directly proportional or otherwise relative to each other.

To determine which of Anzick or MA-1 is closer to Aj58 we'd need a D-stat using both MA-1 and Anzick as a pair and Aj58 and an outgroup as the other pair, which we don't have. The PCA plot in the paper proper, the f3 in S4 and the trees in S6/7 all indicate that MA-1 is closer."

Can you show a calculation that shows that my inference from the D-stats is wrong? Without such a calculation we should stay with what the data shows. Fig. S4 doesn't show the greater proximity of MA-1 to Aj58 over Anzick. The trees can be ambiguous about this particular point. Admixed populations can shift around a tree quite a bit. At S7 Aj58 shows a longer branch than MA-1, which is more consistent in length with Amerindians and Australians. A PCA would show greater proximity between Aj58 and MA-1 on the basis of more frequent and derived West Eurasian alleles. Ancestral connections typically get lost in PCAs.

CarolAST said...

The Saamis are still off on their own island, and everybody ignores them.

Unknown said...

Hmm from the genetic drift data it looks to me like red "Basal Eurasian" originated in the western mediterranean (Sardinia/France) whereas the medium blue (La Brana) has radiated out from Scandinavia.

From the heat maps in the supplement both clearly have a connection to North America (although there were not enough points to deduce point of entry (east or west) or if it was an anomaly. Looks real to me though.

So tentatively::
red = ancient med probably western med since it is surprisingly scarce in the modern eastern med.

medium blue = ancient NW European
light blue = ancient N eurasian
magenta = ancient NE Asian?? maybe
green = ancient S Asian

Someone asked about mtDNA
"haplogroup U
(Ajvide59), U5a1 (the Mesolithic StoraFörvar11), U4d (Ajvide53, Ajvide58, Ajvide70
and Ire8) and V (Ajvide52) (Table 1). These results are consistent with the high
frequencies of U found in hunter-gatherer populations in Scandinavia (5, 7) and in
Europe (6, 10, 56-60). Furthermore, the haplogroups observed among the TRB
individuals, were H (Gökhem4), H1c (Gökhem2), H24 (Gökhem7) and K1e
(Gökhem5)."

Adrian Purcell Heathcote 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.”

I don’t see how this follows, since I must be an earlier Haplogroup than J. Wouldn’t it make more sense to postulate that J must have come from the hunter gatherers, or even better that some ancestral group to both (IJ) split and one went one way and the other the other.

Adrian Purcell Heathcote said...

Tobus, you are wasting your breath on German. One doesn’t get to be an internet crank from following arguments and looking dispassionately at the evidence. One gets there by adopting a bizarre view and then trying to force all evidence into that view. Nothing must be allowed to show that the view is just nonsense.

It’s natural selection at work.

We have to imagine herds of internet cranks drifting across the wireless plains. When one comes upon a watering hole — a forum — he will try to colonise it, by urinating on the nearest stationary object. Then there is a conversion process — everyone must repeat his catechism, exactly as he dictates. Weak and unfit internet cranks will crumble when there is sufficient evidence put against their view. They become extinct — hollow shells mumbling their view to general derision. They are like alkies begging a drink from the nearest bar. The successful internet crank has seen what has happened to these lesser creatures and knows that he must not show even a moment’s rationality. Flash some credentials — but not slowly so that anyone can see what they are, and then repeat often that you have them; insult the majority who disagree with you — why, they are lesser simply by being more numerous. They lack the heroic isolation that true idiocy rewards. And — and this is most important — always go on at far greater length than your opponents. Remember: no one ever got to be a successful internet crank by being succinct. Only if he follows these rules can the internet crank hope to pass on his memes to some poor unfortunate simpleton who believes that ‘hey, the guy must be on to something’.

eurologist said...

Anc contrary to what Eurologist has said, the notion that Neolithic people were peacful little Mediterranean is demonstrably incorrect

Rob,

I have said nothing like that. Also, almost all known sites of conflict are between agriculturalists, not between HGs and farmers.

HGs clearly were the much better skilled fighters, since they practiced related skills everyday, while the farmers were busy sowing, weeding, digging trenches, helping animal births, milking, making cheese, digging clay, shearing wool, making hay, harvesting, tending to their animals during ~4 months of winter indoors, etc.

And the farmers were extremely vulnerable, despite their numbers: all that freshly dried hay and grain and thus the entire livelihood of the extended family living in that longhouse can go up in smoke in no time on a dry and windy autumn night.

There is no way farmers had expanded that swiftly or at all without the cooperation, rather than antagonism, of the HGs.

Tobus said...

@German: Can you show a calculation that shows that my inference from the D-stats is wrong?

The correct question is can *you* show a calculation that shows your inference is *correct*? ... or do expect us to just accept any old assertion without the need to show that it's logically and mathematically valid?

I'd like to point out that I'm not saying the conclusion is definitely wrong (although I'd be very suprised if it were correct), just that the evidence you provided for it is not evidence at all... in the same way that a photography expert doesn't say or prove there's no such thing as UFO's, just that this particular photo is a fake, I'm not saying or proving that Aj58 *isn't* closer to Anzick, just that these particular D-stats don't show it.

So, please explain the rationale behind your inference, in particular:
a) how you think D-stats work - what do you think they measure and how do you think they measure it?
b) how you used the D-stats to arrive at the conclusion that Aj58 is closer to Anzick than to MA-1?

Grey said...

@jackson_montgomery

"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."

Scandinavia is a hotspot for Y DNA I so there's no need for that.

.

Carol
"The Saamis are still off on their own island, and everybody ignores them."

Once the "total replacement from the near east" idea has finally died I expect the Saami will be looked at more.

Unknown said...

German's job is to critically explore the evidence from the perspective of the idea of "Out of America". He is attempting to do that, and I respect him for trying. Albeit I personally think that the evidence is overwhelmingly against this idea. His scientifically critical challenges of the data are good for science and this blog. He just needs to avoid attacking the person rather than the science, as do we all.

I think his different perspective is valuable, particularly in relation to very real flows of population back out of the Americas in more recent times.

Ryan said...

@Tobus

" I'm not saying or proving that Aj58 *isn't* closer to Anzick, just that these particular D-stats don't show it."

It may actually be closer, but only by proxy. There was probably a back-migration around the time of the Arctic Small Tool Tradition's expansion. The Anzick-1 paper shows an affinity between people on the Alaskan coast and those in Northern Europe today. There's linguistic evidence for it too (that Dienekes previously blogged about). If the proto-Uralic homeland was on the west side of the Urals, it seems that they may have mixed with some proto-Dene or proto-Saqqaq from the other side of the Urals.

Doesn't really prove or disprove German's hypothesis though.

Rob said...

Eurologist:
I just don't think you can make sweeping statements about cooperatio, or antagonism. The fact is - the process of Neolithicization was very variable. Every sub-region was different. And although the notion of HG-farmer interaction, acculturation, etc is a hot-topic today, for uch of Europe, this still remains to be demonstrated clearly. Ie many Mesolithic HG assemblages have not been shown wihtouht doubt to lie within Neolithic settlements, and vice-versa.

"HGs were much better fighters". Well, I havent read any papers on the military archaeology of Hunte-gatherers, but (no offence) that sounds like blatant speculation on your behalf, without saying you're inccorect. And it again, smacks of the sweeping unsubstantiated statementship one encounters on this blog time and time again.

Rob said...

Eg, Whilst Zvelibil makes a case case for HG-famer interaction zone in the Eat Baltic, this appears to have *not* been the case in the Low Countries, esp the regions fo modern Nehterlands. Here, there is a notable break between HG and the LBK farmer offshoot, which in turn, the disappears , before new farming groups came somewhat later.

That is what all genetics papers lack, a real , detailed knowledge of the evidence 'from the ground' , and how processes were highly diverse. That is why their broad-brush conclusions are at best, simplistic, at worst, totally off the mark.

terryt said...

"or even better that some ancestral group to both (IJ) split and one went one way and the other the other".

I would think that is the most obvious explanation. The IJ split is at least 37,000 years ago. And J is almost certainly a 'south of the Caucasus' haplogroup and I almost certainly 'north of the Caucasus', or at least from near the Black Sea.

"Tobus, you are wasting your breath on German. One doesn’t get to be an internet crank from following arguments and looking dispassionately at the evidence. One gets there by adopting a bizarre view and then trying to force all evidence into that view. Nothing must be allowed to show that the view is just nonsense".

I second those comments. Carried? I especially liked, 'They lack the heroic isolation that true idiocy rewards'.

Unknown said...

“The Swedish hunter-gatherers appear to be similar to those of Lazaridis et al. (2013) in that their ancestry is a mixture of both European hunter-gatherers like LaBrana1 and ~15% of something related to MA1, which seems quite close to the 19% of ANE ancestry for the older Motala hunter-gatherer also from Sweden.”

In the “Admixture Analysis of ‘World’ Dataset” for K=4:

https://docs.google.com/spreadsheet/ccc?key=0ArJDEoCgzRKedGR2ZWRoQ0VaWTc0dlV1cHh4ZUNJRUE#gid=1

“Swedish_D” breaks down as 91.2% “European” and 8.8% “Amerindian”, and for K=3 its breakdown is 97.5% “European and 2.5% “Asian”.

I’ll bet most of this 8.8% “Amerindian” is from the ANE-like admixture in these Ajvide and Motala hunter-gatherers, and is related at least in part to y-dna Q. I also suspect the Vikings carried it in similar amounts. Here’s a map of the distribution of y-dna Q in Europe:

http://www.eupedia.com/europe/Haplogroup_Q_Y-DNA.shtml

Interestingly from the map, Q peaks in three places: Southern Sweden right near where the Ajvide and Motala samples were found; Sicily, where the Vikings and Normans invaded and mixed in; and in Eastern France, right near Solutre.

Fanty said...

@Grey:

I think what Jackson wants to say is, not an explanation why Scandinavian Hunter Farmers apear to be mixed with WHG, but why EEF in general (Spanish and Italian aswell) apear to be "basal Eurasian mixed with a WHG-like population".

While some people believe thats because the middle eastern Farmers mixed with WHG people in the Balkans before arriving in Italy/Southgermany/Spain, he thinks it could be that the Farmers had been "WHG-like" from the beginning, because of the shared anchestry with the WHG (I)

Then of course.... must be said that there are no Farmers yet that had been "J". All had been "G" but a single (or so) E.

But then again, G and E are anyways somewhat related to J and I. Like stuff like R is related to East Asian/Native American lineages. (thats why "R" would have made more sense for mesolithic Europeans than "I" does...)

Katharós said...

I don’t think HGs could afford to stay for long in the same area to resist farmers , given that HGs were dependent on their Natural environment for food. Through the cultivation of land, farmers slowly fortify themselves in an area. With the long term effect that they would have displaced some HGs groups or at least their lifestyle. I wonder if this displacement effect would have led to an increasing level of competition among HGs themselves.

Tobus said...

@Denis:This fits with a model of the Solutrean Hypothesis:

It may well do. More importantly though it shows that ancient Europeans have varying degrees of affinity with both MA-1 and Native Americans... indicating that these ancient European lineages had already diverged at the time of the European/Native American gene flow.

@Hamar:I agree, this is strange if both populations are related to Anzick via MA-1's population. But if we consider that Aj58 may have received admixture from an ANE source closer -- even just slightly closer -- to that which admixed into Anzick's population

The tree they present in the paper shows LB, Aj58 and MA-1 all with a common ancestor after the farmer lineage had diverged, but with Aj58 receiving additional admixture from MA-1's lineage after the three diverged. This is consistent with LB having more MA-1 than modern Europeans but less MA-1 than Aj58. They also have MA-1 admixing with Anzick after LB's lineage had diverged which is consistent with LB having a similar shared drift with Native Americans as modern Europeans do while Aj58 and MA-1 have more.

@Locrian: We have to imagine herds of internet cranks drifting across the wireless plains...

That brought a smile to my face :)

I do realise I'm unlikely to change German's mind, but I think it's important to point out when people are using the data incorrectly - even if German doesn't see it himself, it might help others avoid the same mistake. (@Dienekes: if you disagree and would prefer if I didn't "feed the troll", please let me know).

Grey said...

@Fanty

"While some people believe thats because the middle eastern Farmers mixed with WHG people in the Balkans before arriving in Italy/Southgermany/Spain, he thinks it could be that the Farmers had been "WHG-like" from the beginning, because of the shared anchestry with the WHG (I)"

Ah, my mistake. I wonder about that myself.


German Dziebel said...

@Tobus

"The correct question is can *you* show a calculation that shows your inference is *correct*? ... or do expect us to just accept any old assertion without the need to show that it's logically and mathematically valid..."

Just read around how these models work. Dienekes's comments policy says: "Google before you ask." For comparison, my comments policy on Anthropogenesis says: Don't post if you don't have a clearly verifiable identity.

You need to get into a habit of refraining from doubting what the data says and starting to supplant better mathematics if you believe the current ones are inadequate, all toward the goal of producing what is (hopefully) better data. If you want to become a scientist, this is a critical first step. There seems to be a clear connection between the D-stats values showing the special proximity between Aj58 and Amerindians and the drift-parameter branch length in Fig. S7. This means something. I think what this means is that MA-1 is more admixed than Aj58 (has South Asian component, for example), so to Hamar Fox's point, Aj58 is purer than MA-1 and hence closer to Amerindians. This also means that the depiction of MA-1 as unadmixed in the graph B in the main body of the paper is nonsense.

"I do realise I'm unlikely to change German's mind..."

Chimps can't change professors' minds, neither can professors change chimps' minds. But at least I can teach you human speech.

@Annie Mouse

"German's job is to critically explore the evidence from the perspective of the idea of "Out of America". He is attempting to do that, and I respect him for trying. Albeit I personally think that the evidence is overwhelmingly against this idea. His scientifically critical challenges of the data are good for science and this blog. He just needs to avoid attacking the person rather than the science, as do we all.

