PDA

View Full Version : "Kennewick-like" analysis of Allentoft's samples from Europe - question on prospects



Tomenable
07-09-2015, 08:50 AM
I'm curious about European samples (especially those from R1a-rich populations), but of course this can apply to any sample. In case of Kennewick Man, the recently published study has established, that his most direct modern descendants, are people of the Colville tribe. If such an accurate ethnic association was possible for Kennewick Man, why should it be any harder to establish e.g. whose ancestor was each of RISE samples?

https://en.wikipedia.org/wiki/Colville_tribe

Here is what authors of the study on Kennewick Man wrote about this:

http://www.nature.com/nature/journal/vnfv/ncurrent/full/nature14625.html


Among the Native American groups for whom genome-wide data are available for comparison, several seem to be descended from a population closely related to that of Kennewick Man, including the Confederated Tribes of the Colville Reservation (Colville), one of the five tribes claiming Kennewick Man. (...) Kennewick Man has ancestry proportions most similar to those of other Northern Native Americans (Fig. 1c, Supplementary Information 7), especially the Colville, Ojibwa, and Algonquin. (...) among the groups for which we have sufficient genomic data we find that the Colville, one of the Native American groups claiming Kennewick Man as ancestral, show close affinities to that individual or at least to the population to which he belonged. (...) These findings can be explained as: (1) the Colville individuals are direct descendants of the population to which Kennewick Man belonged, but subsequently received some relatively minor gene flow from other American populations within the last ~8.5 thousand years, in agreement with our findings above; (2) the Colville individuals descend from a population that ~8.5 thousand years was slightly diverged from the population which Kennewick Man belonged or (3) a combination of both.

So what I would like to see when it comes to East-Central European samples from Allentoft (as well as other ones), is something like this:


Among the European groups for whom genome-wide data are available for comparison, several seem to be descended from a population closely related to that of RISE [number] including the [ethnic group name], one of the five ethnic groups claiming RISE [number]. (...) RISE [number] has ancestry proportions most similar to those of other Europeans (Fig. 1c, Supplementary Information 7), especially the [ethnic group name], [ethnic group name], and [ethnic group name]. (...) among the groups for which we have sufficient genomic data we find that the [ethnic group name], one of the European groups claiming RISE [number] as ancestral, show close affinities to that individual or at least to the population to which he belonged. (...) These findings can be explained as: (1) the [ethnic group name] individuals are direct descendants of the population to which RISE [number] belonged, but subsequently received some relatively minor gene flow from other European populations within the last ~[number] thousand years, in agreement with our findings above; (2) the [ethnic group name] individuals descend from a population that ~[number] thousand years was slightly diverged from the population which RISE [number] belonged or (3) a combination of both.

What do you think, when will we see results like this ??? If ever.

Krefter
07-09-2015, 09:13 AM
That's a good idea. Some of it is already known. Especially via-Y DNA. We can see lots of modern regional Y DNA variation form by 2000 BC. There's a close paternal relation between Bell beaker and West Europeans, Corded Ware and East Europeans, and LN/BA Scandinavians with Germanic-speakers. The R1b-U152 found in Beaker by the Alps is probably from the U152 people who went into Italy.

Generalissimo
07-09-2015, 09:55 AM
What do you think, when will we see results like this ??? If ever.

It's not that difficult. It just takes a bit of fumbling around with qpAdm and cross checking with uniparental markers. For instance, these are solid fits based on high quality transversion sites and they correlate well with Y-DNA results.

Lithuanian
RISE_baCorded_Ware 0.974
EN_MN_LN_European 0.014
Nganasan 0.012

chisq 0.024 tail prob 0.987829

Icelandic
RISE_baYamnaya 0.530
EN_MN_LN_European 0.439
Nganasan 0.031

chisq 0.898 tail prob 0.638166

But the problem is that all Europeans from across the north of the continent share a lot of ancestry with Corded Ware and Yamnaya, because the Early Bronze Age population movements from the steppe into Northern Europe were so massive. This was followed by a lot of local migrations and mixing. It's not like North America, where there's probably still a lot of structure dating back 8,000 years.

Shaikorth
07-09-2015, 10:30 AM
It's not that difficult. It just takes a bit of fumbling around with qpAdm and cross checking with uniparental markers. For instance, these are solid fits based on high quality transversion sites and they correlate well with Y-DNA results.

