PDA

View Full Version : Correlating haplogroups with ancient admixture



sparkey
31-12-13, 22:13
Eurogenes has recently explored a three-component admixture analysis based on ancient samples, in which the three components are named WHG (West European Hunter-Gatherer), EEF (Early European Farmer), and ANE (Ancient North Eurasian). This analysis can be applied to existing European populations. See here (http://bga101.blogspot.com/2013/12/eef-whg-ane-test-for-europeans.html).

Based on existing ancient samples and analysis of modern samples, there's a strong indication that Haplogroup I was important in the source population(s) of WHG, and Haplogroup R (perhaps especially R1a?) was important in the source population(s) of ANE. But do these patterns still hold? Taking the breakdown from Eurogenes and the Y-DNA frequencies from Eupedia, I get the following scatter plots and basic regression lines:

http://img809.imageshack.us/img809/5526/9esy.png

Furthermore, we can see some interesting patterns when looking at the difference between haplogroup and component frequency population-by-population:

http://img20.imageshack.us/img20/6271/wi9n.png

What to take from this? Here are some of my takes:


Haplogroup I doesn't correlate with much of anything, with its strongest positive correlation being with WHG as we may expect, but even with that one, it is a weak correlation. This could be explained by the haplogroup being widespread in Europe early, but with several later expansions into non-WHG-dominant populations.
It's also interesting to note that if we expanded this to a world admixture analysis, beyond WHG+EEF+ANE, then there would likely be a strong correlation between WHG+EEF+ANE combined and Haplogroup I.
Haplogroup R1a's regression fit with ANE shows a good but apparently nonlinear correlation. This could indicate that beyond a certain threshold of ANE percentage, there tends to have been expansions of the R1a within those populations.
Haplogroup R1a actually shows a correlation with WHG as well. This is good evidence for the R1a expansion into Europe (Corded Ware etc.) picking up assimilated hunter-gatherers as it went along, but not picking up farmers so much.
I think that the average (WHG-I) and average (ANE-R1a) are significant. The fact that ave(WHG-I) is quite positive while ave(ANE-R1a) is slightly negative is strong evidence for population absorptions magnifying the Y-lines, but not the autosomal lines, of the absorbing population at the expense of the absorbed population.

LeBrok
01-01-14, 03:04
Heck, very interesting correlations and Y frequency trends. I think they nicely indicate that hgs of I (I think mainly I2a) interacted much sooner with EEF than R1a. With R1a located more North-East clinged to hunting and gathering for much longer.
However, the huge number of I2a-Dinaric in Balkans might be skewing the results. I mean influencing them too much towards EEF. Balkans was an epicenter of EEF, and if I2a moved there relatively recently it would have picked up unproportional (to most Is) amount of EEF.
Regardless, I think in general hg I experienced farming sooner than R1a.




Haplogroup I doesn't correlate with much of anything, with its strongest positive correlation being with WHG as we may expect, but even with that one, it is a weak correlation. This could be explained by the haplogroup being widespread in Europe early, but with several later expansions into non-WHG-dominant populations.


Exactly, unlike R1a.




Haplogroup R1a actually shows a correlation with WHG as well. This is good evidence for the R1a expansion into Europe (Corded Ware etc.) picking up assimilated hunter-gatherers as it went along, but not picking up farmers so much.

But if they stayed for thousands of years in vicinity (far east europe) then just before the time of Corded Ware it might have been mostly WHG already through autosomal component. Looking at the chart, anything population above 20% R1a shows WHG from 0.3-0.5 that's very consistent, and for highest concentration of R1a at 0.4. It might mean that R1a came with 40% of their own WHG.
6155



I think that the average (WHG-I) and average (ANE-R1a) are significant. The fact that ave(WHG-I) is quite positive while ave(ANE-R1a) is slightly negative is strong evidence for population absorptions magnifying the Y-lines, but not the autosomal lines, of the absorbing population at the expense of the absorbed population.

I don't understand positive-negative "values" when describing these charts. Can you elaborate?

This is interesting too:
6156
Above 0.1 (above the noise level?), for higher concentration of R1a the ANE component is very steady around 16%. I think it indicates one time mixture of R1a source population with ANE source, or (probably most likely wide spread event of ANE expansion over huge area of Eurasia.
I wish we could compare these with R1a reach area of Pakistan, Kurdistan, India.

Great job Sparkey
Thanks

sparkey
01-01-14, 03:55
Heck, very interesting correlations and Y frequency trends. I think they nicely indicate that hgs of I (I think mainly I2a) interacted much sooner with EEF than R1a. With R1a located more North-East clinged to hunting and gathering for much longer.

Interesting way of looking at it. It would indeed seem that different groups of I carriers took up farming at different times, and hence some expanded into EEF populations, then different groups expanded into ANE populations, then others waited until late to expand.


However, the huge number of I2a-Dinaric in Balkans might be skewing the results. I mean influencing them too much towards EEF. Balkans was an epicenter of EEF, and if I2a moved there relatively recently it would have picked up unproportional (to most Is) amount of EEF.
Regardless, I think in general hg I experienced farming sooner than R1a.

Northern I2a-Din probably contributed to the I/ANE affinity as well. If we took out late expansions (say post-Bronze Age), then I bet that the I/WHG correlation strengthens somewhat. But then there would still be the EEF-linked I2-M26 subclade and some others. Not to mention that I1 had a strongly anti-EEF expansion around the same time as the I2a-Din expansion, so it may be a wash because of that.


But if they stayed for thousands of years in vicinity (far east europe) then just before the time of Corded Ware it might have been mostly WHG already through autosomal component. Looking at the chart, anything population above 20% R1a shows WHG from 0.3-0.5 that's very consistent, and for highest concentration of R1a at 0.4. It might mean that R1a came with 40% of their own WHG.

It's tough to call this hypothesis at the moment. One piece of evidence against it would be that, if we suppose WHG's source population(s) to have been strongly I dominant as is normally suggested, then the diversity patterns of practically all known I subclades are inconsistent with it. I wouldn't discount an Eastern part-WHG non-I (or extinct-I) group, though, which makes this hypothesis plausible nonetheless.


I don't understand positive-negative "values" when describing these charts. Can you elaborate?

Sorry for the jargon. It's pretty simple:

(WHG-I) = percent WHG minus percent Haplogroup I
(ANE-R1a) = percent ANE minus percent R1a
average (WHG-I) = average (WHG-I) of all the sampled populations
average (ANE-R1a) = average (ANE-R1a) of all the sampled populations

I also took averages of the absolute values of (WHG-I) and (ANE-R1a) but I don't think those say much.


This is interesting too:
6156
Above 0.1 (above the noise level?), for higher concentration of R1a the ANE component is very steady around 16%. I think it indicates one time mixture of R1a source population with ANE source, or (probably most likely wide spread event of ANE expansion over huge area of Eurasia.

I think this is a good analysis. The big vertical tower that stretches the R1a levels higher than their corresponding ANE frequencies could be post-R1a-migration-event founder effects. Looks like that tower is made up of Belorussians, Ukrainians, Lithuanians, and Czechs... perhaps a pan-Balto-Slavic founder effect reinforced by later local founder effects?

GloomyGonzales
01-01-14, 10:00
It's absolutely useless and unscientific analysis. You are trying to find correlation between autosomal data and y-dna haplos while in fact autosomal data of a person (some population) depends on much more factors (e.g. you should at least take into account as well mtDNA).

ElHorsto
01-01-14, 13:40
The (Y-DNA I, ANE) graph seems skewed by the single dot in the upper left corner. Without it it would much resemble the (Y-DNA R1a, ANE) graph I think.
The (Y-DNA R1a, ANE) graph would probably look smoother if R1b is incuded I guess.

I'm still a bit sceptic about correlating haplogroups with large autosomal components. It is much more reliable when applied to sparse, recent or exotic autosomals, like R1b-Gedrosian for example.
However, still interesting experiment.

LeBrok
02-01-14, 02:24
It's tough to call this hypothesis at the moment. One piece of evidence against it would be that, if we suppose WHG's source population(s) to have been strongly I dominant as is normally suggested, then the diversity patterns of practically all known I subclades are inconsistent with it. I wouldn't discount an Eastern part-WHG non-I (or extinct-I) group, though, which makes this hypothesis plausible nonetheless.