I think his different perspective is valuable, particularly in relation to very real flows of population back out of the Americas in more recent times."

This is a fair attitude to out-of-America. Something pseudoscientists such as Tobus, Locrian and TerryT should adopt. I just don't think I'm engaging in personal attacks. I'm just critically exploring the evidence that Tobus, Locrian and TerryT are scientifically minded. Maybe they are but there's some suggestive evidence for their recent backflow into creationism.

Unknown said...

Sorry, after reading further down in that Eupedia link I gave, it appears that the y-dna Q found in Sicily (Q1b1a) isn’t related to the Scandinavian Q, or the Native American Q (which is apparently nearly all Q1a, and overwhelmingly Q-M3).

curious student said...

I think, the expansion of I1 is probably related to farming subsistence, and out of a HG population that picked up farming. Then they expanded in to the HG territory of I2 in scandinavia, and soon outnumbered them.
Farming technology allows up to hundred times higher population density. If two populations differ two hundred years in adopting it, the first will already outnumber the second. When all land is under cultivation, it prevents easy ( without war) expansion,
With the probably very small first populations genetic drift was a stronger factor, Q relative I2 and such.

curious student said...

I think, the expansion of I1 is probably related to farming subsistence, and out of a HG population that picked up farming. Then they expanded in to the HG territory of I2 in scandinavia, and soon outnumbered them.
Farming technology allows up to hundred times higher population density. If two populations differ two hundred years in adopting it, the first will already outnumber the second. When all land is under cultivation, it prevents easy ( without war) expansion,
With the probably very small first populations genetic drift was a stronger factor, Q relative I2 and such.

curious student said...

Forgot my question, If its known what yhaplo the Gök-farmer belongs to?

eurologist said...

I don’t think HGs could afford to stay for long in the same area to resist farmers , given that HGs were dependent on their Natural environment for food. Through the cultivation of land, farmers slowly fortify themselves in an area. With the long term effect that they would have displaced some HGs groups or at least their lifestyle. I wonder if this displacement effect would have led to an increasing level of competition among HGs themselves.

Creative,

But that does not describe an accurate picture of early (say, LBK) farmer live in Central and W and later N Europe. Farming for millennia was done adjacent to rivers, in lands that were periodically flooded, so fertility was high (almost exclusively in Loess soils), but native growth was mostly brush and young and small trees, which were easily eliminated by cutting them in spring and burning the dry remains in the dry fall, and at the same time destroying any seeds of weeds (there were virtually no native seeds that had adopted to fire germination on these wetlands).

There simply was no displacement of HGs. They persisted in the forests and ~80% or so less fertile lands, and persisted having the fishing rights on important river runs and lakes.

Tobus said...

@Annie:His scientifically critical challenges of the data are good for science and this blog.

If (and when) his challenges are scientifically valid then I agree, this can only be good for science. Unfortunately having looked at the data beneath many of his statements I have found that his "scientific" critiques are more often than not semantics and/or misunderstanding (sometimes intentional misrepresentation?) of the data. I certainly started out taking him at face value and found his ideas interesting, but after discovering his sources don't actually support what he's saying a number of times, I now take everything he says with a large grain of salt.

@German:
Just read around how these models work.

I know how D-stats work, which is why I'm contesting your conclusion - can you tell me how *you* think they work and how you came to your conclusion?

Dienekes's comments policy says: "Google before you ask."

Just like it says: "Be polite. Use facts and arguments"?

I just don't think I'm engaging in personal attacks

... and yet you just called me a "chimp" and a "pseudoscientist" in this very same post! Moreover you did so instead of explaining why you think these D-stats show Aj58 has more Anzick than MA-1 - it's exactly the behavior Annie is talking about: "He just needs to avoid attacking the person rather than the science". If you have evidence and reasoning for what you are saying then bring it - if you don't then you'll just have to accept that you might be wrong, but please don't just call people names, it won't help, no matter how "true" you think your insult might be.

ren said...

@Tobus: It shows that there is a relationship between an ancient Amerind and an ancient Scandanavian to the exclusion of Mal'ta, which is essentially the same logical basis for the Mesolithic European-Mal'ta-Amerind link to the exclusion of East Asians.

We are never talking about absolute genetic distances for if that were the case, Mal'ta is just in-between Han and Sardinian and we can say he is a mix of the two.

Again, the simplest model to explain the interlocking math is an Anzick population admixing into Mal'ta and Mesolithic Europeans.

@Harma: And the mtDNA counterpart of that Q seems to be C1e and C1f, which is too Amerind and old to be of Uralic or later waves.

Tobus said...

@ren:
It shows that there is a relationship between an ancient Amerind and an ancient Scandanavian to the exclusion of Mal'ta

Sorry ren, what shows this? If you are talking about the S13 D-stats that German brought up it's a case of misunderstanding the results - none of these D-stats measure Aj58 against MA-1 and Anzick.

the simplest model to explain the interlocking math is an Anzick population admixing into Mal'ta and Mesolithic Europeans.

This model is explicitly ruled out by the high affinity Anzick shows to East Asians, but the relative lack of affinity to East Asians shown by Mal'ta and Mesolithic Europeans. If Anzick gave DNA to Mal'ta, he'd have given him the affinity to East Asians as well. The only model that fits is with the gene flow going the other way, from Mal'ta's lineage into Anzick's.

eurologist said...

just don't think you can make sweeping statements about cooperatio, or antagonism. The fact is - the process of Neolithicization was very variable. Every sub-region was different.

Rob,

Initially, there where only three subregions: Cardium and LBK, and La Hoguette in the NW.

There is an extremely good archaeological record of the cohabitation of agriculturalists and HGs in the Belgium area, for example. It did happen, and by all indications, it was the norm.

Katharós said...

@eurologist

Generally speaking, I’m not sure how many square km of pure forest is needed to survive , but I would assume that it would be difficult for bands of 30 to 50 people depending on such a forest environment.Either way, there must have bin some selective pressure among HGs themselves,increased by Farmers.

HGs clearly were the much better skilled fighters,

This aspect is interesting. I have often herd of accounts that WW2 soldiers who had a pre-war profession of butchering animals, were the most brutal and sadistic soldiers.

Whisper said...

I noticed Anne Mouse said "Furthermore, the haplogroups observed among the TRB
individuals, were H (Gökhem4), H1c (Gökhem2), H24 (Gökhem7) and K1e
(Gökhem5)."
Does H3s fit in the Norse group also?

Unknown said...

Tobus,

The tree they present in the paper shows LB, Aj58 and MA-1 all with a common ancestor after the farmer lineage had diverged, but with Aj58 receiving additional admixture from MA-1's lineage after the three diverged. This is consistent with LB having more MA-1 than modern Europeans but less MA-1 than Aj58. They also have MA-1 admixing with Anzick after LB's lineage had diverged which is consistent with LB having a similar shared drift with Native Americans as modern Europeans do while Aj58 and MA-1 have more.

I agree, but what I meant was that LB and Aj58 differ more in their relationship to Anzick than in their relationship to MA-1. I'd like to see the statistic reproduced using other Swedish samples contrasted with, say, Loschbour. If it turns out to be a trend, then a more detailed model may have admixture going from a different ANE population, maybe better represented by the Afontova Gora population, into both the Swedish HGs and Amerindians. This, I think, would still explain the increased general affinity of Aj58 to MA-1 and Amerindians relative to LB's affinity to the same (as is clear in the tree presented), but may also explain the additional issue of why Aj58 has a (slightly) disproportionately higher relationship to Anzick relative to LB.

Harma,

It shows that there is a relationship between an ancient Amerind and an ancient Scandanavian to the exclusion of Mal'ta, which is essentially the same logical basis for the Mesolithic European-Mal'ta-Amerind link to the exclusion of East Asians.

If you want to argue there was Amerindian admixture specifically into Swedish HGs (which is all that what you say above relates to), feel free. Whether or not that's true, it has no bearing on the wider picture of ANE admixture into Amerindians.

As usual, though, you'd of course have to explain why the 'East Asian' element that would be present in the descendants of any European and modern-Amerindian-like admixture event is not detected. I believe my proposal above better explains the current data.

Again, the simplest model to explain the interlocking math is an Anzick population admixing into Mal'ta and Mesolithic Europeans.

On the basis of physical anthropology, PCA plotting & fst distances in relation to the strength of the Amerindian ADMIXTURE signal, and the distribution of the EDAR gene, it was obvious to me that this hypothesis was false even before the five or six recent papers completely justifying my scepticism.

And the mtDNA counterpart of that Q seems to be C1e and C1f, which is too Amerind and old to be of Uralic or later waves.

Any Y-DNA and mtDNA can be married. All you need is men from one place and women from another.

ren said...

@Harma: I'm not going to address your various claims. This discussion is probably fruitless. Let me just point to one specific thing as an example.

EDAR actually reaches fixation to near-fixation in Amerinds, not in Asia. For example, the Karitiana are 100% EDAR derived.

http://alfred.med.yale.edu/alfred/SiteTable1A_working.asp?siteuid=SI663326A

Tobus said...

@Hamar:
but what I meant was that LB and Aj58 differ more in their relationship to Anzick than in their relationship to MA-1

This is really just a red herring - the "Loci" column shows that the f4 count was done on very different SNPs in the MA-1 and Anzick stats, probably due to MA-1 sharing less than 30% of his DNA with Anzick, and LB/Aj58 both being considerably closer to MA-1 than Anzick (and hence having fewer "AB" differences at MA-1/Yoruba "AB" sites than at Anzick/Yobura "AB" sites). These raw scores aren't directly comparable with each other.

then a more detailed model may have admixture going from a different ANE population, maybe better represented by the Afontova Gora population, into both the Swedish HGs and Amerindians

That fact that MA-1 plots closer to Native Americans than AG does (SI 29 in Raghavan) makes me doubt this - I'd say the ANE population that gave DNA to Native Americans would have to be closer to MA-1 than AG.

Unknown said...

Harma,

I'm not going to address your various claims. This discussion is probably fruitless.

Always keep in mind that one day everything we know about ancient DNA will perfectly explain the genetic relationships among modern populations. A lot of your theories ignore this and require a rewriting of the distances and similarities of moderns.

I'm interested in seeing the data and theories add up to what actually exists. Nothing more. If you have arguments as to why my observations are incorrect, I'd welcome further discussion with you and be fully open to your ideas.

EDAR actually reaches fixation to near-fixation in Amerinds, not in Asia. For example, the Karitiana are 100% EDAR derived.

And is absent in the western descendants of ANE (and WHG) populations. EDAR shows no evidence of deselection. The fixation in Amerindians is no doubt related to the cold adaptations necessary in NE Siberia and Northern N. America during the last ice age, in addition to the Bering bottleneck.

It doesn't seem to have been a burdensome adaptation in other climes, though, as is seen its overall global distribution (which includes Eastern and even in one case West Africa!) and also its distribution in the Americas.

So why would it have been deselected in Europeans? Why does this lack of EDAR coincide with the presence of an Amerindian ADMIXTURE signal in EDAR-less European populations but no East Asian signal in those same populations? Why does it coincide with formal statistics not detecting strong evidence of East Asian admixture in these populations? Why are the ANE and WHG admixed Basques less Asian shifted than almost all other W. Eurasians, including those with minimal HG ancestry and only slight recent E. Eurasian admixture?

And so on.

terryt said...

"And is absent in the western descendants of ANE (and WHG) populations. EDAR shows no evidence of deselection".

Good point. I wonder how German would explain that problem.

German Dziebel said...

@Hamar Fox

"And is absent in the western descendants of ANE (and WHG) populations. EDAR shows no evidence of deselection."

"Why does this lack of EDAR coincide with the presence of an Amerindian ADMIXTURE signal in EDAR-less European populations but no East Asian signal in those same populations?"

The EDAR issue is very interesting indeed. I'm going three ways at it: 1) deselection in West Eurasians (remember shovel shaped incisors are more frequent in some ancient Europeans than in modern Europeans and in some Africans, including San and West Africans they are more frequent than in others); 2) more recent, Late Pleistocene-early Holocene back migration out of North America into East Asia. The higher frequencies of EDAR in the Americas over Asia suggest that it originated in the Americas. In the Americas, too, the derived EDAR allele spread from north to south. In Asia, it marked the emergence of the Mongoloid phenotype and it follows its pattern of distribution with decreasing frequencies westward and southward. Modern East Asians, the way we know them, emerged as a result of a Holocene gene flow from North America, which brought the new EDAR variant with it. 3) ancient structured EDAR positive and EDAR negative population(s) in the Americas: West Eurasians formed on the basis of one, East Asians formed on the basis of the other (but much earlier than Holocene).

I presently most intrigued by the second model as it fits nicely with MA-1 which is EDAR negative and correspondingly doesn't have an East Asian component. It has an Amerindian component proving that ancient Amerindians were EDAR negative. This explains why EDAR-less European populations have an Amerindian but not East Asian admixture.

@TerryT

"Good point. I wonder how German would explain that problem."

See above. With usual ease and elegance. ;)

@Tobus

"and yet you just called me a "chimp" and a "pseudoscientist" in this very same post! Moreover you did so instead of explaining why you think these D-stats show Aj58 has more Anzick than MA-1 - it's exactly the behavior Annie is talking about: "He just needs to avoid attacking the person rather than the science". If you have evidence and reasoning for what you are saying then bring it - if you don't then you'll just have to accept that you might be wrong, but please don't just call people names, it won't help, no matter how "true" you think your insult might be."

I already explained why I think Aj58 is closer to Anzick than to MA-1. Go back and re-read it. If you want to challenge this, go ahead and provide a better formula. For now a comparison between separate D-runs has supplied a meaningful pattern, which is supported by the amount of shared drift metric in the tree.

