search this blog

Saturday, May 14, 2016

D-stats/nMonte open thread #2


I put together a couple of D-stats datasheets featuring the Ice Age European samples from Qiaomei Fu et al. 2016. They're available for download here and here, and can be used as input for nMonte and 4Mix. For instance, a couple of models using nMonte:

Villabruna
ElMiron 99.8
Anatolia_Neolithic 0.2
AfontovaGora3 0
Esan_Nigeria 0
GoyetQ116-1 0
Kostenki14 0
Kotias 0
MA1 0
Neandertal_Altai 0
Ulchi 0
Vestonice16 0

distance%=5.9112 / distance=0.059112

Kotias
Anatolia_Neolithic 56.45
Kostenki14 24.1
AfontovaGora3 18.1
Esan_Nigeria 1.35
ElMiron 0
GoyetQ116-1 0
MA1 0
Neandertal_Altai 0
Ulchi 0
Vestonice16 0
Villabruna 0

distance%=8.3858 / distance=0.083858

Update 17/05/2016: I wanted to see what would happen if I changed the formula from D(Chimp,Rows)(Mbuti,Columns) to D(Chimp,Columns)(Mbuti,Rows). The new datasheets are available here and here. The results are very similar, although never identical, and I'm not sure at the moment which set of sheets works better.

Villabruna
ElMiron 99.5
Anatolia_Neolithic 0.5
AfontovaGora3 0
Esan_Nigeria 0
GoyetQ116-1 0
Kostenki14 0
Kotias 0
MA1 0
Neandertal_Altai 0
Ulchi 0
Vestonice16 0

distance%=5.8334 / distance=0.058334

Kotias
Anatolia_Neolithic 69.5
AfontovaGora3 24.3
Ulchi 5.35
Kostenki14 0.85
ElMiron 0
Esan_Nigeria 0
GoyetQ116-1 0
MA1 0
Neandertal_Altai 0
Vestonice16 0
Villabruna 0

distance%=8.6196 / distance=0.086196

See also...

Yamnaya = Khvalynsk + extra CHG + maybe something else

D-stats/nMonte open thread

D-stats/4mix tour of ancient Eurasia

186 comments:

Krefter said...

Awesome job. Some Europeans might have a little El Miron. The best way to confirm would to have Goyet Q116 as a row population.

Iberia_MN
"Anatolia_Neolithic" 72.9
"Loschbour" 24.05
"ElMiron" 3.05
"Yamnaya_Kalmykia" 0
"Hungary_HG" 0

"distance%=1.5785 / distance=0.015785"

But when Loschbour and Hungary_HG are replaced by Villabruna.

Iberia_MN
"Anatolia_Neolithic" 71.6
"Villabruna" 28.4
"Yamnaya_Kalmykia" 0
"ElMiron" 0

"distance%=1.8276 / distance=0.018276"

Krefter said...

Do these results make sense? Modelling Middle Easterners as Basal Eurasian+Stone age North Eurasia. I took out all Middle Eastern row, Denosovian, and African row populations. My ghost Basal Eurasian has 0.3 closeness to all Eurasians and 0.27 to Papuans.

They make sense to me. Instead of anything close to WHG(like even El Miron(, Middle Easterners being mostly something distant from WHG(like vestonice16)+Basal Eurasian makes the most sense to me. Except Anatolia Neolithic had some WHG admixture, evident in their small amounts of Y DNA I2 and mtDNA U5.