I don't see a problem of WHG being wholly transferable from hg I giving the long timescale from LGM or at least all neolithic. Same way that blond hair was transferred from original population to Baltic area across many haplogroups and cultures. At the time of Bronze Age Corded Ware, WHG admixture could have been spread very far into eastern Europe infiltrating R1a.
It is still plausible that transfer happened later from Bronze Age till current times. From native population to R1a/Corded Ware newcomers. 5k years it is not such short time to make it impossible. I just think it is less likely though, because newcomers were farmers and as such they outnumbered local H-Gs 10 to 1. The WHG signal would have dropped to 10%, same as it happened in south Europe in presence of EEF.
I think Corded Ware period was the time when finally H-Gs of North-East became successful farmers, both R1a and I at the same time. That's why WHG admixture survived in substantial level.
Other reason why R1a already contained WHG is that otherwise they would have to contain only EEF+ANE before Corded Ware. I'm not sure how EEF could have gotten to R1a area circumventing WHGs?




I think this is a good analysis. The big vertical tower that stretches the R1a levels higher than their corresponding ANE frequencies could be post-R1a-migration-event founder effects. Looks like that tower is made up of Belorussians, Ukrainians, Lithuanians, and Czechs... perhaps a pan-Balto-Slavic founder effect reinforced by later local founder effects?
I looked at the table again and to my understanding ANE has to be very ancient all over the europe, so old that it transcends all the haplogroups. It is very evenly spread all over europe with slight gradient from east to west, and this denotes a very ancient event. It might be so ancient that EEF had ANE at 10% already before they arrived. Therefore EEF didn't change much of ANE level in south Europe. Eastern Europe is more elevated with ANE from later migrations of steppe tribes.

ElHorsto
02-01-14, 03:05
I looked at the table again and to my understanding ANE has to be very ancient all over the europe, so old that it transcends all the haplogroups. It is very evenly spread all over europe with slight gradient from east to west, and this denotes a very ancient event. It might be so ancient that EEF had ANE at 10% already before they arrived. Therefore EEF didn't change much of ANE level in south Europe. Eastern Europe is more elevated with ANE from later migrations of steppe tribes.

But the ANE/WHG ratio is much higher in western and southern europe than in north-eastern europe. And in southern europe, only Sardinia has practically 0% ANE and Basque also has very low ANE/WHG ratio, as both are non-IE peoples, which indicates a rather late distribution of ANE at least to west and parts of south europe. I agree only for north and east europe that ANE is very old, because Motala samples were already ANE.
In particular Scotland has almost the same absolute ANE score as Estonia where it peaks (0.182 vs. 0.183), but still has much less WHG than Estonia (0.428 vs. 0.495). Since I doubt that ANE is more ancient in Scotland than in Estonia (Loshbour had no ANE) I think it came later during Bronze age to Scotland. Maybe a part ANE was already there before, but it must have been minor.

(see 'Extended Data Table3' from the paper)

Sile
02-01-14, 03:55
Eurogenes has recently come up with a three-component admixture analysis based on ancient samples, in which the three components are named WHG (West European Hunter-Gatherer), EEF (Early European Farmer), and ANE (Ancient North Eurasian). This analysis can be applied to existing European populations. See here (http://bga101.blogspot.com/2013/12/eef-whg-ane-test-for-europeans.html).

Based on existing ancient samples and analysis of modern samples, there's a strong indication that Haplogroup I was important in the source population(s) of WHG, and Haplogroup R (perhaps especially R1a?) was important in the source population(s) of ANE. But do these patterns still hold? Taking the breakdown from Eurogenes and the Y-DNA frequencies from Eupedia, I get the following scatter plots and basic regression lines:

http://img809.imageshack.us/img809/5526/9esy.png

Furthermore, we can see some interesting patterns when looking at the difference between haplogroup and component frequency population-by-population:

http://img20.imageshack.us/img20/6271/wi9n.png

What to take from this? Here are some of my takes:


Haplogroup I doesn't correlate with much of anything, with its strongest positive correlation being with WHG as we may expect, but even with that one, it is a weak correlation. This could be explained by the haplogroup being widespread in Europe early, but with several later expansions into non-WHG-dominant populations.
It's also interesting to note that if we expanded this to a world admixture analysis, beyond WHG+EEF+ANE, then there would likely be a strong correlation between WHG+EEF+ANE combined and Haplogroup I.
Haplogroup R1a's regression fit with ANE shows a good but apparently nonlinear correlation. This could indicate that beyond a certain threshold of ANE percentage, there tends to have been expansions of the R1a within those populations.
Haplogroup R1a actually shows a correlation with WHG as well. This is good evidence for the R1a expansion into Europe (Corded Ware etc.) picking up assimilated hunter-gatherers as it went along, but not picking up farmers so much.
I think that the average (WHG-I) and average (ANE-R1a) are significant. The fact that ave(WHG-I) is quite positive while ave(ANE-R1a) is slightly negative is strong evidence for population absorptions magnifying the Y-lines, but not the autosomal lines, of the absorbing population at the expense of the absorbed population.


These are not the only haplogroups which exist in the tests. The tests are meant for no haplogroup specifically. Scholars state, the numbers that range close to:
EEF 70
WHG 20
ANE 10
are from the areas of North Portugal, Northern Spain, Southern france, Northern italy, the alpine areas of swiss, austrian and southern germany..............ending on the fringes of Hungaria

I am
EEF 69.20886831
WHG 20.20271451
ANE 10.58841718

I sit between Bergamo and South France ..................which is yours

nordicwarrior
02-01-14, 04:42
...Same way that blond hair was transferred from original population to Baltic area across many haplogroups and cultures...

Surely you jest. Did you really think you could pass this conjecture off unchallenged?

On a larger note, I'm going to have to agree with El Horsto... the science linking haplogroup to autosomal in these samples seems flimsy to me at best.

And please allow me to go off on a related tangent here-- one must look no further than Youtube to see that autosomal readings (as they apply to phenotype) at this point are simply not that accurate. (I recently watched a video where the genetic test missed both eye color and hair texture from a poster who read his results on camera-- and this was from a living, breathing individual.)

**EDIT**
Youtube channel referenced above is called "Heisaswethink Heis" and the video is labelled "23 and me". Start watching at 4:39 to save time. Actually the test missed eye color, hair texture, and male pattern baldness factor (at least 3 out of 13 areas). Video can be found by searching for 23 and Me in youtube search engine.

I would think these tests would be far more accurate with "fresh" testing material vs. DNA that is hundreds/thousands of years old. What does this tell us about the validity of the readings from these ancient samples?

LeBrok
02-01-14, 05:53
But the ANE/WHG ratio is much higher in western and southern europe than in north-eastern europe. And in southern europe, only Sardinia has practically 0% ANE and Basque also has very low ANE/WHG ratio, as both are non-IE peoples, which indicates a rather late distribution of ANE at least to west and parts of south europe. I agree only for north and east europe that ANE is very old, because Motala samples were already ANE.
In particular Scotland has almost the same absolute ANE score as Estonia where it peaks (0.182 vs. 0.183), but still has much less WHG than Estonia (0.428 vs. 0.495). Since I doubt that ANE is more ancient in Scotland than in Estonia (Loshbour had no ANE) I think it came later during Bronze age to Scotland. Maybe a part ANE was already there before, but it must have been minor.

(see 'Extended Data Table3' from the paper)
6158
Looking at this quick map I made, one can argue that ANE was brought to Europe by waves of invasions from the east. However if ANE is ubiquitous and almost at same level in every person from same region that it has to be very old and had time to spread so evenly from 10 to 20 percent in every european. If it was Bronze Age recent you could see bigger disproportions between people of same country, and many people completely missing ANE. Same way you can find lactose intolerant people in central europe. So far we can't find even one person missing ANE component in Europe. It has got to be very ancient.

Sile
02-01-14, 06:48
6158
Looking at this quick map I made, one can argue that ANE was brought to Europe by waves of invasions from the east. However if ANE is ubiquitous and almost at same level in every person from same region that it has to be very old and had time to spread so evenly from 10 to 20 percent in every european. If it was Bronze Age recent you could see bigger disproportions between people of same country, and many people completely missing ANE. Same way you can find lactose intolerant people in central europe. So far we can't find even one person missing ANE component in Europe. It has got to be very ancient.

looking at your map....the magyars came from the urals........makes sense

epoch
02-01-14, 10:18
6158
Looking at this quick map I made, one can argue that ANE was brought to Europe by waves of invasions from the east. However if ANE is ubiquitous and almost at same level in every person from same region that it has to be very old and had time to spread so evenly from 10 to 20 percent in every european. If it was Bronze Age recent you could see bigger disproportions between people of same country, and many people completely missing ANE. Same way you can find lactose intolerant people in central europe. So far we can't find even one person missing ANE component in Europe. It has got to be very ancient.

Scotland is an outlier in the ANE component map and an outlier in SWG map. That is interesting. Do you also have a map for the other two components? If so, can you post them?