"Pseudoscientist" is not a personal attack. It's an accurate description of the way you approach scientific material and discussions. I know it hurts for you to read it but you just have to change your ways. There's no other way around it. Get a degree and it will hopefully go away. I use descriptive terms ("pseudoscientists") and tropes ("chimp") to refer to your blatant lack of a scientific demeanor (with respect to facts and to people) and your arrogance in presuming you can change the mind of a person with all possible academic credentials without furnishing any of your own and without making any appeal to facts or logic.

terryt said...

"See above. With usual ease and elegance".

And usual absolute nonsense. In turn:

"1) deselection in West Eurasians"

You fail completely to explain why West Eurasians would have undergone 'deselection' yet such is notably absent in all other regions East Eurasians managed to expand to. There is an obvious reason why you cannot explain it: because it is not the correct explanation.

"2) more recent, Late Pleistocene-early Holocene back migration out of North America into East Asia".

For which we have no evidence at all, apart from a very minor possible back migration in the extreme north, which would have carried the mutation to Western Eurasia anyway.

"3) ancient structured EDAR positive and EDAR negative population(s) in the Americas: West Eurasians formed on the basis of one, East Asians formed on the basis of the other (but much earlier than Holocene)".

Doesn't make sense, and doesn't fit with any evidence, especially when we consider that many Amerindian populations far removed from the Arctic regions have near fixation levels of the mutation.

"I presently most intrigued by the second model as it fits nicely with MA-1 which is EDAR negative and correspondingly doesn't have an East Asian component".

Yet if the mutation had derived from Amerindians MA-1 would certainly have the mutation prominently, in spite of your convoluted attempt at explanation:

"It has an Amerindian component proving that ancient Amerindians were EDAR negative".

To make that at all convincing you have to explain what gave rise to the near 100% fixation in parts of America. Once again you explanation makes no sense at all.

"I know it hurts for you to read it but you just have to change your ways. There's no other way around it. Get a degree and it will hopefully go away".

I suspect he has a degree, probably in biology. Unlike you. He certainly knows much more biology than you do.

"arrogance in presuming you can change the mind of a person with all possible academic credentials"

Don't you mean, 'arrogance in presuming you can change the mind of a person with an absolutely closed one'?

terryt said...

"I presently most intrigued by the second model as it fits nicely with MA-1 which is EDAR negative and correspondingly doesn't have an East Asian component. It has an Amerindian component proving that ancient Amerindians were EDAR negative. This explains why EDAR-less European populations have an Amerindian but not East Asian admixture".

Once more German completely contradicts himself. He has been arguing continually that Amerindians are not a hybrid between an MA-1-like population and an East Asian population, yet here he is claiming exactly that. He now claims two separate populations in America: an 'original' EDAR negative American population and an EDAR positive population which evolved from that earlier one. Presumably he will claim they each separately evolved in America and then, to reach the mix we see there today, the two populations next mixed. No explanation as to how or why either event happened. The EDAR negative population is exactly equivalent to the MA-1 population and the EDAR positive population is exactly equivalent to the East Asian population, with no crossover in either Eurasian population. That is extremely difficult to fit with any 'out of America' scenario. Surely the most obvious, and most likely, explanation is that the EDAR- and EDAR+ populations evolved separately in the separate regions where each have been found: In Central Asia and in East Asia respectively. And they mixed shortly before entering America with more of the the East Asian element following on after the initial entry. That scenario explains all the available evidence completely. To fit both populations separately into America we are obliged to postulate an extremely convoluted and unlikely series of events. I think German has just shot himself in the foot. Yet again.

Tobus said...

@German:
I already explained why I think Aj58 is closer to Anzick than to MA-1.

...and I already explained to you why that interpretation is unfounded, if you read above you can see that I've explained the problem with it in detail, multiple times. The two D-stats you refer to are counting a largely different set of SNPs within the genome and they have different baseline sizes (see the Loci column!), hence the final figures aren't directly comparable and can't be used to infer LB's or Aj58's relative affinity to MA-1 or Anzick. They are useful in telling us which of LB or Aj58 is closer to MA-1/Anzick, but they can't tell us whether LB or Aj58 themselves is closer to MA-1 or Anzick - we'd need a D(Yoruba; LB/Aj58, MA-1, Anzick) for that.

A basic "smoke test" for any interpretation of the data is whether it provides consistent results across all the data, and in this case if we apply your logic to other D-stats we get conflicting results, confirming that it's an invalid way of reading the results. If you look down to the Gok2 section of Table S13 you can see in D(Y, x; Gok2, Sardinian) that MA-1 gets a lower score than Anzick indicating (according to you) that Gok2 is closer to Anzick than to MA-1, but just below that we have D(Y, x; Gok2, Iceman) showing MA-1 getting a *higher* score than Anzick, indicating the exact opposite. If the maths wasn't enough to convince me your method is flawed, the fact that it gives contradictory answers is absolute proof - this is not a valid way to interpret D-stat results.

"Pseudoscientist" is not a personal attack

On the contrary it's a textbook example... compare with something like: "Your reasoning is pseudoscientific because... (list of logic/data errors)". By placing a negative label on me instead of responding to my critiques you are attacking the person, not the argument.

presuming you can change the mind of a person with all possible academic credentials

I think your use of "all possible" here is a little overstated - do you have a degree in Biology? Genetics? Statistics? An undergraduate degree in any of these would be more relevant to this particular topic than your History and "Dances With Wolves" Anthropology PhDs. Not that any academic credential would excuse you from supplying facts and logic to support your position - even Stephan Hawking makes mistakes.

Gary said...

The finding of a native American component in ancient Scandinavia populations should not be all that surprising. Archeologists and cultural anthropologists have long noted similarities between Scandinavians and Native Americans of Northern North America, including such cultural traits as the importance of the sauna/sweat lodge and mythology. Take for instance the recent find of a well-reserved bow and arrow set that was uncovered by a retreating glacier in Norway:

"No one knows exactly who left these ancient hunting instruments, but the bow and arrows have a design that's strikingly similar to those found thousands of miles away in other frigid landscapes, such as the Yukon, Callanan said." (www.livescience.com/40058-snow-reveals-neolithic-bow-arrow.html).

There is also linguistic evidence of such an ancient connection. The Swedesh word for 'and', och, is strikingly similar to the equivalent in Algonkian Lenape, ok.

Last weekend, I was talking to a Norwegian lady. When I mentioned that I had been researching the possible relationship between Native American and European languages, she immediately volunteered: "We have never had any doubt that we (Scandinavians) are somehow related to Native Americans."

Gary said...

By the way, regarding the ANE connection, note the distribution of YHG I on this map. While the frequency falls off rapidly outside if Europe, it does reappear in central Siberia in the region west of Lake Baikal.

http://en.wikipedia.org/wiki/File:Haplogroup_I_%28Y-DNA%29.PNG

Unfortunately, I have not been able to find a breakdown of the subtypes of YHG I found in Siberian populations.

A corresponding map for YHG Q in Europe can be found at:

http://www.eupedia.com/europe/Haplogroup_Q_Y-DNA.shtml

German Dziebel said...

@Tobus

"and I already explained to you why that interpretation is unfounded, if you read above you can see that I've explained the problem with it in detail, multiple times. The two D-stats you refer to are counting a largely different set of SNPs within the genome and they have different baseline sizes (see the Loci column!), hence the final figures aren't directly comparable and can't be used to infer LB's or Aj58's relative affinity to MA-1 or Anzick. They are useful in telling us which of LB or Aj58 is closer to MA-1/Anzick, but they can't tell us whether LB or Aj58 themselves is closer to MA-1 or Anzick - we'd need a D(Yoruba; LB/Aj58, MA-1, Anzick) for that."

Well, you just have to do the D(Yoruba; LB/Aj58, MA-1, Anzick) then to prove me wrong. All of the above is an apology for your lack of counterevidence.

"A basic "smoke test" for any interpretation of the data is whether it provides consistent results across all the data, and in this case if we apply your logic to other D-stats we get conflicting results, confirming that it's an invalid way of reading the results."

This is an example of pseudoscience. Your statement is meaningless because you haven't provided an apples-to-apples comparison between my inference from available data and an inference from the data generated with a "better" formula.

" If you look down to the Gok2 section of Table S13 you can see in D(Y, x; Gok2, Sardinian) that MA-1 gets a lower score than Anzick indicating (according to you) that Gok2 is closer to Anzick than to MA-1, but just below that we have D(Y, x; Gok2, Iceman) showing MA-1 getting a *higher* score than Anzick, indicating the exact opposite."

It's too premature to call it "noise." There could be an explanation for it. Just put some thinking into it.

"I think your use of "all possible" here is a little overstated - do you have a degree in Biology? Genetics? Statistics? An undergraduate degree in any of these would be more relevant to this particular topic than your History and "Dances With Wolves" Anthropology PhDs. Not that any academic credential would excuse you from supplying facts and logic to support your position - even Stephan Hawking makes mistakes."

I studied all of these sciences. But most importantly, any honor's essay devoted to the study of human beings, whether playing Indian such as some folks in Europe or playing scientist such as you, Tobus, is vastly superior to any Ph.D. earned for a study of Drosophila, which is what the science of genetics was built on. You just naively believe in the omniscience of scientists, whereas, for an anthropologist, they are much more interesting as objects of an ethnographic research than producers of solid knowledge about modern human origins.

German Dziebel said...

@TerryT

"Surely the most obvious, and most likely, explanation is that the EDAR- and EDAR+ populations evolved separately in the separate regions where each have been found: In Central Asia and in East Asia respectively. And they mixed shortly before entering America with more of the the East Asian element following on after the initial entry. That scenario explains all the available evidence completely. To fit both populations separately into America we are obliged to postulate an extremely convoluted and unlikely series of events. I think German has just shot himself in the foot. Yet again."

You didn't just shoot yourself in the foot, Terry, with this scenario. You've suicide-bombed yourself and all of your intellectual buddies. You've just postulated two random places in the Old World in which EDAR+ and EDAR- presumably evolved, then they got mixed in the right amount by a divine bartender just to be delivered on time to a population on a lookout for such an unlikely destination as Beringia.

Just forget about EDAR-. It's ancestral state and it doesn't matter where it "evolved." But the fact that Amerindian admixture in Europe does not seem to be associated with EDAR+ tells us that the Amerindian migration to Europe predated the evolution of EDAR+. The fact that EDAR+ is more frequent in America than in Asia suggests that it evolved in America and not in Asia. And since EDAR+ is found in East Asia, it must have got there from the Americas. And since EDAR + was not around in East Asia 24,000 years ago, it must have got there from America at a later time. There could be other explanations (and I hinted at them above) but this one seems to be the most parsimonious.

terryt said...

"I studied all of these sciences".

Not according to the list of your qualifications I've seen. Mostly anthropology and marketing.

"But most importantly, any honor's essay devoted to the study of human beings, whether playing Indian such as some folks in Europe or playing scientist such as you, Tobus, is vastly superior to any Ph.D. earned for a study of Drosophila, which is what the science of genetics was built on".

You've just unveiled your creationist viewpoint. Do you really believe that the study of Drosophila can tell us nothing about humans? Creationist German showing his complete lack of understanding of the field of evolutionary biology!

"Just forget about EDAR-. It's ancestral state and it doesn't matter where it 'evolved'."

Of course you're keen to avoid considering EDAR. It is yet one more piece of evidence that is impossible to reconcile with your belief.

"You've just postulated two random places in the Old World in which EDAR+ and EDAR- presumably evolved, then they got mixed in the right amount by a divine bartender just to be delivered on time to a population on a lookout for such an unlikely destination as Beringia".

Again you demonstrate you total lack of any understanding of evolutionary biology. We can be sure that EDAR+ evolved in a single region. So what was the point of your comment? The only conclusion is that it was designed simply to avoid having to face the difficulty with your belief that I exposed.

"then they got mixed in the right amount by a divine bartender just to be delivered on time to a population on a lookout for such an unlikely destination as Beringia".

'Unlikely destination'? As a destination Beringia is a far more unlikely destination for an American population facing not population pressure than it is to a Eurasian population coming under pressure during the Eurasian Late Upper Paleolithic. And what on earth do you mean by 'mixed in the right amount by a divine bartender'? You are here showing your basal creationist outlook. Surely any mixture is the product of random events, not the product of pre-destination.

"But the fact that Amerindian admixture in Europe does not seem to be associated with EDAR+ tells us that the Amerindian migration to Europe predated the evolution of EDAR+".

Surely the lack of EDAR in Europe shows that your whole belief system is totally incorrect. There was never any ' Amerindian migration to Europe', either before or after the evolution of the EDAR370A mutation.

"The fact that EDAR+ is more frequent in America than in Asia suggests that it evolved in America and not in Asia".

Rubbish. I agree the mutation is more frequent in America than it is in parts of Asia, but is certainly not more frequent than it is in parts of East Asia. Your claim is without merit. Even a blatant lie designed to mislead.

"since EDAR+ is found in East Asia, it must have got there from the Americas".

You can still claim that in spite of the overwhelming evidence that the mutation didn't reach many regions where you claim Amerindians reached? Your pig-headedness is striking but alarming.

"And since EDAR + was not around in East Asia 24,000 years ago, it must have got there from America at a later time. There could be other explanations (and I hinted at them above) but this one seems to be the most parsimonious."

There is another extremely parsimonious explanation, but as it doesn't fit your creationist belief you refuse to consider it. The scientists who brought the mutation to our attention claimed the mutation arose 35,000 years ago, and so it was in East Asia 24,000 years ago. It was just not present in the population MA-1 was part of. It had not been able to reach so far north. Simple? I'm sure it is but your lack of understanding of evolutionary biology is preventing your understanding of the situation in either East Asia of America.

Tobus said...