"distance%=2.0271 / distance=0.020271"

Kotias
"Vestonice16" 32.1
"Basal" 25.9
"AfontovaGora3" 22.65
"Kostenki14" 17.65
"MA1" 1.7
"GoyetQ116-1" 0
"ElMiron" 0
"Villabruna" 0

Anatolia_Neolithic
"Vestonice16" 42.95
"Basal" 27.3
"Villabruna" 16.6
"AfontovaGora3" 13.15
"GoyetQ116-1" 0
"Kostenki14" 0
"ElMiron" 0
"MA1" 0

"distance%=1.7857 / distance=0.017857"


Cypriot
"Vestonice16" 45.8
"Basal" 27.55
"MA1" 15.2
"AfontovaGora3" 6.95
"Villabruna" 4.5
"GoyetQ116-1" 0
"Kostenki14" 0
"ElMiron" 0

"distance%=1.4812 / distance=0.014812"


BedouinB
"Vestonice16" 53.85
"Basal" 29.55
"AfontovaGora3" 11.3
"Esan_Nigeria" 5.25
"MA1" 0.05
"GoyetQ116-1" 0
"Kostenki14" 0
"ElMiron" 0
"Villabruna" 0

Davidski said...

Yeah, that's not bad. I have this ingrained idea that Basal Eurasian should be much higher then 25% in Anatolia Neolithic and Kotias, because of the 44% Basal estimate for Stuttgart in the major papers. But maybe the major papers are wrong?

How did you make your Basal Eurasian?

Anonymous said...

I have a few D-stats to request if you don't mind =)

Mbuti AfontovaGora3 Han Ust_Ishim
Mbuti AfontovaGora3 Kostenki14 Ust_Ishim
Mbuti Anatolia_Neolithic Han Ust_Ishim
Mbuti Anatolia_Neolithic Vestonice16 Ostuni1
Mbuti Bichon Han Ust_Ishim
Mbuti Bichon Kostenki14 Ust_Ishim
Mbuti ElMiron Han Ust_Ishim
Mbuti ElMiron Kostenki14 Ust_Ishim
Mbuti ElMiron Vestonice16 Ostuni1
Mbuti GoyetQ-116 Kostenki14 Ust_Ishim
Mbuti Karitiana AfontovaGora3 Ust_Ishim
Mbuti Karitiana Kostenki14 Ust_Ishim
Mbuti Karitiana MA1 Ust_Ishim
Mbuti Kotias GoyetQ-116 Vestonice16
Mbuti Kotias Han Ust_Ishim
Mbuti Kotias Kostenki14 ElMiron
Mbuti Kotias Kostenki14 GoyetQ-116
Mbuti Kotias Kostenki14 Ust_Ishim
Mbuti Kotias Kostenki14 Vestonice16
Mbuti Kotias MA1 AfontovaGora3
Mbuti Kotias Vestonice16 Ostuni1
Mbuti LaBrana1 Kostenki14 Ust_Ishim
Mbuti Loschbour Kostenki14 Ust_Ishim
Mbuti MA1 Han Ust_Ishim
Mbuti MA1 Kostenki14 Ust_Ishim
Mbuti MA1 Vestonice16 Ostuni1
Chimp Mbuti ElMiron Ust_Ishim
Chimp Mbuti GoyetQ-116 Ust_Ishim
Chimp Mbuti Han Ust_Ishim
Chimp Mbuti Kostenki14 Ust_Ishim
Chimp Mbuti Ostuni1 Ust_Ishim
Chimp Mbuti Vestonice16 Ust_Ishim
Chimp Mbuti Villabruna Ust_Ishim
Mbuti Ostuni1 Kostenki14 Ust_Ishim
Mbuti Ranchot88 Han Ust_Ishim
Mbuti Rochedane Han Ust_Ishim
Mbuti Vestonice16 Kostenki14 Ust_Ishim
Mbuti Villabruna Han Ust_Ishim
Mbuti Villabruna Kostenki14 Ust_Ishim
Chimp Yoruba ElMiron Ust_Ishim
Chimp Yoruba GoyetQ-116 Ust_Ishim
Chimp Yoruba Han Ust_Ishim
Chimp Yoruba Kostenki14 Ust_Ishim
Chimp Yoruba Ostuni1 Ust_Ishim
Chimp Yoruba Vestonice16 Ust_Ishim
Chimp Yoruba Villabruna Ust_Ishim
Mbuti Tunisian Kostenki14 Ust_Ishim
Mbuti Tunisian GoyetQ-116 Ust_Ishim
Mbuti Tunisian Vestonice16 Ust_Ishim
Mbuti Tunisian Ostuni1 Ust_Ishim
Mbuti Tunisian ElMiron Ust_Ishim
Mbuti Tunisian Han Ust_Ishim
Mbuti Tunisian Villabruna Ust_Ishim

Davidski said...