ElHorsto
02-01-14, 12:44
6158
Looking at this quick map I made, one can argue that ANE was brought to Europe by waves of invasions from the east.
However if ANE is ubiquitous and almost at same level in every person from same region that it has to be very old and had time to spread so evenly from 10 to 20 percent in every european.


I don't think this is significant. R1b for instance also spread quite evenly across large parts of Europe and still it happened recently.



If it was Bronze Age recent you could see bigger disproportions between people of same country, and many people completely missing ANE. Same way you can find lactose intolerant people in central europe. So far we can't find even one person missing ANE component in Europe. It has got to be very ancient.

Well, Sardinia has no ANE but it has WHG instead. Loschbour in Luxembourg had no ANE and was fully WHG. For me this is enough evidence to assume a rather late ANE dispersal to west and south. The higher ANE/WHG > 1 ratio elsewhere in southern europe can be explained by bronze age invaders, because these places 'coincidentally' are accompanied by late haplogroups like R1b and R1a for instance and speak indo-european languages. I'm correlating haplogroups here because ANE is rare in these places and R1b is recent, which makes it more significant.

Also your map shows the absolute ANE levels only, not the ANE/WHG ratio. I agree that WHG would show almost the same east-west gradient as ANE, but that's why I think the ANE/WHG ratio is so important and because of the overlap of both in NE europe.

ElHorsto
02-01-14, 12:54
6158
Looking at this quick map I made, one can argue that ANE was brought to Europe by waves of invasions from the east. However if ANE is ubiquitous and almost at same level in every person from same region that it has to be very old and had time to spread so evenly from 10 to 20 percent in every european. If it was Bronze Age recent you could see bigger disproportions between people of same country, and many people completely missing ANE. Same way you can find lactose intolerant people in central europe. So far we can't find even one person missing ANE component in Europe. It has got to be very ancient.

I could agree partially because of another reason, which is that WHG and ANE are probably very similar or just two ends within the same paleolithic component, because, if I understood correctly, the authors just defined WHG = Loschbour.

sparkey
02-01-14, 18:00
It's absolutely useless and unscientific analysis. You are trying to find correlation between autosomal data and y-dna haplos while in fact autosomal data of a person (some population) depends on much more factors (e.g. you should at least take into account as well mtDNA).

You're missing the point. I don't dispute that autosomal DNA depends on more than Y-DNA haplogroups. This is just a visualization of how well certain haplogroups correlate with certain autosomal components, with an open possibility that they may not correlate at all, or even that data may be contradictory/insufficient. It's not an attempt to break an autosomal component down into its haplogroups or anything like that.

Angela
02-01-14, 19:45
I don't think this is significant. R1b for instance also spread quite evenly across large parts of Europe and still it happened recently.



Well, Sardinia has no ANE but it has WHG instead. Loschbour in Luxembourg had no ANE and was fully WHG. For me this is enough evidence to assume a rather late ANE dispersal to west and south. The higher ANE/WHG > 1 ratio elsewhere in southern europe can be explained by bronze age invaders, because these places 'coincidentally' are accompanied by late haplogroups like R1b and R1a for instance and speak indo-european languages. I'm correlating haplogroups here because ANE is rare in these places and R1b is recent, which makes it more significant.

Also your map shows the absolute ANE levels only, not the ANE/WHG ratio. I agree that WHG would show almost the same east-west gradient as ANE, but that's why I think the ANE/WHG ratio is so important and because of the overlap of both in NE europe.

I agree that a good portion of ANE must have arrived in Europe relatively late, with the exception of some inflow into the northeast by the Mesolithic (Motala) and with historical era migrations as well, as the authors of Lazaridis et al point out.

I suppose everything is in the eye of the beholder...I see a definite east west and particularly north-east/south-west cline. I think the north east can be partially explained by my comment above. It wasn't, however, in western Europe in the late Mesolithic, going by Louschbar.

Spain and southern France are relatively light in it, and so are parts of Italy. It looks to me as if there was another gene flow starting somewhere around the Caucasus/Iran, and petering out by the time it got to central Europe.

I also wonder whether R1b is indeed the primary carrier of these Indo-European languages. Believe me, I know that's the orthodoxy on this Board, but there are some facts that don't quite fit and perhaps might argue for an R1a role in western Europe as well. For instance, if I am reading the map correctly, the ANE light areas in Italy include part of the Etruscan speaking areas, and perhaps the Raetian speaking areas? Or is that half moon area too far west for that? (I know that on the eastern part of the Alps there was some R1a geneflow, but not in the western part of the Alps.) Certainly the Basque speaking area is extremely light in it as well. Btw, these areas-Tuscany, northern Italy, and the Basque region-are very high in R1b, but did not adopt Indo European languages until very late. I'm not sure about the Basques, but there's very little R1a in those regions of Italy.

I also wonder how this fits with not only R1b spreading this component in Europe, but also how, even accepting that premise, it fits with the standard model of Indo-European bearing R1b having a staging ground in the Alps. You would think from that, that the northern areas of Italy would have a lot of ANE if R1b was the primary carrier.

Southern Italy presents another issue. There's much less R1b there, and what there is of it is heavily of the eastern variety, lots of J2a, but also quite a bit of R1a, and it has the most ANE of any area in Italy. Is that just because there is less WHG to "divide up the pie", or is it the result of Bronze Age incursions from Greece(the elite Myceneans were R1a weren't they) and perhaps the Caucasus/Iran by way of Greece carrying R1a and J2.

Just my two cents.

Sile
02-01-14, 20:22
I agree that a good portion of ANE must have arrived in Europe relatively late, with the exception of some inflow into the northeast by the Mesolithic (Motala) and with historical era migrations as well, as the authors of Lazaridis et al point out.

I suppose everything is in the eye of the beholder...I see a definite east west and particularly north-east/south-west cline. I think the north east can be partially explained by my comment above. It wasn't, however, in western Europe in the late Mesolithic, going by Louschbar.

Spain and southern France are relatively light in it, and so are parts of Italy. It looks to me as if there was another gene flow starting somewhere around the Caucasus/Iran, and petering out by the time it got to central Europe.

I also wonder whether R1b is indeed the primary carrier of these Indo-European languages. Believe me, I know that's the orthodoxy on this Board, but there are some facts that don't quite fit and perhaps might argue for an R1a role in western Europe as well. For instance, if I am reading the map correctly, the ANE light areas in Italy include part of the Etruscan speaking areas, and perhaps the Raetian speaking areas? Or is that half moon area too far west for that? (I know that on the eastern part of the Alps there was some R1a geneflow, but not in the western part of the Alps.) Certainly the Basque speaking area is extremely light in it as well. Btw, these areas-Tuscany, northern Italy, and the Basque region-are very high in R1b, but did not adopt Indo European languages until very late. I'm not sure about the Basques, but there's very little R1a in those regions of Italy.

I also wonder how this fits with not only R1b spreading this component in Europe, but also how, even accepting that premise, it fits with the standard model of Indo-European bearing R1b having a staging ground in the Alps. You would think from that, that the northern areas of Italy would have a lot of ANE if R1b was the primary carrier.

Southern Italy presents another issue. There's much less R1b there, and what there is of it is heavily of the eastern variety, lots of J2a, but also quite a bit of R1a, and it has the most ANE of any area in Italy. Is that just because there is less WHG to "divide up the pie", or is it the result of Bronze Age incursions from Greece(the elite Myceneans were R1a weren't they) and perhaps the Caucasus/Iran by way of Greece carrying R1a and J2.

Just my two cents.

The half moon area is noted as the early gallic "invasion" of northern Italy. It does sit in the "western Raeti" area ....now known as Romansch linguistic zone

Aberdeen
02-01-14, 20:28
The (Y-DNA I, ANE) graph seems skewed by the single dot in the upper left corner. Without it it would much resemble the (Y-DNA R1a, ANE) graph I think.
...........



Can you address this, sparky? If the results for I resembled those for R1a with respect to ANE, I think that would require different conclusions than one where the results for I seem fairly flat in all three graphs.

ElHorsto
02-01-14, 21:49
I agree that a good portion of ANE must have arrived in Europe relatively late, with the exception of some inflow into the northeast by the Mesolithic (Motala) and with historical era migrations as well, as the authors of Lazaridis et al point out.

I suppose everything is in the eye of the beholder...I see a definite east west and particularly north-east/south-west cline. I think the north east can be partially explained by my comment above. It wasn't, however, in western Europe in the late Mesolithic, going by Louschbar.

Spain and southern France are relatively light in it, and so are parts of Italy. It looks to me as if there was another gene flow starting somewhere around the Caucasus/Iran, and petering out by the time it got to central Europe.