@German:
Well, you just have to do the D(Yoruba; LB/Aj58, MA-1, Anzick) then to prove me wrong

So we accept any statement as fact now without needing any evidence to back it? I can make up any baloney I like and you have to prove it's false? That's not how science works German. You are proposing a unique and undocumented interpretation of a well-defined mathematical measurement - you need to show what you are doing is mathematically valid and logically consistent before you can present it as fact.

This is an example of pseudoscience. Your statement is meaningless because you haven't provided an apples-to-apples comparison between my inference from available data and an inference from the data generated with a "better" formula.

Then we must disagree on what "Science" is. To me it's a set of logically proven facts that can be used to consistently explain various phenomena. For example "a^2 + b^2 = c^2" is a fact that is consistent across all values of a, b and c as sides of a right-angle triangle. "E = mc^2" is consistent across all values of m as mass. It's very easy to make up a rule based on your own observations, but unless it's consistent across all values then it's not a rule at all, it's a false pattern (in your words, "pseudoscience").

D-stats of "D(A, B; X, Y)" are a genuine "fact" in that they consistently measure X or Y's increased affinity to A or B - you never see a case where positive means B and X have the higher affinity, it's always positive is A-X and negative is B-X... it's a mathematically valid and logically consistent pattern. What you are proposing however, is that you can tell if Aj58 is closer to MA-1 or Anzick by comparing the results of D(Y, MA-1, LB, Aj58) and D(Y, Anzick, LB, Aj58), and this is something outside the bounds of how D-stats are defined or used by any researcher. If this is a mathematically valid and logically consistent interpretation then the same interpretation needs to hold true in all other cases where we have two D-stats with only a different "B" population - essentially you are positing a rule that the difference between D(A, B1, X, Y) and D(A, B2, X, Y) represents the difference in affinity of X to B1 and X to B2. To see whether this rule might be correct we can test it against the data we already have - look through all the D-stats which vary only in the "B" population and see if your interpretation produces consistent results. Since the logical structure and maths of each D-stat we compare is identical, then we are comparing "apples with apples" as you say and the pattern you see should apply globally. If the pattern holds true in all cases then it may well be valid, despite the lack of a formal mathematical proof, but if the pattern is only sometimes there, then it's clearly a case of "pseudoscience" - just a random phenomenon with no meaning behind it.

As I've already pointed out your proposed pattern fails in the case of Gok2, where we get contradictory answers about his relative affinity to MA-1/Anzick when using Otzi compared to Sardinians... it was a nice idea, but sorry, it just doesn't hold up as a logically consistent method of interpretation (ie it's "pseudoscience").

It's too premature to call it "noise." There could be an explanation for it. Just put some thinking into it.

I didn't call it "noise" I called it a contradiction, and any explanation you come up with for the Gok2 case could equally apply to the Aj58 case, undermining the validity of your initial conclusion.

You just made it up, it has no mathematical reasoning and it produces inconsistent results... maybe time to consider it's just wrong?

Tobus said...

@German (cont):
any honor's essay devoted to the study of human beings...is vastly superior to any Ph.D. earned for a study of Drosophila

Well you would say that because that's your chosen field of expertise, "superior" is relative though - a PhD in anything isn't as good a 2-week course in mechanics when you're stuck in the middle of the desert with a broken car. I use a totally different set of criteria when assessing what qualifications are "superior" when deciding who's going to fix my broken leg to when deciding who's going to fix my plumbing.

I'm sure Anthropology is superior in many areas, but in terms of this discussion - using a mathematical formula to measure specific genetic traits - I think you'd have to agree that maths and genetics have more direct relevance.

for an anthropologist, they are much more interesting as objects of an ethnographic research than producers of solid knowledge about modern human origins.

.. and he calls me the science-denier!

We're not talking about human origins here, we're talking about how to interpret a particular mathematical measurement of genetic variation - and (despite your PhDs in topics completely unrelated!), you're doing it wrong.

German Dziebel said...

@Tobus

"Then we must disagree on what "Science" is."

Just simply do the calculation D(Yoruba; LB/Aj58, MA-1, Anzick) , will you? Don't debate what Science is until you provide some data to think about.

"It's very easy to make up a rule based on your own observations, but unless it's consistent across all values then it's not a rule at all, it's a false pattern."

It's a pattern. Whether it will hold the test of time, I don't know. But the science of human origins is an iterative process. You extract a provisional pattern from the data, then test it. You haven't tested my pattern. And my pattern is not based on just a single comparison between two D scores. It's supported by the similar proximity between shared drift for Anzick and Aj58 estimates on the "tree" graph.

"I didn't call it "noise" I called it a contradiction, and any explanation you come up with for the Gok2 case could equally apply to the Aj58 case, undermining the validity of your initial conclusion."

Why? Just come up with a tentative explanation for the Gok2 case and let's see if it would work for Aj58. Sardinian is a population that experienced a full impact of the farmer gene flow, Iceman is an early farmer. Maybe there's something here...

"I use a totally different set of criteria when assessing what qualifications are "superior" when deciding who's going to fix my broken leg to when deciding who's going to fix my plumbing."

Just fix it then. You've been talking a lot but I haven't really seen you in any action.

"We're not talking about human origins here, we're talking about how to interpret a particular mathematical measurement of genetic variation - and (despite your PhDs in topics completely unrelated!), you're doing it wrong."

Above you've decided that you don't need to PROVE me wrong because this is not how you think "science" works. But you continue to SAY that I'm wrong. It's time to stop talking and start doing. You are just one calculation away from getting some respect from me.

German Dziebel said...

@TerryT

"Rubbish. I agree the mutation is more frequent in America than it is in parts of Asia, but is certainly not more frequent than it is in parts of East Asia. Your claim is without merit. Even a blatant lie designed to mislead. "

Just review the table at http://alfred.med.yale.edu/alfred/SiteTable1A_working.asp?siteuid=SI663326A and count the populations in East Asia vs. America in terms of the deciles of frequencies of EDAR+, from highest to lowest. Exclude recently admixed ones in America. Then compare sample sizes for the two regions. Report your findings.

"The scientists who brought the mutation to our attention claimed the mutation arose 35,000 years ago, and so it was in East Asia 24,000 years ago."

You are comparing molecular clock with radiocarbon. Molecular clock is unreliable.

"You can still claim that in spite of the overwhelming evidence that the mutation didn't reach many regions where you claim Amerindians reached?"

For a person whose wits have been blunted by hours and hours of daily psalm singing you're predictably dense: I postulate two migrations out of the Americas: an earlier one to Eurasia in pre-EDAR+ times and a later one carrying EDAR+ to East Asia-only.

Tobus said...

@German:
Just simply do the calculation D(Yoruba; LB/Aj58, MA-1, Anzick) , will you?

What's stopping your from doing it? It's you who's making the claim after all... or is it your habit to state as fact something that you have no evidence for?

It's a pattern. Whether it will hold the test of time, I don't know.

It's already failed in the Gok2 data, as I've pointed out twice now.

Just come up with a tentative explanation for the Gok2 case and let's see if it would work for Aj58.

The explanation is that differences in D-stats with different B populations doesn't represent a consistent measure of relative affinity between then X and B populations used. And yes, it applies to Aj58 as well.

Just fix it then.

I'm trying to, but a person with no skills in the area keeps insisting on doing it his own way.

Above you've decided that you don't need to PROVE me wrong because this is not how you think "science" works

I've already proved you wrong - whether Aj58 is *actually* closer to MA-1 or Anzick isn't part my argument, as I made clear earlier. Whatever the result of such a D-stat might be, your method of assuming relative affinity from two different D-stats in not a valid interpretive method, as demonstrated by different allele sets used in the individual calculations, the lack of any underlying mathematical relationship between the two results, and the contradictory interpretations it produces when applied to a larger sample... that's why we need the suggested D-stat in the first place.

German Dziebel said...

@Tobus

"The explanation is that differences in D-stats with different B populations doesn't represent a consistent measure of relative affinity between then X and B populations used."

Your claim remains unfounded.

"the lack of any underlying mathematical relationship between the two results, and the contradictory interpretations it produces when applied to a larger sample..."

A pseudomathematical objection...

"I'm trying to, but a person with no skills in the area keeps insisting on doing it his own way."

D(Yoruba; LB/Aj58, MA-1, Anzick) is patiently waiting for you.

terryt said...

"Just review the table at http://alfred.med.yale.edu/alfred/SiteTable1A_working.asp?siteuid=SI663326A and count the populations in East Asia vs. America in terms of the deciles of frequencies of EDAR+, from highest to lowest".

Absolute numbers of populations are irrelevant, surely. There are fare more American samples than from any other single region and so naturally we will finish up with more American populations with a high level of the mutation than any other regions. Anyway the problem you have (apart from the lack of a basic knowledge of statistics) is that you want to include all samples listed as 'East Asian'? That's a ridiculous idea. Just take a look at the East Asian samples for yourself. If you exclude those from southern and western China you will see levels of the mutation much higher than most from America. The original study noted a gradual drop off in the level as they considered progressively more southern populations. That supports other evidence that shows a movement south of the Mongoloid phenotype from northern China during the Neolithic. If you're going to insist on lumping all the East Asian samples you are obviously going to minimise the mutation's level in East Asia. I presume that is why you're insisting on doing so.

"You are comparing molecular clock with radiocarbon. Molecular clock is unreliable".

The calculations have nothing to do with radiocarbon dating. If you have a problem with the authors' method of calculating the age of the mutation why don't you take it up with them? Meanwhile I am sure their method is superior to any method you are likely to come up with. Incidentally the original study noted diversity was much greater among northern Han than in any other population. That would surely be unlikely if the mutation had arisen somewhere other than in northern China.

"I postulate two migrations out of the Americas: an earlier one to Eurasia in pre-EDAR+ times and a later one carrying EDAR+ to East Asia-only".

That is a ridiculous claim, especially for someone who wrote a few days ago:

"You've just postulated two random places in the Old World in which EDAR+ and EDAR- presumably evolved, then they got mixed in the right amount by a divine bartender just to be delivered on time to a population on a lookout for such an unlikely destination as Beringia".

You have just now 'postulated two random places in the ... [New] World in which EDAR+ and EDAR- presumably evolved, then they got ... [separated] in the right amount by a divine bartender just to be delivered on time to [two] population[s] on a lookout for such an unlikely destination as [East Asia]".

What mechanism do you propose drove this apparent complete population separation in America. You propose a most unlikely scenario, but that is exactly the sort of thing we have all come to expect from you.

"or is it your habit to state as fact something that you have no evidence for?"

I'm sure you're only pretending surprise here.

Unknown said...

It might have been addressed elsewhere, but I've been wondering how archaic admixture levels fit with an out of America hypothesis.

I've not been able to find any comprehensive estimates of Neanderthal admixture in unadmixed Amerindian populations, but Sankararaman et al.'s recent paper shows Mexicans (a 3-way mix)to be more Neanderthal than either Europeans or Africans, suggesting that Amerindians are more Neanderthal than either:

http://dienekes.blogspot.co.uk/2014/01/neandertal-admixture-in-modern-humans.html

The other two Latin American populations have lower Neanderthal estimates than Mexicans, but also have 1) lower levels of Amerindian ancestry and 2) more African ancestry, especially Puerto Ricans.

So my question is when did this admixture between Amerindians and Neanderthals occur, or, alternatively, how did Europeans and especially Africans lose this admixture? Did admixture occur after the first out-of-America wave but before the second? Was there an extra bit of mixing after the second wave?

If Neanderthals introgressed into the Americas, we should expect to see some definite clines of Neanderthal admixture in Amerindians unmixed with Europeans or Africans, and this admixture should peak at a feasible Neanderthal entry point, while the out of Africa hypothesis would predict no cline, unless Amerindians mixed with a more aboriginal population with its own unique levels of archaic admixture (incidentally, I've seen some evidence of a Denisovan hotspot in S. America), or particular Neanderthal genes had a greater selective advantage in some environments than others. Still, introgression leaves a different pattern than selection, so we should be able to advance the debate once we have some figures to work with. If there are no clines or regional differences in the Americas, then no admixture occurred in the Americas, and so either Europeans and Africans lost some admixture somehow or they do not descend from Americans.

Of course, I don't have any evidence either way, so I'm not issuing challenges from a place of safety. I just think it's a healthy exercise to discuss what we should expect to see before we know the answers, to help us not keep talking in endless circles once the data actually becomes known.

So I predict no clines or regional differences, except for a difference between Arctic populations known to have more recent links with East Asia. There may also be a hotspot related to the observed Denisovan hotspot, which is not in a possible place of entry for a Neanderthal population, and so, if it's not a statistical error, could indicate admixture with an aboriginal population.

Tobus said...

@German:
Your claim remains unfounded

It has a mathematical explanation and is consistent across all the data, something that can't be said for the alternative.

A pseudomathematical objection...

So it's "pseudo"-mathematical to point out the lack of a mathematical basis and the contradictory results when applied to a wider sample? Who knew! I wonder what your idea of genuine maths is - unproven and inconsistent interpretations?




German Dziebel said...

@Hamar Fox

"I've not been able to find any comprehensive estimates of Neanderthal admixture in unadmixed Amerindian populations."

I don't have them either. But my suspicion is that it's higher than in East Asia (and of course higher than in Europe and Africa). Prufer (http://www.nature.com/nature/journal/v505/n7481/extref/nature12886-s1.pdf), Table S13.1 shows Karitiana as the least divergent at 1.21 from Altai Neandertals among modern humans (save Australian A at 1.09). They are less divergent from Denisovans than Africans, Europeans and Han but more divergent than Papuans, Australians and, with a caveat, Dai. (Dai likely absorbed a Papuan-like substrate and their Denisovan-affinity is therefore not genuine but socondary.) So, in both cases, Amerindians are closer to the two archaic species than East Asians, West Eurasians or Africans and they seem to belong to a Circumpacific substrate (Amerindians, Papuans and Australians as best modern examples thereof) that show the deepest connection to Eurasian archaics (descent and/or admixture). Prufer, Lazaridis and a number of earlier studies also concur that Amerindians (followed by Papuans) are the closest among modern humans to the homozygosity levels seen in Denisovans and Neandertals. The cline of proximity to Eurasian archaics (both in terms of allele sharing and homozygosity parity) from the highest among Amerindians/Papuans in the extreme east to the lowest among Europeans /Africans in the extreme west is a very solid cline.