Lithuanian
RISE_baCorded_Ware 0.974
EN_MN_LN_European 0.014
Nganasan 0.012

chisq 0.024 tail prob 0.987829

Icelandic
RISE_baYamnaya 0.530
EN_MN_LN_European 0.439
Nganasan 0.031

chisq 0.898 tail prob 0.638166



Does using Corded Ware instead of Yamnaya improve the Icelandic fit? That was the case with Lithuanians, so in theory it should work for others too.

parastais
07-09-2015, 11:16 AM
Does using Corded Ware instead of Yamnaya improve the Icelandic fit? That was the case with Lithuanians, so in theory it should work for others too.
Additionally, is there big if any difference if Estonians are checked in place of Lithuanians? There should be some, just wonder if noticeable.

Tomenable
07-09-2015, 11:33 AM
But the problem is that all Europeans from across the north of the continent share a lot of ancestry with Corded Ware and Yamnaya, because the Early Bronze Age population movements from the steppe into Northern Europe were so massive. This was followed by a lot of local migrations and mixing. It's not like North America, where there's probably still a lot of structure dating back 8,000 years.

I think that in North America there have also been a lot of local migrations and mixing. Maybe some relatively massive migrations too.

Yet it was still possible to determine which of the modern populations is to the greatest extent descended from Kennewick Man, or from the clan or tribe to which he belonged. Obviously most if not all of native North Americans have some ancestry from that guy, but certain tribes have significantly more than other tribes. Let's keep in mind that the Colville tribes, who are perhaps the most direct descendants of Kennewick Man's prehistoric "clan", number only around 8700 people, according to this link:

https://en.wikipedia.org/wiki/Colville_Indian_Reservation

8700 is but a tiny proportion of all Amerindians in North America (there are at least several million people with some Native ancestry living in the USA and in Canada). So they found out that these 8700 people are to a larger extent descended from Kennewick Man than perhaps any other group of ~9000 Amerindian people. It would be like saying, that people in - for example - West Midlands, are significantly more descended from Hinxton-1, than people in any other part of the UK, let alone of Europe.

Generalissimo
07-09-2015, 11:59 AM
Does using Corded Ware instead of Yamnaya improve the Icelandic fit? That was the case with Lithuanians, so in theory it should work for others too.

Sure, because some of the CWC samples from RISE are almost modern North Euros.

Icelandic
RISE_baCorded_Ware 0.771
EN_MN_LN_European 0.205
Nganasan 0.024

chisq 0.239 tail prob 0.887253


Additionally, is there big if any difference if Estonians are checked in place of Lithuanians? There should be some, just wonder if noticeable.

The Nganasan goes up, for one.

Estonian
RISE_baCorded_Ware 0.831
EN_MN_LN_European 0.111
Nganasan 0.059

chisq 0.232 tail prob 0.890445


I'm fairly sure that in North America there have also been a lot of local migrations and mixing. Maybe some relatively massive migrations too.

I really think that's unlikely becuse North America is huge and most of it was still inhabited by hunter-gatherers when Europeans got there. So it was probably easier for ancient substructures to survive there until the present. In contrast, Northern Europe was homogenized a couple times over by massive long range population movements.

Megalophias
07-09-2015, 11:59 AM
There were certainly major language spreads in North America. How much that translates into genetic spread isn't really known yet, I don't think.

Almost all of Canada is covered by just three language families: Algonquian, Athapaskan, and Inuit. These all extend into the US as well. The similarity in place names thousands of miles apart is striking. The Algonquian languages are spoken from the Rockies to the Atlantic, but the other members of the larger Algic family to which the Algonquian branch belongs are spoken in California. Athapaskan speakers are most concentrated in northwestern Canada and interior Alaska, but patches extend south to California and of course Arizona and New Mexico. The other branches of the Na-Dene family to which they belong are from southern Alaska. Closely-related Inuit dialects stretch all the way from Alaska to Greenland, and basically form a branch of the Yupik languages, spoken in Alaska and even across the Bering Strait in Russia; together they are more distantly related to Aleutian.

Interestingly, none of these expansions was agricultural (some eastern Algonquians and the Navajo did take up farming later). Siouan and Iroquoian people were farmers and do extend into parts of southern Canada as well, but mainly are found in the US. Only west of the Rockies in BC is there considerable linguistic diversity (5 families and isolates besides Na-Dene).