I also wonder whether R1b is indeed the primary carrier of these Indo-European languages. Believe me, I know that's the orthodoxy on this Board, but there are some facts that don't quite fit and perhaps might argue for an R1a role in western Europe as well. For instance, if I am reading the map correctly, the ANE light areas in Italy include part of the Etruscan speaking areas, and perhaps the Raetian speaking areas? Or is that half moon area too far west for that? (I know that on the eastern part of the Alps there was some R1a geneflow, but not in the western part of the Alps.) Certainly the Basque speaking area is extremely light in it as well. Btw, these areas-Tuscany, northern Italy, and the Basque region-are very high in R1b, but did not adopt Indo European languages until very late. I'm not sure about the Basques, but there's very little R1a in those regions of Italy.

I also wonder how this fits with not only R1b spreading this component in Europe, but also how, even accepting that premise, it fits with the standard model of Indo-European bearing R1b having a staging ground in the Alps. You would think from that, that the northern areas of Italy would have a lot of ANE if R1b was the primary carrier.

Southern Italy presents another issue. There's much less R1b there, and what there is of it is heavily of the eastern variety, lots of J2a, but also quite a bit of R1a, and it has the most ANE of any area in Italy. Is that just because there is less WHG to "divide up the pie", or is it the result of Bronze Age incursions from Greece(the elite Myceneans were R1a weren't they) and perhaps the Caucasus/Iran by way of Greece carrying R1a and J2.

Just my two cents.

I agree. The haplo-autosomal matching possibly again shows its limits even for R1b-ANE. After all it all could be even a false positive, but rather unlikely at the moment. The authors mention a sudden mtDNA change in Central europe during Bronze-age to be indicative for the time of ANE distribution. Looking at single Y-HG or mtDNA-HG alone is certainly risky, but it fits so nicely with the IE theory here that it is tempting to make the bet at the moment. And if it does not fit exactly for some regions it always can be easily explained away by HG-drift, you know. That being said, Y-HG R1b history is by no means 100% certainly IE for me as well.

The north-east/south-west cline you mention is obviously strong for ANE, but so it is also for WHG. But my point is that the ANE/(ANE+WHG) ratio for instance is higher in Scotland, France, Orkney and most southern Europe than in Belarus,Ukraine and Estonia, despite ANE has a more eastern ancestry than WHG.

epoch
02-01-14, 21:59
Scotlands relative high ANE admixture combined with a relative high WHG admixture makes it an interesting place. The British isles were in contact with the mainland as recent as 10.000 years ago. Recently a map has been produced of so called Doggerland, the name of the middle North Sea area that was above sea level then.

http://www.nature.com/news/2008/080709/full/454151a.html?s=news_rss

In the North Sea mesolithic finds are found by fisherman, as that article states. North Scotland could have been a refuge of these people, like the Baltic.

LeBrok
02-01-14, 22:23
I don't think this is significant. R1b for instance also spread quite evenly across large parts of Europe and still it happened recently.



Well, Sardinia has no ANE but it has WHG instead. Loschbour in Luxembourg had no ANE and was fully WHG. For me this is enough evidence to assume a rather late ANE dispersal to west and south. The higher ANE/WHG > 1 ratio elsewhere in southern europe can be explained by bronze age invaders, because these places 'coincidentally' are accompanied by late haplogroups like R1b and R1a for instance and speak indo-european languages. I'm correlating haplogroups here because ANE is rare in these places and R1b is recent, which makes it more significant.

Also your map shows the absolute ANE levels only, not the ANE/WHG ratio. I agree that WHG would show almost the same east-west gradient as ANE, but that's why I think the ANE/WHG ratio is so important and because of the overlap of both in NE europe.

I just looked at Shtudgard low Asian admixtures and I have to agree that most ANE had to be introduced much later than I thought.
WHG case is easy forward. They were just pushed up north by EEF and it is easily viasable in gradual decline towards south. It corresponds exactly with history of agriculture. WHG doesn't exist in appriciable quantities outside Europe. It was only in Europe, didn't go anywhere much, only slowly declined, and was replaced by EEF and ANE. Not much mistery here. Same easy story with EEF.
However ANE doesn't have a simple story to tell. Maybe because ANE is 24 thousand years old, and EEF and WHG is 7-8k old. Today I suspect that EEF ancestors from Middle East brought about 5% of ANE with them, plus there have been a slight level of ANE in Europe already, although not very well mixed yet. The rest came in few waves with IE and Hunnic tribes. I think only multiple source can explain distribution of ANE in Europe.


I would love to see ancient Latin DNA to check who they correspond to. Anyone have Kurdish WHG reading, or rather WHG in Kurdish DNA?

LeBrok
02-01-14, 22:39
if I am reading the map correctly, the ANE light areas in Italy include part of the Etruscan speaking areas, and perhaps the Raetian speaking areas? .
You are right Angela. This map was quick (too quick) based on table from post 1, completed with my guesses and research from other sites with people posting their results and locations. It is more about ANE major trends than percentile precision. I'm sure with more data and time Maciamo will make a beautiful new map.

Overall many intriguing thoughts in your posts. And I agree that it is so hard to find IE using autosomal admixtures. If not lots of Y hg R1 and IE language it would be hard to know that they existed, lol.

Angela
02-01-14, 22:51
Scotlands relative high ANE admixture combined with a relative high WHG admixture makes it an interesting place. The British isles were in contact with the mainland as recent as 10.000 years ago. Recently a map has been produced of so called Doggerland, the name of the middle North Sea area that was above sea level then.

http://www.nature.com/news/2008/080709/full/454151a.html?s=news_rss

In the North Sea mesolithic finds are found by fisherman, as that article states. North Scotland could have been a refuge of these people, like the Baltic.

Epoch, I think I read somewhere that the samples come from Argyle. Did that area experience a higher level of gene flow from Vikings/Norwegians, and could that explain part of the ANE in them. Of course, it's also only four people, so I don't know how representative they are for all of Scotland.

Goga
02-01-14, 23:00
Anyone have Kurdish WHG reading, or rather WHG in Kurdish DNA?
This test was made for the Europeans or for the 'New World' people with the European ancestry.


My GEDmatch Eurogenes K13 Admixture Proportions:


North_Atlantic

2.30%



Baltic

3.21%



West_Med

7.71%



West_Asian

39.60%



East_Med

31.77%



Red_Sea

5.03%



South_Asian

8.50%



East_Asian

-



Siberian

1.01%



Amerindian

0.38%



Oceanian

-



Northeast_African

0.06%



Sub-Saharan

0.42%





My results of EEF/WHG/ANE:


EEF

91,55903



WHG

-0,66906



ANE

9,110035





Anatolian Kurdish average (n=10) : http://www.anthrogenica.com/showthread.php?1781-Post-Your-EEF-WHG-and-ANE-Admixture-Proportions :

EEF: 89.891057
WHG: 0.6981958
ANE: 9.4272976

epoch
02-01-14, 23:05
Epoch, I think I read somewhere that the samples come from Argyle. Did that area experience a higher level of gene flow from Vikings/Norwegians, and could that explain part of the ANE in them. Of course, it's also only four people, so I don't know how representative they are for all of Scotland.

Argyll is West-Scotland. Parts of it have been part of the Norse kingdom once. The Orkney islands - inhabitants are Orcadians - has been known to be Norwegian settling grounds, and were also part of Norway once. They even spoke a Norse language called Norn.

http://en.wikipedia.org/wiki/Norn_language
http://en.wikipedia.org/wiki/Scandinavian_Scotland


Yet they have far less.


EEF WHG ANE
0.457 0.385 0.158 -- Orcadian

I have no idea how relevant this is, mind you. I would love to see more tested against this.

Angela
02-01-14, 23:14
I just looked at Shtudgard low Asian admixtures and I have to agree that most ANE had to be introduced much later than I thought.
WHG case is easy forward. They were just pushed up north by EEF and it is easily viasable in gradual decline towards south. It corresponds exactly with history of agriculture. WHG doesn't exist in appriciable quantities outside Europe. It was only in Europe, didn't go anywhere much, only slowly declined, and was replaced by EEF and ANE. Not much mistery here. Same easy story with EEF.
However ANE doesn't have a simple story to tell. Maybe because ANE is 24 thousand years old, and EEF and WHG is 7-8k old. Today I suspect that EEF ancestors from Middle East brought about 5% of ANE with them, plus there have been a slight level of ANE in Europe already, although not very well mixed yet. The rest came in few waves with IE and Hunnic tribes. I think only multiple source can explain distribution of ANE in Europe.


I would love to see ancient Latin DNA to check who they correspond to. Anyone have Kurdish WHG reading, or rather WHG in Kurdish DNA?