We do have further genomic support for a special proximity between Amerindians and Neandertals (e.g., the basal B0006 clade of a human X chromosome gene was ascertained in Neandertals and is at highest worldwide frequencies in the New World, with the same decreasing cline toward Europe and Africa; the two Neandertals tested for blood groups turned up blood group O, which is almost fixed in South American Indians, etc.). So, the article you linked to is not aberrant. See also a PCA at http://anthropogenesis.kinshipstudies.org/2012/03/american-indians-neanderthals-and-denisovans-pca-views/.

Denisovan alleles do have a peak in South America. They haven't been detected in North America to date. I have no data on Neandertal admixture to match it up. The distribution of X chromosome hg B0006 in America shows North America as the highest but parts of South America are not too far out. http://anthropogenesis.kinshipstudies.org/2012/03/american-indians-neanderthals-and-denisovans-pca-views/. In the Prufer study referenced above Mixe has more divergence from Neandertals (and Denisovans) than Karitiana meaning that South American Indians may end up being consistently more Neandertal (and Denisovan) than North American Indians. The contrast between North and South American Indians is visible across genetic systems.

What may also transpire in the data over time is that "Amerindian admixture" mimics "Neandertal admixture" across Eurasia: all Eurasians are affected, it's higher in the east, lower in the west; higher in northern Europe, lower in southern Europe. Both seem to be giving the human genome that "eastern" pull not envisioned by classical out-of-Africa. (Incidentally, people were easy enough to accept Neandertal and Denisovan admixture but they continue to be reluctant to accept Amerindian admixture and try to explain it away as substructure among Eurasians.)

German Dziebel said...

@Hamar Fox (contd.)

An ideal out of America II model predicts that modern humans are descended from an East Eurasian hominin, went through a bottleneck (still seen in modern Amerindians) or inherited a depressed demographic condition from their source East Eurasian population, speciated into "us" in the New World and then some of the descendants migrated back to the Old World, expanded in size and either admixed with local hominins in Eurasia or did not. If they didn't admix then we would expect a progressive decline in Neandertal alleles from East Eurasians to west Eurasians to Africans. if they re-admixed with local Eurasian hominins then some populations may have higher than Amerindians concentration of archaic alleles. So far Neandertal alleles seem to decrease progressively as we depart America, while Denisovan alleles show the same pattern when it comes to West Eurasians and Africans but, in addition, they show signs of re-admixture as ancient populations went down the coast from America to the Sahul.

So, basically my hypothesis is that Neandertal and Denisovan admixture in Eurasia was brought there by Amerindians. It wasn't absorbed directly from Neandertals or Denisovans because they had mostly gone extinct by then. Only the populations that went down to the Sahul replenished their Denisovan ancestry, so Denisovans must have survived until the return of modern human to the Old World. Amerindians as a refugium population used to be a unique repository of Neandertal and Denisovan alleles after the Neandertal populations had gone extinct and prior to the back migration of now-modern humans into the Old World. This is why Europeans have less Neandertal admixture than East Asians or Amerindians, although they settled in the traditional lands of Neandertals.

Does this answer some of your questions? Happy to go back and forth on this.

Unknown said...

I decided to estimate Amerindian levels of Neanderthal admixture based on the (admittedly sketchy) data available.

Looking at Dodecad's globe4 calculator, the 1000 Genomes Mexican sample is 4.8% SSA; 4.8 i.e. Neanderthal-less, so without SSA admixture, Mexicans would have 1.22/95.2 X 100 = 1.28 Neanderthal admixture. They're also an almost symmetrical European/Amerindian mixture. We have an estimate of Iberian Neanderthal levels, but considering Iberians also have a small (1.5-3.5%) degree of SSA admixture, and since we need to exclude that admixture to reach a realistic estimate, it would be safer to use an intermediate European value instead. Let's go with 1.15%. 1.28 is the average of 1.15 and 1.41, so 1.41 is my rough estimate of Neanderthal admixture in Amerindians.

That figure matches or surpasses East Asian levels, but it may be an overestimate. But it is still unequivocally higher than European levels.



German Dziebel said...

@Tobus

" I wonder what your idea of genuine maths is - unproven and inconsistent interpretations?"

Genuine math is to respect existing math, seek alternative explanations for seeming disparity between a couple of calculations and providing better and better calculations. I know we'll get there!

German Dziebel said...

@Hamar Fox

"1.28 is the average of 1.15 and 1.41, so 1.41 is my rough estimate of Neanderthal admixture in Amerindians.

That figure matches or surpasses East Asian levels, but it may be an overestimate. But it is still unequivocally higher than European levels."

First of all, thank you for doing this. For once there's a doer out there, not just a talker. Your results are consistent with what I pulled above. It doesn't seem to be an overestimation. On the contrary, it's the elevated East Asian levels that may need an explanation. I wonder if the special proximity between East Asians and Amerindians in terms of the amount of Neandertal ancestry compared to Europeans comes from a secondary, late Pleistocene-early Holocene back migration from North America to East Asia (EDAR+, some of the Mongoloid skull features, Sinodonty dental pattern, etc.), which restored it from West Eurasian levels to more Amerindian levels. The lower amount of Denisovan alleles in East Asians (Han) compared with Amerindians may contradict that later back migration, but then Denisovan alleles haven't been detected in North America and the Holocene back migration from America to east Asia must have happened from North America.

German Dziebel said...

@Tobus

"D-stats of "D(A, B; X, Y)" are a genuine "fact" in that they consistently measure X or Y's increased affinity to A or B - you never see a case where positive means B and X have the higher affinity, it's always positive is A-X and negative is B-X... it's a mathematically valid and logically consistent pattern. What you are proposing however, is that you can tell if Aj58 is closer to MA-1 or Anzick by comparing the results of D(Y, MA-1, LB, Aj58) and D(Y, Anzick, LB, Aj58), and this is something outside the bounds of how D-stats are defined or used by any researcher. If this is a mathematically valid and logically consistent interpretation then the same interpretation needs to hold true in all other cases where we have two D-stats with only a different "B" population - essentially you are positing a rule that the difference between D(A, B1, X, Y) and D(A, B2, X, Y) represents the difference in affinity of X to B1 and X to B2. To see whether this rule might be correct we can test it against the data we already have - look through all the D-stats which vary only in the "B" population and see if your interpretation produces consistent results. Since the logical structure and maths of each D-stat we compare is identical, then we are comparing "apples with apples" as you say and the pattern you see should apply globally. If the pattern holds true in all cases then it may well be valid, despite the lack of a formal mathematical proof, but if the pattern is only sometimes there, then it's clearly a case of "pseudoscience" - just a random phenomenon with no meaning behind it.

As I've already pointed out your proposed pattern fails in the case of Gok2, where we get contradictory answers about his relative affinity to MA-1/Anzick when using Otzi compared to Sardinians..:

I'm comfortable with your writeup of the way I think we should interpret and scale f3/f4 stats across runs. Your exception to the rule, however, can be explained by reference to the well-known phenomenon whereby MA-1 is closer to East Asians than Sardinians are (see Lazaridis). One explanation is that Sardinians are heavily "Basal Eurasian" admixed through Neolithic gene flow, while MA-1 is not. Now over to your exception from my rule: Under the hypothesis that Amerindians are closer to Stuttgart than MA-1 because they are demonstrably closer to Stuttgart than ENA and ENA are demonstrably closer to MA-1 than to Stuttgart, Anzick is closer to Gok (20% Basal Eurasian) than MA-1 if compared against a more Basal Eurasian-admixed population (Sardinians) vs. a less Basal Eurasian-admixed population (Oetzi).

Tobus said...

@German:
I'm comfortable with your writeup of the way I think we should interpret and scale f3/f4 stats across runs.

The problem with your interpretation is that it assumes the sole cause of the difference is more ABAB sites (closer affinity of B and X), when the ABBA sites (closer affinity of B and Y) may be having an equal or greater effect. In the case in point you are saying the higher score has to be because Aj58 is closer to Anzick than to MA-1, but we know that LB is further from Anzick than MA-1 and hence the (ABAB - ABBA) result will be increased due to the ABBA count being smaller. If the ABAB count declined by a lesser amount then the overall score would be higher, even if both X and Y are closer to MA-1. I strongly suspect (have no doubt in fact) that the effect you are seeing is due to LB having more difference between his MA-1 and Anzick affinity than Aj58 does (consistent with LB branching before the admixture and Aj58 either in the in-between time and/or receiving additional ANE after the event).

If we want to compare the stats via your method, it's only viable if we can ensure that all of the difference in the results is due to changes in the ABAB count and not in the ABBA count - ie. that the relationship between the two B populations used only varies for the X populations but not the Y population. If this were the case then any increase in the score can only be attributed to increased alleles between B and X, and not due to decreased alleles between B and Y. Lazaridis uses comparisons like these to make the "f4 ratios" in EDFs 5 and 7. They use Bedouin as the Y population which should be roughly symmetrical in relation to MA-1 and Anzick. They also use Stuttgart as the Y in SI 14.11, although she is slightly more MA-1 than Amerindians (see EDF 5) so the results would be slightly skewed towards Karitiana - evidently not enough to reverse any of the ratios though.

So, I'm happy to accept it's possible to use the difference in two f4(A, B1/B2, X, Y) stats to infer the relative affinity of X to the B1 and B2 populations IF, AND ONLY IF, the Y population is effectively an outgroup to the two B populations being tested. This is because the greater the difference between Y's affinities to each B population, the greater impact the ABBA score will have and less reliable the inference of X's affinities to B1/B2 will be.

Under the hypothesis that Amerindians are closer to Stuttgart than MA-1

This is disproved by Lazaridis EDF 5 - Stuttgart is far from both but slightly closer to MA-1.

Now over to your exception from my rule: ....Anzick is closer to Gok (20% Basal Eurasian) than MA-1 if compared against a more Basal Eurasian-admixed population (Sardinians) vs. a less Basal Eurasian-admixed population (Oetzi).

This fits my analysis above - both Sardinians and Otzi have significantly differing affinities to MA-1 and to Anzick, hence these f4 ratios for Gok will be unreliable (any difference isn't solely due Gok's relative affinity to MA-1/Anzick).

You've noticed that the results change "if compared against a more [xxx]-admixed population", which is the issue I have with your interpretation - it varies based on the choice of Y population. I hope you can see that the only reliable and objective comparison is when we have a Y population that has an equal genetic relationship to the two B populations.

German Dziebel said...

@Tobus

"I strongly suspect (have no doubt in fact) that the effect you are seeing is due to LB having more difference between his MA-1 and Anzick affinity than Aj58 does."

So, in essence you're saying that LB is closer to MA-1 than to Anzick, while Aj58 is closer to Anzick than to MA-1? Both can be true, right? Especially since Aj58 seems to be closer to Anzick on the drift paramemter in Skoglund Fig. S7.

"consistent with LB branching before the admixture and Aj58 either in the in-between time and/or receiving additional ANE after the event."

Why would additional ANE create a difference in affinities for Anzick vs. MA-1?

"This is disproved by Lazaridis EDF 5 - Stuttgart is far from both but slightly closer to MA-1."

Yes, I can see what you're referring to. Stuttgart has Mesolithic admixture, though, and ANE is found in the Middle East from where the Neolithic wave to Europe originated, so it's unclear if Stuttgart is a tad closer to MA-1 than to Anzick because of Stuttgart's Mesolithic substrate or Neolithic superstrate.

"This fits my analysis above - both Sardinians and Otzi have significantly differing affinities to MA-1 and to Anzick, hence these f4 ratios for Gok will be unreliable (any difference isn't solely due Gok's relative affinity to MA-1/Anzick)."

They are far from being unreliable. In fact they are very telling. It's just a matter of interpreting what the scores mean, not saying that they don't mean anything because they vary depending on the choice of Y population. The scores are reflective of certain population processes, they are not just mathematical possibilities.



Tobus said...

@German:
So, in essence you're saying that LB is closer to MA-1 than to Anzick, while Aj58 is closer to Anzick than to MA-1?

Nope. I'm saying we can't tell from the D-stats results alone, although I strongly suspect both are closer to MA-1 than to Anzick.

Both can be true, right?

In theory yes, but if you look at the "Loci" column you can see there's a 312k deficit in alleles between the MA-1 calculation and the Anzick one. Since Yoruba is an outgroup, the bulk of these are alleles where both Aj53 and LB have the same allele as MA-1. If you calc out the figures, the D-stats counted Aj53 as having ~592k alleles in the Anzick calcs, and ~429k with MA-1. If we add the 312k skipped from the MA-1 count because LB has the same alleles, we have Aj53 having some 742k shared alleles with MA-1, ~150k more than with Anzick.

Why would additional ANE create a difference in affinities for Anzick vs. MA-1?

Because the closer to the ANE/Amerindian admixture event the additional admixture is, the more similar it will be to what MA-1 gave to Amerindians. Imagine two populations diverging from the MA-1 lineage at say 30kya. There's 6000 years of MA-1-specific drift that Amerindians get that these populations don't. If one of these population then receives additional ANE admixture at say 22kya, it's going to have the original DNA plus the extra drift that Amerindians got - hence it will have different proportions of MA-1/Anzick affinity than the original non-admixed population.

so it's unclear if Stuttgart is a tad closer to MA-1 than to Anzick because of Stuttgart's Mesolithic substrate or Neolithic superstrate

But what *is* clear is that she is indeed closer to MA-1.