8700 is but a tiny proportion of all Amerindians in North America (there are at least several million people with some Native American ancestry living in the USA and in Canada). So they found out that these 8700 people are perhaps to a larger extent descended from Kennewick Man than any other group of ~9000 people in North America. It would be like saying, that for example people in - for example - West Midlands, are significantly more descended from Hinxton-1, than people in any other part of the UK, let alone Europe.

This is not even remotely true unfortunately. They only had data from a handful of groups north of Mexico, most of them very far away from Kennewick Man - from the northern part of the Pacific coast in northern BC and southern Alaska; from the Great Lakes region; from the boreal forests of northern Canada; from the high arctic; and from Arizona. There were also two other groups from the Plateau region, further to the north in Canada, who are linguistically and genetically related to the Colville people; they were not quite as close to Kennewick Man, though.

It was a totally inadequate sampling, which the authors made clear both in the paper itself and in interviews with the media.

Shaikorth
07-09-2015, 12:12 PM
What's in the EN_MN_LN sample? Mainly interested in the LN part.

Generalissimo
07-09-2015, 12:14 PM
What's in the EN_MN_LN sample? Mainly interested in the LN part.

CO1 and Iceman. Also Gok2, but that's borderline MN/LN.

Tomenable
07-09-2015, 01:31 PM
I really think that's unlikely becuse North America is huge and most of it was still inhabited by hunter-gatherers when Europeans got there. So it was probably easier for ancient substructures to survive there until the present. In contrast, Northern Europe was homogenized a couple times over by massive long range population movements.

When Europe was inhabited by hunter-gatherers, there was actually more of homogeneity, less of genetic diversity and more of IBD sharing than today, at least according to prof. Pinhasi: http://www.dienekes.blogspot.com/2015/03/ice-age-europeans-on-brink-of-extinction.html


"In some cases, small bands of potentially as few as 20 to 30 people could have been moving over very large areas, over the whole of Europe as a single territory, according to Professor Ron Pinhasi, principal investigator on the EU-funded ADNABIOARC project. Prof. Pinhasi’s team has found that the genomes sequenced from hunter-gatherers from Hungary and Switzerland between 14 000 to 7 500 years ago are very close to specimens from Denmark or Sweden from the same period. These findings suggest that genetic diversity between inhabitants of most of western and central Europe after the ice age was very limited, indicating a major demographic bottleneck triggered by human isolation and extinction during the ice age."

Generalissimo
07-09-2015, 01:39 PM
When Europe was inhabited by hunter-gatherers, there was actually more of homogeneity, less of genetic diversity and more of IBD sharing than today, at least according to prof. Pinhasi: http://www.dienekes.blogspot.com/2015/03/ice-age-europeans-on-brink-of-extinction.html

Parts of Europe, yes, but as we've learned recently in Sweden the hunter-gatherers were somewhat different, and in Russia they were very different.

Tomenable
07-09-2015, 01:54 PM
There were also two other groups from the Plateau region, further to the north in Canada, who are linguistically and genetically related to the Colville people; they were not quite as close to Kennewick Man, though.

Generally, IIRC, speakers of Salishan and Algonquian languages were found to be close to Kennewick Man, but the closest of them all seem to be the Colville people - at least out of populations analysed in that study (though as you wrote, their sampling was probably inadequate).

Tomenable
07-09-2015, 02:11 PM
When Europe was inhabited by hunter-gatherers, there was actually more of homogeneity, less of genetic diversity and more of IBD sharing than today, at least according to prof. Pinhasi

Parts of Europe, yes, but as we've learned recently in Sweden the hunter-gatherers were somewhat different, and in Russia they were very different.

Yes indeed.

When it comes to those in Russia - the Samara hunter and the Karelia hunter were very similar to each other, despite a dozen hundred or so of kilometers (and a few hundred years) separating them from each other. So it seems that not all populations across the continent were homogeneous, but there were several large areas, each with a relatively homogeneous population of its own. That could also be the case in North America. Hunter-gatherers tend to be highly mobile groups, but small in numbers relative to size of their habitats (very sparse population density).

Shaikorth
07-09-2015, 02:13 PM
Generally, IIRC, speakers of Salishan and Algonquian languages were found to be close to Kennewick Man, but the closest of them all seem to be the Colville people - at least out of populations analysed in that study (though as you wrote, their sampling was probably inadequate).