Here...

Mbuti AfontovaGora3 Han Ust_Ishim -0.0231 -4.312 232022
Mbuti AfontovaGora3 Kostenki14 Ust_Ishim -0.0507 -6.78 217817
Mbuti Anatolia_Neolithic Han Ust_Ishim -0.011 -3.21 961098
Mbuti Anatolia_Neolithic Vestonice16 Ostuni1 0.0069 1.372 242872
Mbuti Bichon Han Ust_Ishim -0.0075 -1.447 966917
Mbuti Bichon Kostenki14 Ust_Ishim -0.0741 -11.719 904064
Mbuti ElMiron Han Ust_Ishim -0.0032 -0.604 548229
Mbuti ElMiron Kostenki14 Ust_Ishim -0.0743 -10.802 522879
Mbuti ElMiron Vestonice16 Ostuni1 0.0066 0.776 207010
Mbuti GoyetQ116-1 Kostenki14 Ust_Ishim -0.0676 -9.014 629888
Mbuti Karitiana AfontovaGora3 Ust_Ishim -0.1129 -19.329 232022
Mbuti Karitiana Kostenki14 Ust_Ishim -0.0166 -3.107 906084
Mbuti Karitiana MA1 Ust_Ishim -0.0912 -16.436 683873
Mbuti Kotias GoyetQ116-1 Vestonice16 -0.0011 -0.175 512010
Mbuti Kotias Han Ust_Ishim -0.0201 -4.247 969412
Mbuti Kotias Kostenki14 ElMiron 0.0104 1.773 523887
Mbuti Kotias Kostenki14 GoyetQ116-1 0.0047 0.777 631108
Mbuti Kotias Kostenki14 Ust_Ishim -0.0449 -7.843 905549
Mbuti Kotias Kostenki14 Vestonice16 0.0009 0.15 611440
Mbuti Kotias MA1 AfontovaGora3 0.002 0.238 170270
Mbuti Kotias Vestonice16 Ostuni1 0.0072 0.941 242802
Mbuti LaBrana1 Kostenki14 Ust_Ishim -0.076 -12.223 848504
Mbuti Loschbour Kostenki14 Ust_Ishim -0.0776 -12.336 897446
Mbuti MA1 Han Ust_Ishim -0.0148 -2.894 683873
Mbuti MA1 Kostenki14 Ust_Ishim -0.0528 -8.094 643188
Mbuti MA1 Vestonice16 Ostuni1 -0.007 -0.776 178615
Chimp Mbuti ElMiron Ust_Ishim -0.0121 -2.738 528284
Chimp Mbuti GoyetQ116-1 Ust_Ishim -0.0051 -1.183 631569
Chimp Mbuti Han Ust_Ishim -0.0182 -6.131 934129
Chimp Mbuti Kostenki14 Ust_Ishim -0.005 -1.32 872372
Chimp Mbuti Ostuni1 Ust_Ishim -0.0035 -0.662 267551
Chimp Mbuti Vestonice16 Ust_Ishim -0.0045 -1.041 613349
Chimp Mbuti Villabruna Ust_Ishim -0.0116 -2.978 734472
Mbuti Ostuni1 Kostenki14 Ust_Ishim -0.0917 -12.303 258576
Mbuti Ranchot88 Han Ust_Ishim -0.0142 -2.71 332387
Mbuti Rochedane Han Ust_Ishim -0.0018 -0.32 188559
Mbuti Vestonice16 Kostenki14 Ust_Ishim -0.0914 -12.839 610307
Mbuti Villabruna Han Ust_Ishim -0.0011 -0.219 762594
Mbuti Villabruna Kostenki14 Ust_Ishim -0.0752 -11.295 734000
Chimp Yoruba ElMiron Ust_Ishim -0.0116 -2.658 528284
Chimp Yoruba GoyetQ116-1 Ust_Ishim -0.0075 -1.771 631569
Chimp Yoruba Han Ust_Ishim -0.0216 -6.874 934129
Chimp Yoruba Kostenki14 Ust_Ishim -0.0077 -1.951 872372
Chimp Yoruba Ostuni1 Ust_Ishim -0.0048 -0.953 267551
Chimp Yoruba Vestonice16 Ust_Ishim -0.0077 -1.717 613349
Chimp Yoruba Villabruna Ust_Ishim -0.0152 -3.83 734472
Mbuti Tunisian Kostenki14 Ust_Ishim -0.0378 -9.273 561097
Mbuti Tunisian GoyetQ116-1 Ust_Ishim -0.042 -10.046 411854
Mbuti Tunisian Vestonice16 Ust_Ishim -0.0487 -11.799 436796
Mbuti Tunisian Ostuni1 Ust_Ishim -0.0478 -10.486 201864
Mbuti Tunisian ElMiron Ust_Ishim -0.0595 -14.698 385109
Mbuti Tunisian Han Ust_Ishim -0.0171 -5.247 592408
Mbuti Tunisian Villabruna Ust_Ishim -0.0736 -19.729 505608