They are far from being unreliable.

They are unreliable for the purpose you are using them for - inferring relative affinity between the X and B populations.

They can reliably tell us is which of X or Y is closer to B, but because the absolute score is dependent on both X's *and* Y's relationship to B, it's impossible to single out whether a difference between scores for B1 and B2 is due to X, Y or some combination of both. X could be closer to, further from, or equidistant with B1 than B2, and the score might go up, down or stay the same depending on what Y's relationships to B1 and B2 are.

German Dziebel said...

@Tobus

"Nope. I'm saying we can't tell from the D-stats results alone, although I strongly suspect both are closer to MA-1 than to Anzick."

It wouldn't matter to me either way, but you're running out of metrics to nail what can actually measure your ephemeral genetic distance obsession.

"Since Yoruba is an outgroup, the bulk of these are alleles where both Aj53 and LB have the same allele as MA-1."

As MA-1 AND Anzick. So, "the D-stats counted Aj53 as having ~592k alleles in the Anzick calcs, and ~429k with MA-1" should still hold.

"Because the closer to the ANE/Amerindian admixture event the additional admixture is, the more similar it will be to what MA-1 gave to Amerindians. Imagine two populations diverging from the MA-1 lineage at say 30kya. There's 6000 years of MA-1-specific drift that Amerindians get that these populations don't. If one of these population then receives additional ANE admixture at say 22kya, it's going to have the original DNA plus the extra drift that Amerindians got - hence it will have different proportions of MA-1/Anzick affinity than the original non-admixed population."

Nice thinking. Too bad, ANE is a fictitious population.

"But what *is* clear is that she is indeed closer to MA-1."

It's not clear yet but there's some indication that this may be true. Geographic proximity predicts genetic distance. It doesn't change much in the bigger picture, though. MA-1 is an Amerindian-derived population.

"They are unreliable for the purpose you are using them for - inferring relative affinity between the X and B populations. They can reliably tell us is which of X or Y is closer to B, but because the absolute score is dependent on both X's *and* Y's relationship to B, it's impossible to single out whether a difference between scores for B1 and B2 is due to X, Y or some combination of both. X could be closer to, further from, or equidistant with B1 than B2, and the score might go up, down or stay the same depending on what Y's relationships to B1 and B2 are."

This is precisely what makes them reliable. You tease out the population history meaning of these statistical relationships as you go on a case by case basis. Your "absolute score" is an unrealistic expectation.

Tobus said...

@German:
As MA-1 AND Anzick.

What?!? I'm talking about the *difference* in alleles used in the MA-1 and Anzick count - if it was alleles shared with both the the counts would be the *same*!

The fact that there's ~300k *less* alleles used in the MA-1 count means that LB and Aj58 share 300k more alleles in common with LB than they do with Anzick - there's 300k more "ABBB" alleles in the MA-1 count.

, ANE is a fictitious population

Yes, because MA-1 came from a UFO right?

This is precisely what makes them reliable. You tease out the population history meaning of these statistical relationships as you go on a case by case basis.

Only if you have some method of knowing how many of the alleles counted are due to X and Y's respective affinities to B... otherwise it's just guesswork.


German Dziebel said...

@Tobus

"LB and Aj58 share 300k more alleles in common with LB than they do with Anzick"

You mean "in common with MA-1" or indeed with LB?

"Yes, because MA-1 came from a UFO right?"

That's right! If you claim that it's not of East Asians, nor of West Eurasian, nor of Amerindian, nor of Papuan nor it's admixed between some or all of them, then you must be thinking it's of UFOs. And your logic is right. A UFO landed and contributed genes to all of these populations. Then took off. Too bad, we don't have any evidence for UFO landings. I know you'd rather do for UFOs than for Amerindians, but I can't. I need to work with real populations.

"Only if you have some method of knowing how many of the alleles counted are due to X and Y's respective affinities to B... otherwise it's just guesswork."

I call it analysis, not guesswork, but for a science denier like you all science is guesswork. Only the Scriptures are certain. What does your religion recommend as an ideal method of measuring genetic distance?

Tobus said...

@German:
You mean "in common with MA-1" or indeed with LB?

Yes, Aj53 and LB have some 300k more shared alleles in common with MA-1 than with Anzick... making both of them closer to MA-1 than to Anzick based on these D-stats.

I call it analysis, not guesswork, but for a science denier like you all science is guesswork.

Perhaps you can show me this analysis then, instead of just showboating?

If you claim that it's not of East Asians, nor of West Eurasian

MA-1 is an ancient West Eurasian genetically - that's how the West Eurasian affinity got into Amerindians... did you miss that somewhere along the line?

German Dziebel said...

@Tobus

"MA-1 is an ancient West Eurasian genetically - that's how the West Eurasian affinity got into Amerindians... did you miss that somewhere along the line?"

Your other idea, namely that MA-1 is UFO-derived, at least is not explicitly contradicted by any data. In fact I like its logical simplicity and coherence. MA-1 as a purely West Eurasian population that contributed genes to Amerindians is plain counterfactual. On all the admixture charts Amerindians and not West Eurasians are the closest to MA-1. And Amerindians don't have either the European BLUE component in ADMIXTURE runs or West Eurasian haploid lineages detected in MA-1. This suggests that MA-1 is an admixed population - ancestrally Amerindian with derived West Eurasian affinities. Predictably, all West Eurasian ancient samples, including the westernmost ones such as LB and Stuttgart, show Amerindian affinity.

"Yes, Aj53 and LB have some 300k more shared alleles in common with MA-1 than with Anzick... making both of them closer to MA-1 than to Anzick based on these D-stats."

Hmm, how do you explain then that LB doesn't show any affinity to MA-1 or Anzick, while Aj53 shows affinity to both in Skoglund Fig. 2?

terryt said...

@ German:

"MA-1 as a purely West Eurasian population that contributed genes to Amerindians is plain counterfactual".

'Counterfactual'? Your two PhDs in Dances with Indians is severely hampering your ability to consider the data objectively.

"Your other idea, namely that MA-1 is UFO-derived"

That was your idea, not Tobus'. And from your comment on the ancient American DNA post I see you still accept the Neanderthal element reached America via UFO.

"On all the admixture charts Amerindians and not West Eurasians are the closest to MA-1".

Logic has completely deserted you. One minute you're claiming MA-1 is not completely Amerindian because of admixture with a Eurasian population you otherwise claim doesn't exist ('This suggests that MA-1 is an admixed population'), and now you're denying the extreme likelihood that Amerindians are closer to MA-1 than are West Eurasians for the simple reason that the latter have become admixed with a population we actually do know existed. Why did you waste your time doing your PhDs on Dances with Indians?

German Dziebel said...

@TerryT

"'Counterfactual'? Your two PhDs in Dances with Indians is severely hampering your ability to consider the data objectively."

My background is described at http://anthropogenesis.kinshipstudies.org/sample-page/. There's no such thing as Ph.D. in dances with Indians. But the lack of a Ph.D. definitely leads to out-of-Antarctica, UFO as source of Mal'ta, Pygmies are one of the earliest populations to branch off from the human tree, Adam and Eve lived in Africa and all the other myths that you and Tobus have absorbed and propagated.

"That was your idea, not Tobus'."

No, this one is Tobus's. MA-1 is a pure East Asian population, which didn't absorb genes from any known continental population, from which everybody from West Eurasians to Amerindians originated but then it disappeared because modern East Asians or Siberians didn't really descend from it. Must be a UFO.

"One minute you're claiming MA-1 is not completely Amerindian because of admixture with a Eurasian population you otherwise claim doesn't exist ('This suggests that MA-1 is an admixed population'), and now you're denying the extreme likelihood that Amerindians are closer to MA-1 than are West Eurasians for the simple reason that the latter have become admixed with a population we actually do know existed."

You quoted me as saying "On all the admixture charts Amerindians and not West Eurasians are the closest to MA-1". You play it back to me as "now you're denying the extreme likelihood that Amerindians are closer to MA-1 than are West Eurasians." You are a rare specimen of intellectual deficiency, Terry, which will make any museum of anthropology worth a trip.

Tobus said...

@GErman:
On all the admixture charts Amerindians and not West Eurasians are the closest to MA-1

No, only on the f3 "shared drift" chart - the admixture runs and PCA show MA-1 is much closer to West Eurasians than to Amerindians.

And Amerindians don't have either the European BLUE component in ADMIXTURE runs or West Eurasian haploid lineages detected in MA-1.

Amerindians underwent a bottleneck and 15kya isolation that has "fine-tuned" their unique genetic identity so it gets isolated by ADMIXTURE easily. On the other hand West Europeans underwent extensive interbreeding which "muddies" their distinct components somewhat.

Hmm, how do you explain then that LB doesn't show any affinity to MA-1 or Anzick

I don't because it's not true. LB show increased affinity to MA-1 (we're arguing about it on a different thread - duh!), and has the same affinity to Anzick that modern Europeans have.

I suspect you're making unfounded interpretations of the data again - trying reading the details of how these data are calculated and you'll gain a better understanding of how we can and can't interpret them.

No, this one is Tobus's. MA-1 is a pure East Asian population

Please don't tell people what I said unless you're going to get it right - I've never said anything like MA-1 is a "pure East Asian population".

This shows you are just gain-saying every point I make, and not taking the time to give it serious consideration. If you actually made an effort to understand my position perhaps this series of rambling discussions would actually get somewhere.



terryt said...

"There's no such thing as Ph.D. in dances with Indians".

"German Dziebel holds a B.A. in History from St. Petersburg State University (Russia), a Ph.D. in Ethnology from the Peter the Great Museum of Anthropology and Ethnology (St. Petersburg, Russia), an M.A. in Sociology from Central European University (Warsaw, Poland), an M.A. in Anthropology and a Ph.D. in Anthropology from Stanford University (Stanford, U.S.A)".

Which of these qualifications has anything at all to do with either genetics or evolutionary biology? From what I can see your PhD might as well be in Dances with Indians. As demonstrated by:

"MA-1 is a pure East Asian population"

MA-1 is completely different from any modern East Asian population. That should be obvious, even to someone with PhD in Dances with Indians.

" Must be a UFO".

I've yet to see any hypothesis concerning how modern humans happened to evolve in America. The only explanation that seems possible is that they arrived there by UFO.

German Dziebel said...

@Tobus

"No, only on the f3 "shared drift" chart - the admixture runs and PCA show MA-1 is much closer to West Eurasians than to Amerindians."

Raghavan and Olalde both provided compelling data showing the greater proximity of Amerindians and MA-1 over others. PCAs simply show West Eurasian admixture in MA-1, which is secondary.

"Amerindians underwent a bottleneck and 15kya isolation that has "fine-tuned" their unique genetic identity so it gets isolated by ADMIXTURE easily."

Nonsense. A bottleneck after a split from east Asians, admixture with West Eurasians, then a bottleneck to become Amerindians. This is not science. Just your wishful thinking.

"I don't because it's not true. LB show increased affinity to MA-1 (we're arguing about it on a different thread - duh!), and has the same affinity to Anzick that modern Europeans have."

I referenced a specific chart in Skoglund in which LB is shown as not having the same MA-1-like/Amerindian component as Aj58. If you want to go back to Olalde, I proved that LB has greater affinity to MA-1 and Amerindians than modern Europeans. Stop recycling pseudoscience.

" I've never said anything like MA-1 is a "pure East Asian population".

ANE stands for Ancient Northeast Asian. You agree with this construct. And you believe MA-1 a pure population. So I'm right as always. It's nearly impossible to misrepresent what you're saying, Tobus, because you do all the misrepresentation in the first place.

@terryT

""German Dziebel holds a B.A. in History from St. Petersburg State University (Russia), a Ph.D. in Ethnology from the Peter the Great Museum of Anthropology and Ethnology (St. Petersburg, Russia), an M.A. in Sociology from Central European University (Warsaw, Poland), an M.A. in Anthropology and a Ph.D. in Anthropology from Stanford University (Stanford, U.S.A)".

That's more than necessary to solve modern human origins. Genetics and evolutionary anthropology are part of anthropology. And human origins is a historical, not life science.

terryt said...

"That's more than necessary to solve modern human origins. Genetics and evolutionary anthropology are part of anthropology. And human origins is a historical, not life science".

If that s the case why is it that you consistently demonstrate a complete ignorance of evolutionary biology and genetics? Even basic logic appears to elude you, as shown by:

"A bottleneck after a split from east Asians, admixture with West Eurasians, then a bottleneck to become Amerindians. This is not science. Just your wishful thinking".

And that is not what Tobus is suggesting.

Tobus said...

@German:
Raghavan and Olalde both provided compelling data showing the greater proximity of Amerindians and MA-1 over others. PCAs simply show West Eurasian admixture in MA-1, which is secondary.

Are you saying, despite data to the contrary, that MA-1 is *more* Amerindian than non-Amerindian?

A bottleneck after a split from east Asians, admixture with West Eurasians, then a bottleneck to become Amerindians

Only one bottleneck: Split from East Eurasians, admixture from ANE, then a bottleneck due to migration through Beringia.

I referenced a specific chart in Skoglund in which LB is shown as not having the same MA-1-like/Amerindian component as Aj58.

And you misconstrued it as "LB doesn't show any affinity to MA-1 or Anzick" and asked me how to explain that. I said I don't explain it because it's a misreprentation of the data. LB *does* have affinity to both MA-1 and Anzick, so your question is stupid.

ANE stands for Ancient Northeast Asian. You agree with this construct. And you believe MA-1 a pure population. So I'm right as always.

You're mixing your definitions as always. MA-1/ANE were West Eurasians located geographically in East Asia (like AG-2 and the Tarim mummies in more recent times). The way you say "pure East Asian" makes it sound like a genetic, not geographic reference. MA-1 is "purely" West Eurasian in genetic terms - something widely attested in Eastern Eurasia throughout prehistory, so no need for your UFO theory.