Colville seemed to belong in a bigger group of natives that were close to Kennewick, but I don't know if they could be singled out. The other tribes claiming Kennewick are from the same area, so they'd probably have gotten similar results if they had participated in the study. The scientists only managed to convince Colville.

If anything the study shows that Native groups should have nothing to fear when someone asks them for samples to be used in comparisons like this. Ancient native americans aren't going to turn out anything but like their modern counterparts.

Tomenable
07-10-2015, 10:33 AM
Which of modern groups show close affinities to Allentoft's Unetice samples (RISE 577, 586, 109, 139, 145, 150, 154, 431)?

Tolan
07-10-2015, 04:07 PM
Which of modern groups show close affinities to Allentoft's Unetice samples (RISE 577, 586, 109, 139, 145, 150, 154, 431)?

There are some differences between Unetice samples.
In general, it's the Northern Europe:
Oracle 4 with Eurogenes K15:
Rise 577:
1 NW_France @ 4,976696
2 Southwest_English @ 6,367392
3 South_Dutch @ 7,551647
4 Southeast_English @ 7,764888
5 Irish @ 8,971978
6 West-Scottish @ 9,924036
7 North_German @ 9,994099
8 French @ 10,558937
9 North_Dutch @ 10,956637
10 Danish @ 11,057187

Rise150:
1 Danish @ 7,437832
2 Norwegian @ 8,275071
3 West-Norwegian @ 8,404675
4 North_Dutch @ 8,423931
5 West-Scottish @ 9,140144
6 Swedish @ 9,232579
7 Orcadian @ 9,61955
8 North_German @ 9,91944
9 Irish @ 10,070117
10 Southeast_English @ 11,259533

Tomenable
07-10-2015, 05:20 PM
Thank you! What about Allentoft's Central European CW samples (such as RISE 1, 431, 434, 435, 436, 446) as well as Corded Ware samples from previous studies (such as those from Eulau, Esperstedt, Jagodno, Oberwiederstedt, Karsdorf, Quedlinburg, Spreitenbach-Moosweg) and Urnfield samples (such as Haak's sample HAL 36 from Halberstadt, and samples from Lichtenstein Cave)? Is autosomal data available for them?

Is there also autosomal data for U48a, U54, U110 and U131 Medieval samples from Usedom (Freder 2010)?

I'm also curious about autosomal data for RISE 598 from Lithuania (I've already seen results for RISE 00 from Estonia).

parastais
07-10-2015, 08:55 PM
Estonian
RISE_baCorded_Ware 0.831
EN_MN_LN_European 0.111
Nganasan 0.059

chisq 0.232 tail prob 0.890445

What if Lithuanians are again used, but instead of Baltic Corded Ware Sintashta is used? Since they look similar in K8, I wonder how much worse would model get?

Tomenable
07-10-2015, 09:08 PM
Such an autosomal comparison from Davidski's blog:

http://eurogenes.blogspot.com/2015/06/k8-results-for-selected-allentoft-et-al.html

I've added some descriptions so it is more transparent:

http://oi61.tinypic.com/33bflzn.jpg

http://oi61.tinypic.com/33bflzn.jpg

Tomenable
07-10-2015, 10:51 PM
^ Here is how user Arvistro from Eupedia forum commented on the above posted PCA:


Geographically Yamna in this PCA looks located to support Kortlandt's Indo-Uralic theory:

The Dutch linguist Frederik Kortlandt (https://en.wikipedia.org/wiki/Frederik_Kortlandt) supports a model of Indo-Uralic in which the original Indo-Uralic speakers lived north of the Caspian Sea, and the Proto-Indo-European speakers began as a group that branched off westward from there to come into geographic proximity with the Northwest Caucasian languages (https://en.wikipedia.org/wiki/Northwest_Caucasian_languages), absorbing a Northwest Caucasian lexical blending before moving farther westward to a region north of the Black Sea where their language settled into canonical Proto-Indo-European (2002:1). Allan Bomhard (https://en.wikipedia.org/wiki/Allan_R._Bomhard) suggests a similar schema in Indo-European and the Nostratic Hypothesis (1996).

Generalissimo
07-11-2015, 02:47 AM
What if Lithuanians are again used, but instead of Baltic Corded Ware Sintashta is used? Since they look similar in K8, I wonder how much worse would model get?

It doesn't work, but 95% Sintashta/5% Nganasan is an OK fit. The high Nganasan is compensating for something, so I would have to try different reference groups to bring it down.