Krefter said...

@Davidski,

"How did you make your Basal Eurasian"

I made him a little more distant from Eurasians than Ust-Ishim is. So, 0.30 results for all Eurasians and 0.27 for Papuans.

Krefter said...

It's very difficult to discern what the non-Basal Eurasian side of Middle Easterners is. This is one part of European origins we haven't cracked and we need ancient DNA from the Middle East to solve it.

At this point it looks like non-Basal Eurasian in the Middle East is mostly Paleo-European(who might have originated in the Middle East)+WHG+ANE.

Here are some results I've gotten. I modeled Cypriot, CHG, and Anatolia_Neolithic as 30% Basal Eurasian, and then modeled their other 70% as a mixture of Stone age North Eurasians.

Cypriot: 61% Paleo Europe(47.35% Vestonice16, 13.8% Kostinki14), 30.05% Karelia_HG, 8.8% ANE: @ 0.024135
CHG: 49% Paleo Europe(5.35% Vestonice16, 43.5% Kostinki14), 27.95% Karelia_HG, 23.5% ANE: @ 0.028634
Anatolia Neolithic: 56% Paleo Europe(Vestonice16), 30.75% Karelia_HG, 13.6% WHG: @ 0.022152

IMO, we have to take into account the possibility of East Asian-related and (extinct)Crown Eurasian admixture in ancient West Eurasians. (extinct)Crown Eurasian admixture would be impossible to detect. It wouldn't have the affect of anti-East Asian affinity that Basal Eurasian has and wouldn't have the affect of pro-East Asian affinity that East Asian ancestry would.

Anonymous said...

Chimp Yoruba Villabruna Ust_Ishim -0.0152 -3.83 734472

Chimp Mbuti Vestonice16 Ust_Ishim -0.0045 -1.041 613349
Chimp Mbuti Villabruna Ust_Ishim -0.0116 -2.978 734472

There it is again. Villabruna looks more related to anything in the world than earlier UP Europeans.

Davidski said...

Ust-Ishim is not an UP European. And you're also not taking into account archaic ancestry, which has a big impact on the D-stats.

Anonymous said...

@David

Let me rephrase: Villabruna looks to have more affinity with anything post-LGM than pre-LGM. Mal'ta was pre-LGM, AG3 post-LGM.