German Dziebel said...

@Tobus

"Are you saying, despite data to the contrary, that MA-1 is *more* Amerindian than non-Amerindian?"

One doesn't contradict the other. MA-1 is both Amerindian and non-Amerindian. That's what an admixed population is. Fits the data. Anything else doesn't. But you can keep fishing.

"Only one bottleneck: Split from East Eurasians, admixture from ANE, then a bottleneck due to migration through Beringia."

OK. It's half the your past nonsense.

"And you misconstrued it as "LB doesn't show any affinity to MA-1 or Anzick" and asked me how to explain that. I said I don't explain it because it's a misreprentation of the data. LB *does* have affinity to both MA-1 and Anzick, so your question is stupid."

Just look at the chart in Skoglund. Forget Lazaridis. I'm just checking how well your interpretation of Skoglund numbers works.

"MA-1 is "purely" West Eurasian in genetic terms - something widely attested in Eastern Eurasia throughout prehistory, so no need for your UFO theory. "

LOL. You can't have the cake and eat it, too. It's either ANE which is a genetic, not just geographic term. Or it's West Eurasian. If it's ANE, it must be from UFO. If it's West Eurasian then it's counterfactual. If you go for UFO, I'll upgrade your theory to "Science Fiction." If you go for West Eurasian, it'll remain "Pseudoscience." Your choice.

terryt said...

"MA-1 is both Amerindian and non-Amerindian. That's what an admixed population is".

You are still carefully avoiding telling us exactly what MA-1 is admixed with. Why the reluctance?

"It's either ANE which is a genetic, not just geographic term".

Correct, 'Ancient North Eurasian' is a genetic population and is certainly not 'East Eurasian', although MA-1, a member of that ANE population, lived in geographically 'East Eurasia'. Perhaps you are beginning to see the facts.

"Or it's West Eurasian".

Logic has deserted you once more. It is not an either/or situation. How do you extract 'West Eurasian' from 'ANE'?

"If it's ANE, it must be from UFO".

As things stand the population MA-1 admixed with after leaving America must have arrived by UFO. By your own logic your belief has just become 'Science Fiction'. I wish you well in your newly-chosen career.

Tobus said...

@German:
One doesn't contradict the other. MA-1 is both Amerindian and non-Amerindian.

You missed the word "more".

OK. It's half the your past nonsense.

I have no idea what that means - but where did you get the idea of two bottlenecks from? I certainly never said it and as far as I'm aware the data only points to one.

Just look at the chart in Skoglund. Forget Lazaridis.

No, I like to look at *all* the data - you should try it! You'll make fewer stupid blanket statements that way.

I'm just checking how well your interpretation of Skoglund numbers works.

Fig. 2 shows LB has a more recent MCRA with MA-1 than with Anzick, as does the Aj sample, consistent with them both having more ABBB alleles in the MA-1 D-stat than in the Anzick one. Is that the chart you mean?

You can't have the cake and eat it, too. It's either ANE which is a genetic, not just geographic term. Or it's West Eurasian.

Unless you're just using semantic definitions to artificially separate what are essentially the same thing. Consider 30-40kya ago, a Tianyuan-like population splits into two groups, one which (eventually) gives rise to modern Europeans, South Asians and Central Asians (ie a "West Eurasian" branch), the other to modern East Asians and Amerindians (ie an "East Eurasian" branch). ANE is an ancient example of the former branch, thus it's *both* ANE (a specific sub-branch) and "West Eurasian" (the parent lineage). Pretty simple really, I'm guessing you're using a different definition of "West Eurasian" that's throwing you out.

German Dziebel said...

@Tobus

"I have no idea what that means - but where did you get the idea of two bottlenecks from? I certainly never said it and as far as I'm aware the data only points to one."

All conventional theories of Amerindian origins postulate a bottleneck as Amerindians diverged from east Asians. I assume since you agree with everything that mainstream says, you agree with that. Then come West Eurasians, they admixed with a bottlenecked proto-Amerindian population. Then this West Eurasian admixed proto-Amerindian population goes through another bottleneck to become modern Amerindian.

"No, I like to look at *all* the data - you should try it! You'll make fewer stupid blanket statements that way."

No, you don't . Only if I force you to, you start looking at the data and reporting what you're finding. Scroll up the Dienekes post we are commenting on and you'll see on the left hand side Fig. 2 with a tree with attached pie charts in which MA-1 is colored all GREY, Anzick 1/2 GREY, Avj58 1/4 GREY and LB is colored all BLUE. How does it jibe with the conclusion that there are more "ABBB alleles in the MA-1 D-stat than in the Anzick one."

"Unless you're just using semantic definitions to artificially separate what are essentially the same thing. Consider 30-40kya ago, a Tianyuan-like population splits into two groups, one which (eventually) gives rise to modern Europeans, South Asians and Central Asians (ie a "West Eurasian" branch), the other to modern East Asians and Amerindians (ie an "East Eurasian" branch). ANE is an ancient example of the former branch, thus it's *both* ANE (a specific sub-branch) and "West Eurasian" (the parent lineage). Pretty simple really, I'm guessing you're using a different definition of "West Eurasian" that's throwing you out."

I can of course imagine anything you want me to. But I prefer to stick with facts. ANE is shifted toward Amerindians on all admixture charts and it doesn't cluster with any other West Eurasians. Amerindians are the most MA-1/ANE among all modern human populations. ANE is precisely a one-population taxon that's both Amerindian (part of your East Eurasian) and West Eurasian. So you can say that MA-1 is Eurasian and its ANE but you can't say that it's both West Eurasian and ANE. It's just counterfactual.

terryt said...

"Then come West Eurasians, they admixed with a bottlenecked proto-Amerindian population".

I thought you claimed to be intelligent. Surely it is obvious that West Eurasians did not mix with 'a bottlenecked proto-Amerindian population'. If anything they mixed with a proto-Amerindian population before that population had became bottlenecked, in other words before it had entered America. In reality of course the West Eurasians mixed with a population that later formed just part of the Amerindian ancestry but you are so unintelligent you begin to comprehend that scenario. As demonstrated by:

"Then this West Eurasian admixed proto-Amerindian population goes through another bottleneck to become modern Amerindian".

That statement is only necessary if you have decided in advance that humans emerged from America. Once you let that belief go it will all become obvious even to you. Dienekes' post on Y-DNA K explains it all, although I notice you are still indulging in contortions in order to make the data in that paper conform to your creationist belief.

"I can of course imagine anything you want me to".

Incorrect. You are incapable of imaging anything other than what your creationist belief demands even when it conflicts with every single fragment of data.

"But I prefer to stick with facts".

Great. So you are now going to change tack? Imagine. After all this time.

German Dziebel said...

@Tobus

"Consider 30-40kya ago, a Tianyuan-like population splits into two groups, one which (eventually) gives rise to modern Europeans, South Asians and Central Asians (ie a "West Eurasian" branch), the other to modern East Asians and Amerindians..."

Also, as Raghavan showed, Tianyuan shows the same pattern of divergence from MA-1 as other East Asians. Under your model, East Asians and West Eurasians should be equidistant from Tianyaun.

German Dziebel said...

@Tobus (contd.)

Your use of the terms "ANE" and "West Eurasian" as a daughter vs. parent population is not how the two terms are presently used. E.g., in lazaridis (http://biorxiv.org/content/early/2013/12/23/001552) ANE and West Eurasian are two daughter lineages derived from an unnamed node.

Also here http://dienekes.blogspot.com/2013/12/europeans-neolithic-farmers-mesolithic.html

Tobus said...

@German:
All conventional theories of Amerindian origins postulate a bottleneck as Amerindians diverged from east Asians.

Yes, but also after admixture from the West Eurasian lineage... one bottleneck, during the migration through Beringia, after (or while?) separating from East Asians and getting admixture from West Eurasians.

Scroll up the Dienekes post we are commenting on and you'll see on the left hand side Fig. 2 with a tree with attached pie charts in which MA-1 is colored all GREY, Anzick 1/2 GREY, Avj58 1/4 GREY and LB is colored all BLUE. How does it jibe with the conclusion that there are more "ABBB alleles in the MA-1 D-stat than in the Anzick one."

Because those circles are only indicators of admixture between the various lineages since divergence, not overall genetic affinity. You can see from the tree that both LB and Aj58 have more shared ancestry with MA-1 than they do to Anzick, hence an increased common affinity and a correspondingly higher ABBB count.

ANE is shifted toward Amerindians on all admixture charts and it doesn't cluster with any other West Eurasians

It clusters with West Eurasian lineages in TreeMix, in ADMIXTURE and in PCA (inside C/S Asians in PC1/3, only just outside Europeans in PC2). It's shifted toward Amerindians because it contributed DNA to them.

ANE is precisely a one-population taxon that's both Amerindian (part of your East Eurasian) and West Eurasian.

That would be fair enough if it were a modern population, but assigning ancestry from modern populations to an ancient context is misleading. ANE definitely shares ancestry with both modern West Eurasians and modern Amerindians, but the data shows that instead of ANE being "part-Amerindian" as you put it, Amerindians are actually "part-ANE". Most notably, ANE lacks any significant East Asian affinity, indicating he didn't receive admixture from a member of the East Eurasian lineage.

So you can say that MA-1 is Eurasian and its ANE but you can't say that it's both West Eurasian and ANE. It's just counterfactual.

If you look at the tree again you'll see that the combined "Eurasian" lineage branches off two "Western" lineages before the common Amerindian/East Asian ancestor ("Eastern" branch) diverged. What I am saying is that MA-1 (and ANE) is from a "Western" lineage, with no discernible ancestry from the "Eastern" branch.

Also, as Raghavan showed, Tianyuan shows the same pattern of divergence from MA-1 as other East Asians.

Sorry German, which part of Raghavan are you talking about? I can only see Raghavan using Tianyuan to prove that the increased Amerindian/European affinity is not due to East Asians being pulled away.

Under your model, East Asians and West Eurasians should be equidistant from Tianyaun.

Rasmussen (2014) EDF 5F and G. Note that I'm only using "Tianyuan-like" as an example, the actual common ancestor may have been in a different place with different genetics to the Tianyuan sample (although probably quite similar given his fairly even affinity to all modern Eurasians).

Your use of the terms "ANE" and "West Eurasian" as a daughter vs. parent population is not how the two terms are presently used

Yes, as I suggested above it seems your issue is purely one of definition. I think I've explained fairly clearly what I mean by "MA-1 is 'purely' West Eurasian in genetic terms" - he's derived solely from a "Western" branch of the original Eurasian population, with no ancestry from the "Eastern" branch. You shouldn't need a UFO to understand that.

terryt said...

"Yes, but also after admixture from the West Eurasian lineage... one bottleneck, during the migration through Beringia, after (or while?) separating from East Asians and getting admixture from West Eurasians".

Exactly, but German's limited intellect and understanding of genetics or evolutionary biology will prevent him from grasping that simple concept, I'm afraid.

"the data shows that instead of ANE being 'part-Amerindian' as you put it, Amerindians are actually 'part-ANE'".

I don't hold out much hope German will understand the difference. We have both tried to explain that to him many times.

German Dziebel said...

@Tobus

"Yes, but also after admixture from the West Eurasian lineage... one bottleneck, during the migration through Beringia, after (or while?) separating from East Asians and getting admixture from West Eurasians."

Separation from East Asians? Well, that's a bottleneck. Then admixture from West Eurasians. Then another bottleneck. Totally ad hoc. Here's a better model. Amerindians have always been "bottlenecked." A bottleneck is postulated for modern humans overall. That's what we see among Mid-Pleistocene hominins in Eurasia. Amerindians have maintained its signal the longest, while West Eurasians and East Asians evolved from that small deme base by population growth, admixture, etc.

"It clusters with West Eurasian lineages in TreeMix, in ADMIXTURE and in PCA (inside C/S Asians in PC1/3, only just outside Europeans in PC2). It's shifted toward Amerindians because it contributed DNA to them."

Sure, they all share derived West Eurasian affinity. But MA-1 wouldn't have been shifted toward Amerindians if it contributed genes to it. Amerindians would have been shifted toward the West Eurasian cluster.

"That would be fair enough if it were a modern population, but assigning ancestry from modern populations to an ancient context is misleading. ANE definitely shares ancestry with both modern West Eurasians and modern Amerindians, but the data shows that instead of ANE being "part-Amerindian" as you put it, Amerindians are actually "part-ANE". Most notably, ANE lacks any significant East Asian affinity, indicating he didn't receive admixture from a member of the East Eurasian lineage."

That's exactly right! Amerindians are not part of the East Eurasian lineage (one easy indicator of that is that they are closer to even westernmost West Eurasians than any of East Eurasians). They are Amerindians from whom West Eurasians and East Eurasians derived without any subsequent gene flow between these two branches. It doesn't matter if ANE is an ancient population. It's a matter of attestation. We were lucky to find a fossil, but let's not assume that we found the Ark of Noah.

"If you look at the tree again you'll see that the combined "Eurasian" lineage branches off two "Western" lineages before the common Amerindian/East Asian ancestor ("Eastern" branch) diverged. What I am saying is that MA-1 (and ANE) is from a "Western" lineage, with no discernible ancestry from the "Eastern" branch."

What tree are you looking at? MA-1 is shifted toward Amerindians. That's pretty clear from all F runs. Amerindians are not an Eastern branch, I totally agree with that.

German Dziebel said...

@Tobus (contd.)

"Sorry German, which part of Raghavan are you talking about?"

From Ragahavan's main article: "reveal significant evidence (Z > 3) for Middle Eastern, European, central Asian and south Asian populations being closer to Karitiana than to Han Chinese. Similar signals were also observed when we replaced modern-day Han Chinese with data from chromosome 21 from a 40,000-year-old east Asian individual (Tianyuan Cave, China), which has been found to be ancestral to modern-day Asians and Native Americans. Thus, if the gene flow direction was from Native Americans into western Eurasians it would have had to spread subsequently to European, Middle Eastern, south Asian and central Asian populations, including MA-1 before 24,000 years ago."