I've thought that as well...that most of the WHG was pushed north and perhaps north east, despite the study that showed Mesolithic fishers inhabiting the same area as LBK people for about seventy years. That might have been an anomaly, or a late re-introduction. It just doesn't make sense to me that they would have done much more than replace their very small numbers until they adopted agriculture.

Somewhere on the net I saw a speculation that these HG's only started to expand in terms of population with Corded Ware, when they finally adopted agriculture. Might that be true, and would it explain why prior studies found that admixture between HG's and "farmers" didn't happen on a large scale for at least a thousand years? The WHG element could then have moved back into central and western Europe?

Believe me, no one wants to see ancient dna studies on "Etruscans" and "Romans" more than I do...I've been studying the ethnogenesis of Italians for years, and while some information has come out...there's still so much I don't understand. Actually, drawing conclusions from a totally sequenced Etruscan male might be easier to interpret than a "Roman" one, because depending on the area and era, the results might be quite different. Some early Empire landowner from the area of Lazio where the Flavians were born might be quite different from a landowner from the area and time of Livy...born in Cisalpine Gaul in approximately 60 B.C. although he was probably born in a Roman colony.

All I know for certain is that the phenotype snps must be extremely dominant, because there are tons of people walking around who look just like those old statues, lol.

sparkey
02-01-14, 23:25
Can you address this, sparky? If the results for I resembled those for R1a with respect to ANE, I think that would require different conclusions than one where the results for I seem fairly flat in all three graphs.

You guys aren't just seeing things. Without Sardinians, ANE has a much stronger correlation with Haplogroup I among the provided sample populations:

http://img833.imageshack.us/img833/4951/qqrx.png

That could indicate that the largest Haplogroup I expansions either expanded significantly into ANE carrying populations, or coincided with expansions of ANE. Germanic ANE/I1 and Slavic ANE/I2a-Din expansions perhaps? However, I don't think that it's entirely honest to simply drop Sardinians as insignificant outliers, as they show importantly that Haplogroup I expansions haven't necessarily coincided with ANE.

One difference I notice between (R1a, ANE) and (I, ANE) is that there is more of a sign in the (I, ANE) graph that the trend isn't increasing ever upward, but rather tailing off at the end. The Uralic speakers (Hungarians, Estonians) in particular form the right side of what looks like a Christmas tree pattern in the (I, ANE) graph. I suspect that continuing eastward into eastern Uralic populations would flatten the regression some more by completing the Christmas tree. I don't know if it would do the same for (R1a, ANE)... what are the R1a rates in eastern Uralic populations?

epoch
02-01-14, 23:34
WHG case is easy forward. They were just pushed up north by EEF and it is easily viasable in gradual decline towards south. It corresponds exactly with history of agriculture. WHG doesn't exist in appriciable quantities outside Europe. It was only in Europe, didn't go anywhere much, only slowly declined, and was replaced by EEF and ANE. Not much mistery here. Same easy story with EEF.

I understand what you mean, LeBrok. But in large parts of Europe WHG survived in somewhat of one third of the gene pool; that is no small admixture. There is archeological evidence that hunter-gatherers lived alongside the farmers for two thousand years. They started to emulate agriculture - kept pigs - and were absorbed in the newer neolithic culture that emerged. Possibly these even had characteristics derived from the hunter-gatherers turned farmers.

If we assume that farmers had a slightly larger survival rate for children they grew faster in population size. That would mean that the later farmers arrived in a spot the larger the WHG admixture, only because the farmer population did not outgrow the WHG long enough. So possibly no need for a scenario where farmers kept pushing hunter-gatherers north. The latter simply kept doing what they did as long as the farmers stuck to the very fertile bits of Europe.


http://dienekes.blogspot.nl/2013/10/ancient-central-european-mtdna-across.html

LeBrok
02-01-14, 23:48
Here is a map of WHG cline, straight South/North direction, in agreement with history of farming in Europe. Nothing much, EEF had outbred the WHG. Only Basque shows as anomaly. There will be few more but the resolution of this map is very low.

6167

epoch
03-01-14, 00:06
Here is a map of WHG cline, straight South/North direction, in agreement with history of farming in Europe. Nothing much, EEF had outbred the WHG. Only Basque shows as anomaly. There will be few more but the resolution of this map is very low.

6167

Thank you, LeBrok.

epoch
03-01-14, 01:12
You guys aren't just seeing things. Without Sardinians, ANE has a much stronger correlation with Haplogroup I among the provided sample populations:

http://img833.imageshack.us/img833/4951/qqrx.png



So Sardinia skews the relation between I and ANE. That is odd. Y-DNA I starts to look mesolithic hunter-gatherer if we consider recent published samples. Y-DNA I2 is about of a third of Sardnia's Y-DNA, yet WHG admixture in Sardinia is absent. This does look like reversed Basques, who have large hunter-gatherer admixture yet seem to have an nonhunter-gatherer Y-DNA prevalence (R1b).

Aberdeen
03-01-14, 01:25
You guys aren't just seeing things. Without Sardinians, ANE has a much stronger correlation with Haplogroup I among the provided sample populations:

http://img833.imageshack.us/img833/4951/qqrx.png

That could indicate that the largest Haplogroup I expansions either expanded significantly into ANE carrying populations, or coincided with expansions of ANE. Germanic ANE/I1 and Slavic ANE/I2a-Din expansions perhaps? However, I don't think that it's entirely honest to simply drop Sardinians as insignificant outliers, as they show importantly that Haplogroup I expansions haven't necessarily coincided with ANE.

One difference I notice between (R1a, ANE) and (I, ANE) is that there is more of a sign in the (I, ANE) graph that the trend isn't increasing ever upward, but rather tailing off at the end. The Uralic speakers (Hungarians, Estonians) in particular form the right side of what looks like a Christmas tree pattern in the (I, ANE) graph. I suspect that continuing eastward into eastern Uralic populations would flatten the regression some more by completing the Christmas tree. I don't know if it would do the same for (R1a, ANE)... what are the R1a rates in eastern Uralic populations?

The fact that Sardinia is an outlier may just mean that ANE expanded into I territory without reaching one island. The flattening as you move east may mean that there was already ANE in the east but it later expanded into the west. Would that work? I also notice that Sardinia is the only place in Europe, other than Cyprus, that's all I2 but I don't think that means anything in terms of ANE, I think it just means that I1 expanded later than I2, which the contrasting ages of I1 and I2 already indicates - probably I1 came from the late I* found in mesolithic Scandinavia.

LeBrok
03-01-14, 01:42
Ashkenazi case:
Ashkenazi_Jew EEF-0.931 WHG-0 ANE-0.069
2,000 years in Europe and they didn't pick any WHG?! Possibly they've isolated themselves very well as a community. From Old Testament we know that they were population isolates for a long time even before coming to Europe. In this case they might actually carry admixture proportions of First Neolithic Farmers from Fertile Crescent. At least how I see it, that first farmers brought first few percent ANE to Europe.

Still can't explain what happened in Sardinia to ANE. Maybe a small sample size?

Edit:
From Sparkey's table Sardinia shows ANE at 0.041, almost as much as I could expect early farmers to bring with them.
Extended Data Table 3 (from here http://arxiv.org/ftp/arxiv/papers/1312/1312.6639.pdf) shows ANE at 0.008, five times less.

kamani
03-01-14, 03:28
What I get out of this analysis is that ANE(R1a) and EEF have mixed relatively recently, as opposed to WHG who mixed with the farmers in the Neolithic. So in the late Neolithic WHG became farmers, and the only hunter-gatherers were the ANE. Obviously the lifestyles of hunter-gatherer and farmer do not co-exist well (if you're a farmer, the hunter-gatherer will come to gather your crops and be all surprised when you come to fight him). It might be that R1b+EEF pushed R1a further North-East in Eurasia, as farming expanded, then R1a made a comeback in the Metal-Ages and later with the Huns.

LeBrok
03-01-14, 03:29
Somewhere on the net I saw a speculation that these HG's only started to expand in terms of population with Corded Ware, when they finally adopted agriculture. Might that be true, and would it explain why prior studies found that admixture between HG's and "farmers" didn't happen on a large scale for at least a thousand years? The WHG element could then have moved back into central and western Europe? I would say so too. Only in Bronze age with development of bronze axe, the farmers were able to chop/de-wood heavy northern forests on industrial scale, clearing the land for white fields. Till then copper axe was too soft for the job. Also breeding new varieties of wheat, like rye, for northern climate was paramount for farmers to spread north successfully. The process was long enough to allow natives of I and R1a varieties to slowly adapt to farming, hence bigger proportions of WHG when going north.