Shaikorth
07-11-2015, 06:03 AM
Nganasan should perhaps be replaced with EHG and Han/She/Dai since that's what they basically are.

Tomenable
07-11-2015, 11:03 AM
There are some differences between Unetice samples.
In general, it's the Northern Europe:
Oracle 4 with Eurogenes K15:
Rise 577:
1 NW_France @ 4,976696
2 Southwest_English @ 6,367392
3 South_Dutch @ 7,551647
4 Southeast_English @ 7,764888
5 Irish @ 8,971978
6 West-Scottish @ 9,924036
7 North_German @ 9,994099
8 French @ 10,558937
9 North_Dutch @ 10,956637
10 Danish @ 11,057187

Rise150:
1 Danish @ 7,437832
2 Norwegian @ 8,275071
3 West-Norwegian @ 8,404675
4 North_Dutch @ 8,423931
5 West-Scottish @ 9,140144
6 Swedish @ 9,232579
7 Orcadian @ 9,61955
8 North_German @ 9,91944
9 Irish @ 10,070117
10 Southeast_English @ 11,259533

Interesting, here we can see very different results (maybe other Unetice samples were used?):

http://www.sciencemag.org/content/342/6155/257.abstract

http://www.sciencemag.org/content/342/6155/257/suppl/DC1

Edit: But this data is about mitochondrial DNA:

http://i1076.photobucket.com/albums/w443/priwas/4qkq.png

Were there significant differences between various sub-groups of the Unetice culture ???:

https://en.wikipedia.org/wiki/Unetice_culture#Sub-groups


Ten local sub-groups can be distinguished in its classical phase:[2]

Bohemia Group[3]
Moravia Group[4]
Slovakia Group; following the so-called Nitra Group[5]
Lower Austria Group[6]
Central Germany Group[7]
Lower Saxony Group[8][9]
Lower Lusatia Group[10]
Silesia Group[11][12]
Greater Poland (Kościan) Group[13][14]
Galicia (Western Ukraine) Group[15]

Krefter
07-11-2015, 11:15 AM
Interesting, here we can see very different results (maybe other Unetice samples were used?):

http://www.sciencemag.org/content/342/6155/257.abstract

http://www.sciencemag.org/content/342/6155/257/suppl/DC1

http://i1076.photobucket.com/albums/w443/priwas/4qkq.png

Were there significant differences between various sub-groups of the Unetice culture ???:

https://en.wikipedia.org/wiki/Unetice_culture#Sub-groups

Those ancient Unetice samples were only covered in the HV1 region and sometimes HV1 and HV2. Not much can be learned about their deep maternal relation with any modern Europeans with such low coverage. You need the entire mtDNA sequenced to do that, and there's a good chance that might not even work. Plus, those maps of similarity are based on mtDNA haplogroup freuqnecies. They're almost completely unreliable.

Generalissimo
07-11-2015, 11:29 AM
Interesting, here we can see very different results (maybe other Unetice samples were used?]

Please don't believe everything you see posted online about these samples.

Forcing ancient genomes into ADMIXTURE clusters created with modern populations, especially in tests that have a lot of clusters, isn't a useful way to test their population affinities. It's like putting the cart before the horse. In other words, it's the modern populations that should be tested with clusters created by ancient samples.

There are also other problems, like the fact that the Rise genomes should only be run using transversion (high confidence) SNPs.

Christina
07-15-2015, 12:37 AM
Just remember that the Colville tribe scored the highest because they were the only one of the tribes and bands of the Pacific Northwest to consent to and provide their DNA.

Agamemnon
07-15-2015, 12:58 AM
The Dutch linguist Frederik Kortlandt supports a model of Indo-Uralic in which the original Indo-Uralic speakers lived north of the Caspian Sea, and the Proto-Indo-European speakers began as a group that branched off westward from there to come into geographic proximity with the Northwest Caucasian languages, absorbing a Northwest Caucasian lexical blending before moving farther westward to a region north of the Black Sea where their language settled into canonical Proto-Indo-European (2002:1). Allan Bomhard suggests a similar schema in Indo-European and the Nostratic Hypothesis (1996).

The main problem here is that a genetic relationship between IE and Uralic isn't the only explanation available, then there's all the unjustified talk about "Nostratic" (which if you ask me, is a crackpot theory).