"Rasmussen (2014) EDF 5F and G. Note that I'm only using "Tianyuan-like" as an example, the actual common ancestor may have been in a different place with different genetics to the Tianyuan sample (although probably quite similar given his fairly even affinity to all modern Eurasians)."

Rasmussen EDF 5F and G show Amerindians as Amerindians, while West Eurasians and East Asians as equally Tiyuanyuan like. What this tells me is that Amerindians must have branched off prior to the Tianyuan times. In Fu et al. Tianyuan is most similar to Karitiana.

"Yes, as I suggested above it seems your issue is purely one of definition. I think I've explained fairly clearly what I mean by "MA-1 is 'purely' West Eurasian in genetic terms" - he's derived solely from a "Western" branch of the original Eurasian population, with no ancestry from the "Eastern" branch."

So, you just renamed ANE into West Eurasian because I just caught you with your logical pants down. Regardless, it still won't work because MA-1 is not just West Eurasian. That's why geneticists invented the hypothetical ANE population. It's affinity to Amerindians is clear, but there are no typical West Eurasian markers in Amerindians (hence ADMIXTURE's BLUE, mtDNA U and Y-DNA R are missing in America). Look at the Lazaridis's best tree and you'll see West Eurasian going all the way down to modern Europeans, with ANE contributing along the way to modern Europeans. West Eurasians are not contributing to Amerindians. Instead you could say that ANE and West Eurasians coalesce at a higher "northern Eurasian" node, which is different from the East Eurasian node, so MA-1 is basal to all of West Eurasians and Europeans, and then Karitiana ancestors are basal to all of those higher "northern Eurasians" plus "East Eurasians." So, lazaridis should've drawn a tree in which Karitiana is its own lineage parallel to "Basal Eurasian." A step down from the Karitiana-leading node there is a fork leading to East Eurasians and West Eurasians. I would agree with that. But that's not what you're saying, unfortunately.

terryt said...

"Amerindians have always been 'bottlenecked.'"

Do you actually understand what a population 'bottleneck' is?

"But MA-1 wouldn't have been shifted toward Amerindians if it contributed genes to it".

Please explain how you come to that conclusion. Or is it another example of your complete lack of understanding of genetics and evolutionary biology? Surely if a population has contributed genes to another it would be 'shifted towards' that population.

"Amerindians would have been shifted toward the West Eurasian cluster".

Which they are compared to East Asians.

"They are Amerindians from whom West Eurasians and East Eurasians derived without any subsequent gene flow between these two branches".

That is a completely impossible scenario. If both West Eurasians and East Eurasians are derived from Amerindians, even allowing for complete lack of subsequent gene flow between the two, they would be far more similar to each other than they are. And they would be equally distant from Amerindians.

Tobus said...

@German:
Separation from East Asians? Well, that's a bottleneck.

No, that's expansion/migration.

But MA-1 wouldn't have been shifted toward Amerindians if it contributed genes to it.

MA-1 shifted towards Amerindians, Amerindians shifted towards MA-1... same thing, different viewpoint.

Amerindians would have been shifted toward the West Eurasian cluster

As they are in the Sardinian and LB axes in Olalde EDF 5a, b, c, d, and e.

That's exactly right! Amerindians are not part of the East Eurasian lineage

That's exactly the opposite of what I'm saying.

one easy indicator of that is that they are closer to even westernmost West Eurasians than any of East Eurasians

*Sigh* Olalde EDF 5a, b and e - Amerindians are closer to East Asians than Europeans, and vice versa. See also Raghavan's TreeMix run - Amerindians and East Asians are both from the "East Eurasian" branch.

What tree are you looking at

Raghavan S11.

MA-1 is shifted toward Amerindians. That's pretty clear from all F runs.

... and Amerindians are likewise shifted towards MA-1 in all F runs.

Amerindians are not an Eastern branch, I totally agree with that.

Not sure who you think you're agreeing with, nobody else is saying that. Amerindians are most definitely an "Eastern" branch, that's what precludes ANE from being Amerindian-admixed - they lack the corresponding affinity with the rest of the Eastern branch.

Rasmussen EDF 5F and G show Amerindians as Amerindians, while West Eurasians and East Asians as equally Tiyuanyuan like. What this tells me is that Amerindians must have branched off prior to the Tianyuan times.

That's certainly one possible interpretation, another is that Amerindians contain DNA from two different post-Tianyuan populations and hence have more non-Tianyuan drift than either of them separately. The first option is rejected by the TreeMix runs which show Amerindians diverging after the West/East Eurasian splits, so the second option (also supported by ADMIXTURE, PCA and D-stats) is more likely correct.

So, you just renamed ANE into West Eurasian

No, I used "West Eurasian" to refer to three separate lineages that are genetically distinct from the "East Eurasian" lineages. ANE is only one of these "West Eurasian" lineages.

Regardless, it still won't work because MA-1 is not just West Eurasian.

On the contrary, it works remarkably well. So well in fact that's it's accepted as fact by every expert in the field. You don't get it because you insist on putting ancestry back to front - despite what you might think from the way the pretty colours are divided, MA-1's "Amerindian" component proves that Amerindians contain MA-1 DNA, but not necessarily the other way around. You seem to think it means modern DNA is "attested" at 24kya, but what it really means is that some 24kyo DNA is attested today. Time goes forwards remember.

So, lazaridis should've drawn a tree in which Karitiana is its own lineage parallel to "Basal Eurasian."

Lazaridis didn't "draw a tree" - the tree is programmatically generated to represent the most parsimonious model given the data. Saying they "should've" made changes to it so it fits your preconceived notions completely undermines the point of the tree in the first place. If it doesn't say what you want it to say, it's most likely because what you want it to say is wrong.

German Dziebel said...

@Tobus

"No, that's expansion/migration."

Pseudoscientist Tobus is back! If it was an expansion, then hetreozygosity would have gone up. Admixture would have added even more heterozygosity to proto-Amerindians. In reality Amerindians are the least heterozygous of all. So you are postulating an empty model. But you are factually wrong, too: Y-DNA and mtDNA clearly show that Amerindians did not pickup many of East Asian haplogroups. Correspondingly, in ADMIXTURE runs Amerindians don't have East Asian YELLOW just like they don't have West Eurasian BLUE. So you have to postulate a bottleneck if you want to derive Amerindians from East Asians.

"MA-1 shifted towards Amerindians, Amerindians shifted towards MA-1... same thing, different viewpoint."

Amerindians are the most MA-1-like population and the other way around. This means MA-1 is West Eurasian in only its derived alleles.

"As they are in the Sardinian and LB axes in Olalde EDF 5a, b, c, d, and e."

Wrong. We need to differentiate geographic vs. genetic aspects of the plots. All continental populations occupy their respective corners. But Amerindians are more divergent on the Karitiana-anchored axis than West Eurasians are on the Sardinian-anchored axis. This means that it's West Eurasians who are shifted toward Amerindians genetically, not the other way around.

" and Amerindians are likewise shifted towards MA-1 in all F runs."

That's the point! You can spell it right but you're still missing its meaning. MA-1 is an Amerindian population with West Eurasian admixture.

"Amerindians are most definitely an "Eastern" branch, that's what precludes ANE from being Amerindian-admixed - they lack the corresponding affinity with the rest of the Eastern branch."

Flawed logic. Amerindians are their own branch and hence, if they contributed genes to East Asians and West Eurasians, the latter two won't show as particularly closer to each other.

"The first option is rejected by the TreeMix runs which show Amerindians diverging after the West/East Eurasian splits,"

Nonsense. Amerindians are linked to both West Eurasians and East Asians. This means they predate the split between them. It's just the software is designed to show them as a mix between the two.

"so the second option (also supported by ADMIXTURE, PCA and D-stats) is more likely correct."

"Amerindians contain DNA from two different post-Tianyuan populations and hence have more non-Tianyuan drift than either of them separately..so the second option (also supported by ADMIXTURE, PCA and D-stats) is more likely correct."

It's disproven by Fu data showing Tianuayn as the closest to karitiana." ADMIXTURE runs have no support for a dula origin of Amerindians, as I have showed you multiple times.

"No, I used "West Eurasian" to refer to three separate lineages that are genetically distinct from the "East Eurasian" lineages. ANE is only one of these "West Eurasian" lineages."

That's not how people use the term "West Eurasian." It's clearly opposed to ANE in all the graphs. MA-1 is not geographically in West Eurasia and it's not genetically of West Eurasian stock. It has a secondary layer of West Eurasianness.

German Dziebel said...

@Tobus (contd.)

"On the contrary, it works remarkably well. So well in fact that's it's accepted as fact by every expert in the field. You don't get it because you insist on putting ancestry back to front - despite what you might think from the way the pretty colours are divided, MA-1's "Amerindian" component proves that Amerindians contain MA-1 DNA, but not necessarily the other way around. You seem to think it means modern DNA is "attested" at 24kya, but what it really means is that some 24kyo DNA is attested today. Time goes forwards remember.

I don't care what the "experts" say. I challenge them and I'm a better expert as my background testifies (two doctorates, two books in all the relevant fields - find another one with these credentials). By your logic, MA-1 can't be called "West Eurasian" either then. It's an East Asian population from which West Eurasians evolved. And since modern East Asians are further removed from West Eurasians than Amerindians and MA-1 had to come from somewhere, it most obviously came from America.

"Lazaridis didn't "draw a tree" - the tree is programmatically generated to represent the most parsimonious model given the data. Saying they "should've" made changes to it so it fits your preconceived notions completely undermines the point of the tree in the first place. If it doesn't say what you want it to say, it's most likely because what you want it to say is wrong."

It's very easy to show historiographically that the origin of Amerindians from Asia IS a pre-conceived idea. It's this pre-scientific idea that's guiding your "experts." I have dispelled this myth to make our ideas more in line with scientific facts and reasoning.

terryt said...

"I don't care what the 'experts' say. I challenge them and I'm a better expert as my background testifies (two doctorates, two books in all the relevant fields"

The only possible conclusion is that you are an arrogant idiot. Such as demonstrated:

"Y-DNA and mtDNA clearly show that Amerindians did not pickup many of East Asian haplogroups."

Only Y-DNA C2 along with mt-DNAs A, B, C and D. A further example:

"Amerindians are linked to both West Eurasians and East Asians. This means they predate the split between them. It's just the software is designed to show them as a mix between the two".

And admixture analyses show West Eurasians and East Asians are hardly linked at all. An impossible situation if both descend from Amerindians. But perfectly logical if Amerindians are a hybrid between West Eurasians and East Asians.

Tobus said...

@German:
If it was an expansion, then hetreozygosity would have gone up. Admixture would have added even more heterozygosity to proto-Amerindians.

So you agree - separation from East Asians does not mean a bottleneck. Your idea of two bottlenecks is erroneous, the (single) bottleneck was *after* the events you describe, when Amerindian ancestors spent some 10,000 years in extremely adverse condition in Beringia during the LGM before entering America.

Amerindians are the most MA-1-like population and the other way around.

No, not the other way round - East Asians are the most Amerindian-like population. That's the contradiction that your model fails to adequately address.

But Amerindians are more divergent on the Karitiana-anchored axis than West Eurasians are on the Sardinian-anchored axis

Consistent with Amerindians combining shared drift from two separate lineages.

Amerindians are linked to both West Eurasians and East Asians. This means they predate the split between them

I am linked to both my mother and my father - does these mean I predate them? For someone claiming intellectual superiority your logic really does suck.

ADMIXTURE runs have no support for a dula origin of Amerindians, as I have showed you multiple times.

Neither do they reject it, and as I've had to remind *you* an equal number of times, the flow of time goes forwards not backwards. These runs are consistent with modern Amerindians containing DNA from MA-1, not other way round.

That's not how people use the term "West Eurasian."

I defined my usage of the term very clearly, perhaps you'd prefer "non-East Eurasian"?

MA-1 is not geographically in West Eurasia and it's not genetically of West Eurasian stock. It has a secondary layer of West Eurasianness.

TreeMix has MA-1 being of "West Eurasian stock", ADMIXTURE shows MA-1 as "West Eurasian stock", Lazaridis S12.3 shows MA-1 as "West Eurasian stock" etc. etc.

I don't care what the "experts" say.

You should - they know a lot more about it than you and typically provide a whole raft of empirical data to back up what they are saying. Try listening to what they say, you might learn a thing a two.

I challenge them and I'm a better expert as my background testifies

We've been through your background multiple times - you're an Arts major who's only "scientific" achievement is dressing up like a Native American and traipsing round Russia for a few years - you have no training or experience in genetics and are way out of your depth here. Note also that "expert" is generally a title given to one by others, not something you just give to yourself. Good to see you have a healthy ego though :)

MA-1 can't be called "West Eurasian" either then. It's an East Asian population from which West Eurasians evolved.

From Lazaridis:
"K=2 separates African from non-African populations."
"K=3 reveals a West Eurasian ancestry component."

MA-1 is clearly "West Eurasian" from it's inception... not sure where you're getting the "East Asian" from - are you redefining "East Asian" to mean "non-African" now?

It's very easy to show historiographically that the origin of Amerindians from Asia IS a pre-conceived idea. It's this pre-scientific idea that's guiding your "experts."

No, as I've pointed out repeatedly, the data is what determines the TreeMix results, and it's this result that forms the basis of the rest of the paper.

So what motivates you to reject the TreeMix result - do *you* have a long-held preconceived idea that this result contradicts?

terryt said...

"So what motivates you to reject the TreeMix result - do *you* have a long-held preconceived idea that this result contradicts?"

German? With a long-held preconceived idea? Surely not!