The WHG element could then have moved back into central and western Europe?Most population movements, that we know of, happened from North to South and West to East. We should conclude that these movements elevated WHG and ANE level somewhat. Deep South of Europe is at around 10% level of WHG, so how much was there before many invasions, 5%?

Fan with math:
Let's say population of Italy is 5 million at the end of Roman Empire (at 5% of WHG). Germanic tribes of 500 thousand people (WHG at 40%), invaded Italy. I think math goes like this: (5*10+0.5*40)/10.5=8.19% of WHG for all mixed population of Italy, giving enough time for mixing.
It is not an accurate example by no means, but it illustrates how numbers of WHG can go up every time there was an invasion from North or North East, and how much effort it would take to make a dent.


Believe me, no one wants to see ancient dna studies on "Etruscans" and "Romans" more than I do...I've been studying the ethnogenesis of Italians for years, and while some information has come out...there's still so much I don't understand. Actually, drawing conclusions from a totally sequenced Etruscan male might be easier to interpret than a "Roman" one, because depending on the area and era, the results might be quite different. Some early Empire landowner from the area of Lazio where the Flavians were born might be quite different from a landowner from the area and time of Livy...born in Cisalpine Gaul in approximately 60 B.C. although he was probably born in a Roman colony.
All I know for certain is that the phenotype snps must be extremely dominant, because there are tons of people walking around who look just like those old statues, lol.
I would look for ancient "Roman" village with any DNA to sequence, before the rise of the Rome. As long as the village could be identified as Latin, and older the better. Later people started mixing a lot to be sure of results.

Angela
03-01-14, 07:12
I would say so too. Only in Bronze age with development of bronze axe, the farmers were able to chop/de-wood heavy northern forests on industrial scale, clearing the land for white fields. Till then copper axe was too soft for the job. Also breeding new varieties of wheat, like rye, for northern climate was paramount for farmers to spread north successfully. The process was long enough to allow natives of I and R1a varieties to slowly adapt to farming, hence bigger proportions of WHG when going north.
Most population movements, that we know of, happened from North to South and West to East. We should conclude that these movements elevated WHG and ANE level somewhat. Deep South of Europe is at around 10% level of WHG, so how much was there before many invasions, 5%?

Fan with math:
Let's say population of Italy is 5 million at the end of Roman Empire (at 5% of WHG). Germanic tribes of 500 thousand people (WHG at 40%), invaded Italy. I think math goes like this: (5*10+0.5*40)/10.5=8.19% of WHG for all mixed population of Italy, giving enough time for mixing.
It is not an accurate example by no means, but it illustrates how numbers of WHG can go up every time there was an invasion from North or North East, and how much effort it would take to make a dent.


I would look for ancient "Roman" village with any DNA to sequence, before the rise of the Rome. As long as the village could be identified as Latin, and older the better. Later people started mixing a lot to be sure of results.


Those figures for the Ashkenazim don't seem to support very much admixture in central or eastern Europe with "European" women. I would have a hard time believing that the WHG component wasn't present in the Rhineland at the time of the Crusades at figures close to what they are today. If, and to what extent admixture occurred, it looks as if the majority of it would have had to have been admixture with Greek and/or southernItalian/Sicilian and and/or Balkan type populations of the Iron Age.

I'm getting confused with all of the data floating around...and accepting other people's statements without checking for myself, for which I apologize. Although the numbers for ANE are all close, Sicilians don't have the highest ANE of the Italian populations. It's actually the Tuscans, which means this is one of those situations where the components are not on the predicted cline. (I'll also correct the record on the original post.)

Here are the actual proportions for the ten tested populations:
EEF/WHG/ANE:

Bergamo: .715, .177, .108
Tuscan: .746, .136, .118
Sicilian: .903, 0, .097
Sardinians: .817, .175, .008

The virtual lack of ANE in Sardinians is striking, but in keeping with all the ADMIXTURE analyses run in the past on them which show no "West Asian". At the same time, the WHG is as high as Bergamo, or about 18%. The question for me is, were there WHG's on the island who were absorbed by incoming EEF (who, of course, like all EEF people, also carry another hidden 20% or so WHG?) or did it arrive later. Jean Manco is on record as saying that there were no actual H/G settlements on the island, and the few traces found were transitory camp sites. She then proposes that the yDNA I2a and E1b1b on the island comes from young clades that arrived from the Balkans in the Copper Age. You guys would know more about the yDNA than I do, but if it did come from the Balkans from the period after the collapse of "Old Europe" and the movement into the area of Indo-European speaking steppe people, shouldn't they also have carried some ANE into Sardinia? Not even mentioning they'd bring some sort of PIE. I'm going to have to research the dates for these movements. Another possibility is that it came by way of Spain and then the southern French coast. There is another very old yDNA 12a strain in western Europe, and E1b1b was also found, of course. Also, from the northern Mediterranean into Sardinia is far easier than approaching it from the east.

As for when I2a became involved with agriculture and admixed with the EEF people, I think one good bet would be the Balkans. There was a paper that found evidence that there was admixture between the female EEF and the male HG/s within a couple of generations. Perhaps some of the I2a H/G's returned the complement and became part of the farming community. This wouldn't have impacted any H/G YDNA I further north for a long time.

I can't wait until we get the results from the Bean project. They're studying not only the Mesolithic but also the Neolithic peoples of both the Balkans and Anatolia. I'm very interested to see how different the people in Thessaly, for example, could have been from the people in NW Anatolia. I do hope they're also testing people from further east in Anatolia, near the Zagros mountains, and also down the coast a bit to northern Syria.

Perhaps we'll be surprised and there is another HG group which has to be taken into consideration. A study by a Greek academic, which was posted by Dienekes, shows no mtdna U whatsoever, not even in the Mesolithic samples. And the mtDNA she found is what we have become accustomed to call "European" or at least "Neolithic" dna: J1b, H, lots of K, and even X. What will that do to all of our calculations?

And, will we find that the EEF population didn't mostly come from the Near East or the very southern Levant, as we all have been postulating, but came from southern and south eastern European Mesolithic populations with only a small input from further south or east. Or, as another possibility, as Hunter Gatherers from Siberia to the gates of Europe weren't that different from one another, as the AF finds may show, perhaps the HGs, soon to be farmers of the European southeast, were not very different from the people in Anatolia. Who knows...I'm sure it will be a surprise.

One final thought before I fall asleep over the computer...the Bollingino study that was cited above made some pretty broad claims based on one incidence of HG's and EEF sharing time and space in central Europe. I don't know if that finding can be extrapolated to the situation over all of Europe, i.e. that there were a lot of HG's who remained basically in place, tending to their fishing and ignoring the farmers tilling the good loess soils. Actually, they were'nt totally ignoring each other. The mtDNA data doesn't show any EEF females being absorbed into the HG community, but a good number of the WHG mtDNA lines did show up among the members of the Neolithic farming community. So, it's those U mtDNA signatures that show that admixture was indeed taking place.

The big question is, what was the Y signature of these EEF men...G2a, and some I2a, doubtless, and if they were happily outbreeding all the HG males, you would have a big increase in G2a for example. So what happened? Are R1b and R1a, ANE carriers for the sake of the argument, responsible for the decrease in the EEF males lineages? They certainly didn't replace the autosomal make-up of these people however. I read somewhere that the median in Europe is over 56 for EEF.

And here I was thinking that some good ancient DNA would clear everything up? LOL

LeBrok
03-01-14, 09:10
Here are the actual proportions for the ten tested populations:
EEF/WHG/ANE:

Bergamo: .715, .177, .108
Tuscan: .746, .136, .118
Sicilian: .903, 0, .097
Sardinians: .817, .175, .008

Sicilian 0 of WHG doesn't make sense. Please give me at least 0.001 lol. What happened to WHG from Greeks, Romans, Vandal, or even some tourists from North who settled there. Are you sure sampling wasn't from one village only? (My comical rant and a question are towards scientists of the paper not to you personally)
Sardinian ANE is also suspicious, it should be at least 0.04, and Ashkenazi WHG is ridiculous, at least 0.002 should be more appropriate. We should wait for more samples.



As for when I2a became involved with agriculture and admixed with the EEF people, I think one good bet would be the Balkans. There was a paper that found evidence that there was admixture between the female EEF and the male HG/s within a couple of generations. Perhaps some of the I2a H/G's returned the complement and became part of the farming community. This wouldn't have impacted any H/G YDNA I further north for a long time.

I can't wait until we get the results from the Bean project. They're studying not only the Mesolithic but also the Neolithic peoples of both the Balkans and Anatolia. I'm very interested to see how different the people in Thessaly, for example, could have been from the people in NW Anatolia. I do hope they're also testing people from further east in Anatolia, near the Zagros mountains, and also down the coast a bit to northern Syria.

Perhaps we'll be surprised and there is another HG group which has to be taken into consideration. A study by a Greek academic, which was posted by Dienekes, shows no mtdna U whatsoever, not even in the Mesolithic samples. And the mtDNA she found is what we have become accustomed to call "European" or at least "Neolithic" dna: J1b, H, lots of K, and even X. What will that do to all of our calculations?

And, will we find that the EEF population didn't mostly come from the Near East or the very southern Levant, as we all have been postulating, but came from southern and south eastern European Mesolithic populations with only a small input from further south or east. Or, as another possibility, as Hunter Gatherers from Siberia to the gates of Europe weren't that different from one another, as the AF finds may show, perhaps the HGs, soon to be farmers of the European southeast, were not very different from the people in Anatolia. Who knows...I'm sure it will be a surprise.

One final thought before I fall asleep over the computer...the Bollingino study that was cited above made some pretty broad claims based on one incidence of HG's and EEF sharing time and space in central Europe. I don't know if that finding can be extrapolated to the situation over all of Europe, i.e. that there were a lot of HG's who remained basically in place, tending to their fishing and ignoring the farmers tilling the good loess soils. Actually, they were'nt totally ignoring each other. The mtDNA data doesn't show any EEF females being absorbed into the HG community, but a good number of the WHG mtDNA lines did show up among the members of the Neolithic farming community. So, it's those U mtDNA signatures that show that admixture was indeed taking place.

The big question is, what was the Y signature of these EEF men...G2a, and some I2a, doubtless, and if they were happily outbreeding all the HG males, you would have a big increase in G2a for example. So what happened? Are R1b and R1a, ANE carriers for the sake of the argument, responsible for the decrease in the EEF males lineages? They certainly didn't replace the autosomal make-up of these people however. I read somewhere that the median in Europe is over 56 for EEF.

And here I was thinking that some good ancient DNA would clear everything up? LOL
Denm ancestors kept moving around too much, lol.

All great inquiries Angela, I can't wait either. From now on I'm going to vote for politicians who promise more genetic tests of ancient DNA, lol.

LeBrok
03-01-14, 09:38
This test was made for the Europeans or for the 'New World' people with the European ancestry.


My GEDmatch Eurogenes K13 Admixture Proportions:


North_Atlantic
2.30%


Baltic
3.21%


West_Med
7.71%


West_Asian
39.60%


East_Med
31.77%


Red_Sea
5.03%


South_Asian
8.50%


East_Asian
-


Siberian
1.01%


Amerindian
0.38%


Oceanian
-


Northeast_African
0.06%


Sub-Saharan
0.42%




My results of EEF/WHG/ANE:


EEF
91,55903


WHG
-0,66906


ANE
9,110035




Anatolian Kurdish average (n=10) : http://www.anthrogenica.com/showthread.php?1781-Post-Your-EEF-WHG-and-ANE-Admixture-Proportions :

EEF: 89.891057
WHG: 0.6981958
ANE: 9.4272976

Wow, that's much lower than I expected for WHG, and ANE not high either. I wonder if the calculator is doing the job right. If I understood correctly, in original paper, scientists used comparable statistics of whole (all they could sequence) genomes of Stuttgart, Loschbour and MA1 (mal'ta boy?). On other hand we amateurs use K13 for calculating admixtures, and from this 3 ancestral groups. It is a shortcut and contains inaccuracies, especially for some populations. Someone correct me if I'm wrong.

Garrick
03-01-14, 14:27
I don't see a problem of WHG being wholly transferable from hg I giving the long timescale from LGM or at least all neolithic. Same way that blond hair was transferred from original population to Baltic area across many haplogroups and cultures. At the time of Bronze Age Corded Ware, WHG admixture could have been spread very far into eastern Europe infiltrating R1a.
It is still plausible that transfer happened later from Bronze Age till current times. From native population to R1a/Corded Ware newcomers. 5k years it is not such short time to make it impossible. I just think it is less likely though, because newcomers were farmers and as such they outnumbered local H-Gs 10 to 1. The WHG signal would have dropped to 10%, same as it happened in south Europe in presence of EEF.
I think Corded Ware period was the time when finally H-Gs of North-East became successful farmers, both R1a and I at the same time. That's why WHG admixture survived in substantial level.
Other reason why R1a already contained WHG is that otherwise they would have to contain only EEF+ANE before Corded Ware. I'm not sure how EEF could have gotten to R1a area circumventing WHGs?

I looked at the table again and to my understanding ANE has to be very ancient all over the europe, so old that it transcends all the haplogroups. It is very evenly spread all over europe with slight gradient from east to west, and this denotes a very ancient event. It might be so ancient that EEF had ANE at 10% already before they arrived. Therefore EEF didn't change much of ANE level in south Europe. Eastern Europe is more elevated with ANE from later migrations of steppe tribes.

LeBrok
For a comprehensive analysis it can be very simplistic that hunter-gatherers in the Europe, Balkans and beyond could not learn and take knowledge from the early farmers. On another topic, you gave an example of contact of cultures of Europeans and Indians in America, but letting the opportunity to not always be such a scenario. In other words, the native populations of hunter-gatherers (most likely I carriers maybe plus R) could also learn agriculture. It is quite possible. Take for example the Vinca culture on the territory of Serbia. It was culture of farmers and hunter-gatherers (and they also had domestic animals). In any case such a scenario would be possible, then the percentages are quite different.

ElHorsto
03-01-14, 14:39
Are R1b and R1a, ANE carriers for the sake of the argument, responsible for the decrease in the EEF males lineages? They certainly didn't replace the autosomal make-up of these people however.

I see no reason to assume that ANE carriers did not carry EEF. I personally think that many IE-tribes were slightly more EEF than anything, similar to old-europe average, except they had more ANE than WHG. I also think like you that West Asian admixture is strongly related to ANE. This makeup fits well to the north Caucasus region before the Russian colonization. (One possible wild guess could be 50%EEF+40%ANE+10%WHG, or in Admixture terms aproximately: 40% Atlantic_Med, 40% North_European, 20% West_Asian).

I don't understand how one can claim that Bronze age invaders did not alter the autosomal admixture. Check almost every important admixture run around, for instance the most recent (http://biorxiv.org/highwire/filestream/383/field_highwire_adjunct_files/1/001552-2.pdf) one from the paper. Check out the light yellow color from K=14..20 which peaks in all Caucasus peoples (Georgians, Lezgins, Chechens, Abkhazians,...) and you'll see that it occurs in all european countries except Basques and Sardinians. But this is no news actually as K12b for instance was telling basically the same for some time already.

The inconsistencies are minimal: The R1b in Basques is a big inconsistency for a long time already, but I think they have been impacted only partially by the invaders, namely by the male Y-line only but not maternal (see their high U5 frequency), such that their autosomal has been changed much less than surrounding peoples and the mothers kept speaking Basque. K12b still showed Gedrosian in them, while Lazaridis et al Admixture runs show no West Asian at all. I think this incompletely appearing West Asian signal in Basques could be the hint for their weaker yet non-zero Bronze-age impression. The Finns and Saami do have some West Asian like other Volga-Ural peoples, but it must come from an aboriginal non-IE source (from more 'pure' ANE-hunter-gathers) because they are so low in EEF.

And any West_Asian admixture was completely absent in paleolithic and neolithic finds so far, except Ötzi showed 20% in K12b where even Sardinians showed 20%. If K12b was not too oversensitive here, it might indicate that very very tiny part West_Asian was present in some neolithic peoples already.

ElHorsto
03-01-14, 14:49
That being said, I also wonder whether some IEans or other might have had 0% "North_European" of any kind and came straight from Anatolia to Italy. Their smaller ANE would show up perhaps as "West_Asian" admixture then.

ElHorsto
03-01-14, 15:17
Sicilian 0 of WHG doesn't make sense. Please give me at least 0.001 lol. What happened to WHG from Greeks, Romans, Vandal, or even some tourists from North who settled there. Are you sure sampling wasn't from one village only? (My comical rant and a question are towards scientists of the paper not to you personally)
Sardinian ANE is also suspicious, it should be at least 0.04, and Ashkenazi WHG is ridiculous, at least 0.002 should be more appropriate. We should wait for more samples.


If I understood correctly, the 0% and 100% are predefined limits such that new samples can still exceed these limits. That's why most near-easteners show > 100% EEF and < 0% WHG for instance. This means the "true" range is not exactly known and having 0% of something is more realistic then because 0% is not the border of the "true" unknown range.

If I remember correctly, the authors also state somewhere that the figures for Ashkenazi, Jews and Sicilians are not so useful.

Wilhelm
03-01-14, 18:23
You guys don't get it, the Sicilians/Maltese/Jews and also Morodvians/Finns are out of range for this test, since they have additional layers of admixture out of the EEF/WHG/ANE ancestries-. In the first group, is because of near-east ancestry, in the second because of Siberian.

Angela
03-01-14, 20:13
If I understood correctly, the 0% and 100% are predefined limits such that new samples can still exceed these limits. That's why most near-easteners show > 100% EEF and < 0% WHG for instance. This means the "true" range is not exactly known and having 0% of something is more realistic then because 0% is not the border of the "true" unknown range.

If I remember correctly, the authors also state somewhere that the figures for Ashkenazi, Jews and Sicilians are not so useful.

Thanks, I had forgotten that the Ashkenazi were included with the Sicilians and some far north-eastern Europeans in that group. I suppose therefore that we really can't draw any conclusions about "European" admixture from the results in this paper.

LeBrok
03-01-14, 20:18
If I understood correctly, the 0% and 100% are predefined limits such that new samples can still exceed these limits. That's why most near-easteners show > 100% EEF and < 0% WHG for instance. This means the "true" range is not exactly known and having 0% of something is more realistic then because 0% is not the border of the "true" unknown range.

If I remember correctly, the authors also state somewhere that the figures for Ashkenazi, Jews and Sicilians are not so useful.
Yes I noticed this, it makes sense, but the 0 was confusing for me yesterday. If they gave reading +0.005 or -0.004 that would mean they took a measurement, but 0.00 means they somehow gave up on measuring at all, I guess.

Other thing that might be hiding the complete truth of admixtures is that they (for now) selected people who might not be a pure farmers or pure hunter-gatherers. The samples are from 7-8k years ago. At this time the stuttgart farmer most likely mixed somewhat with local H-Gs and carries already some of WHG admixture. Granted that they call this admixture EEF and not NEF (near eastern farmer). This might skew the results showing less WHG though, or rather native/ancient european component level.

From the paper:

Recognizing the challenge posed by the lack of accurate surrogates for the ancestral populations, we
hypothesized that Stuttgart is a mixture of an unknown hunter-gatherer population that forms a clade
with Loschbour and an unknown Near Eastern population (NE) in proportions 1- and . W


I think I'm going to retract my thoughts that 0.05 level of ANE showing in south europe was brought by near eastern farmers. If any it was already included in EEF, because EEF was 100%=stuttgart sample, and it wouldn't show separately as ANE. Sardinian sample might be a very original of EEF with some later WHG influence.

Part (or maybe most) ANE in south might be of IE (R1b) origin. It spikes in Greece, entry point, and falls to the West as they continue the migration.

LeBrok
03-01-14, 21:05
More from paper:

and this suggests
that MA1 and Motala12 share more common drift than MA1 and Loschbour.
Scandinavian hunter-gatherers could indeed be fit if they were modeled as a mixture of Loschbour and MA1.


From article:

Scandinavian Hunter-Gatherer (SHG): this is a meta-population made up of Swedish Mesolithic and Neolithic forager samples from Motala and Gotland, respectively. It's a more easterly variant of WHG, with probable ANE admixture

Now we have to look for SHG in our genome. :)

epoch
03-01-14, 21:50
One final thought before I fall asleep over the computer...the Bollingino study that was cited above made some pretty broad claims based on one incidence of HG's and EEF sharing time and space in central Europe. I don't know if that finding can be extrapolated to the situation over all of Europe, i.e. that there were a lot of HG's who remained basically in place, tending to their fishing and ignoring the farmers tilling the good loess soils. Actually, they were'nt totally ignoring each other. The mtDNA data doesn't show any EEF females being absorbed into the HG community, but a good number of the WHG mtDNA lines did show up among the members of the Neolithic farming community. So, it's those U mtDNA signatures that show that admixture was indeed taking place.

But there is more than just paleogenetics to backup that story. There is the Dutch situation were loess grounds were used by LBK farmers in a small part, yet the hunter-gatherers maintained, even as late as 2600 BC, the Vlaardingen culture. Before, a local agricultural culture evolved which still remained largely hunter-gatherer (It might be good to remember that we all are still hunter-gatherers: We fish the seas, we eat game at Christmas) that was called Swifterband culture. Remains of burials are recovered. We also have Ertebolla culture, from about the same time.

But yes, there is also evidence that they weren't completely ignoring each other. I found a document that states there is some archeological evidence that some of the mesolithic burial rituals were copied at the fringes of the LBK culture, so the more hybrid farmers may simply not have been sampled yet. And even EEF Stuttgart had WHG admixture, as is visible in this:

http://biorxiv.org/content/biorxiv/suppl/2013/12/23/001552.DC1/001552-2.pdf



The big question is, what was the Y signature of these EEF men...G2a, and some I2a, doubtless, and if they were happily outbreeding all the HG males, you would have a big increase in G2a for example. So what happened? Are R1b and R1a, ANE carriers for the sake of the argument, responsible for the decrease in the EEF males lineages? They certainly didn't replace the autosomal make-up of these people however. I read somewhere that the median in Europe is over 56 for EEF.

And here I was thinking that some good ancient DNA would clear everything up? LOL

Y-DNA and mtDNA obviously are more prone to drifts than autosomal DNA. As a thought experiment take a mother with four healthy sons. She will not spread her mtDNA but will contribute largely to the autosomal pool, especially if the sons all take more than one wife.

LeBrok
05-01-14, 03:39
http://cdn.eupedia.com/forum/images/misc/quote_icon.png Originally Posted by ElHorsto http://cdn.eupedia.com/forum/images/buttons/viewpost-right.png (http://www.eupedia.com/forum/showthread.php?p=423394#post423394)
If I understood correctly, the 0% and 100% are predefined limits such that new samples can still exceed these limits. That's why most near-easteners show > 100% EEF and < 0% WHG for instance. This means the "true" range is not exactly known and having 0% of something is more realistic then because 0% is not the border of the "true" unknown range.

If I remember correctly, the authors also state somewhere that the figures for Ashkenazi, Jews and Sicilians are not so useful.



Yes I noticed this, it makes sense, but the 0 was confusing for me yesterday. If they gave reading +0.005 or -0.004 that would mean they took a measurement, but 0.00 means they somehow gave up on measuring at all, I guess.

My initial suspicion is confirmed in the paper:

While our three-way mixture model fits the data for most European populations, two sets of populations are poor fits. First, Sicilians, Maltese, and Ashkenazi Jews have EEF estimates beyond the 0-100% interval (SI13) and they cannot be jointly fit with other Europeans (SI12). These populations may have more Near Eastern ancestry than can be explained via EEF admixture (SI13), an inference that is also suggested by the fact that they fall in the gap between European and Near Eastern populations in the PCA of Fig. 1B.

Goga
09-01-14, 17:39
Wow, that's much lower than I expected for WHG, and ANE not high either. I wonder if the calculator is doing the job right. If I understood correctly, in original paper, scientists used comparable statistics of whole (all they could sequence) genomes of Stuttgart, Loschbour and MA1 (mal'ta boy?). On other hand we amateurs use K13 for calculating admixtures, and from this 3 ancestral groups. It is a shortcut and contains inaccuracies, especially for some populations. Someone correct me if I'm wrong.According to a modified calculator for West Asians, I've 17.606% of ANE and according to DODECAD K12b calculator I've got only 4.87% of North_European component, 28.04% of Gedrosia and 37.76% of Caucasus component. So ANE in me is for the greatest part correlated with Gedrosia and Caucasus component, I guess.

pyromatic
11-01-14, 01:50
probably I1 came from the late I* found in mesolithic Scandinavia.

There was no I* found in mesolithic Sweden. The two samples that weren't I2 were simply I. They did not test any of the SNPs phyloequivalent to M253, none of the SNPs that define I2, I2a, I2b, or I2c; nor did they test all of the I-defining SNPs.

The I2-M423 individuals were subsequently found to be derived for about half and ancestral for the other half of the SNPs phyloequivalent to M423, putting them on a parallel branch to that from which all modern I2-M423 individuals descend, one that has subsequently gone extinct; or they are members of the population that was ancestral to all modern I2-M423. Additionally, the ancient I2-M423 samples shared the same derived/ancestral state on the phyloequivalent SNPs that could be read.

Motala2 and Motala9 are just I. They are not true generics, and, based on the results of the extended testing on the I2 individuals, I suspect Motala2 and Motala9 would similarly be I2.

I1 and pre-I1 are yet to be found in ancient DNA.