LGM was a bottle neck. Can this explain it? I recall a paper showing mesolithic HGs having a lower diversity than Asians. However, if you look at the initial UP mtDNA diversity (I know, not a one on one tracker for autosomal DNA) it was far larger.

Does the idea that Villabruna has affinity with anything post-LGM somehow fit into this?

Alberto said...

@Krefter

While in theory the distance of Basal Eurasian to all Crown Eurasians should be equal, I think that's only possible in an ideal world. Each population will have different level of drift after branching off from Basal Eurasian, and that drift will follow different paths, that when placed on space will make them be closer or further from Basal Eurasian.

For example, Ust-Ishim's distance to some populations:

Ust-Ishim <--> Villabruna = 0.2737
Ust-Ishim <--> Atayal = 0.2002

More important is that the results you get with the ghost are consistent with the D-stats that we can actually run. So for example you model:

Anatolia_Neolithic
"Vestonice16" 42.95
"Basal" 27.3
"Villabruna" 16.6
"AfontovaGora3" 13.15
"GoyetQ116-1" 0
"Kostenki14" 0
"ElMiron" 0
"MA1" 0

Seems to be at odds with what we know. Anatolia_Neolithic shares significantly more drift with Villabruna than with Vestonice16, and also a lot more than with AfontovaGora3.

Here is with a ghost Basal Eurasian based on Hungary_HG and Anatolia_Neolithic:



Anatolia_Neolithic
"Villabruna" 50.95
"Basal_Eurasian" 48.1
"AfontovaGora3" 0.95
"Atayal" 0
"Esan_Nigeria" 0
"GoyetQ116-1" 0
"Kostenki14" 0
"MA1" 0
"Ust_Ishim" 0
"Vestonice16" 0
distance=0.015999

Difficult to say if actually Anatolia_Neolithic is 50% Villabruna or if it has another West Eurasian ancestry that we simply don't have yet. But for now that's all we can do.

The distance from this Basal Eurasian ghost and Atayal is 0.3589 and to Villabruna is 0.3210.

Here a model for Kotias:

Kotias
"Basal_Eurasian" 32.1
"Kostenki14" 31.25
"AfontovaGora3" 26.5
"Villabruna" 10.15
"Atayal" 0
"Esan_Nigeria" 0
"GoyetQ116-1" 0
"MA1" 0
"Ust_Ishim" 0
"Vestonice16" 0
distance=0.079073

This ghost is for the first sheet (D-stats1.txt) provided by Davidski in the post

Basal_Eurasian,0.46679,0.4057,0.42572,0.31455,-0.58369,0.41674,0.31372,0.41227,0.4107,0.31501,0.27507,0.33202,0.29588,0.32064,0.28032,0.32174,0.37897,0.33901,0.09475

Alberto said...

With the UP genomes I haven't been able to do much. One can force a model for Villabruna removing ElMiron where it will be 82% Vestonice16 and 18% AfontovaGora3, but that doesn't look too good. And I can't see any ElMiron ancestry so far (when Villabruna is included).

Another observation so far is that when including the 4 Villabruna (WHG) samples, most populations get Hungary_HG and Loschbour, and occasionally some Bichon. But I didn't find yet one that takes Villabruna.

If anyone has any request to try something with the sheets provided in the post I'll try to help running them.

Alberto said...

Another interesting observation that I'll need more time to check carefully, but in case anyone else might also want to look into it:

The pattern of MA1 vs. AG3 looks interesting. Both Karelia_HG and Kotias prefer AG3. So does Yamnaya_Samara. However, Yamnaya_Kalmykia and Afanasievo also take MA1. Then Corded_Ware_Germany takes only AG3, but Bell_Beaker_Germany also takes MA1. Then moving onto modern populations, Northern Europeans take AG3, but Southern Europeans also take MA1. The line dividing north and south is interesting: Lithuanian, Norwegian and English_Cornwall only take AG3, but English_Kent already takes MA1, as does French and Spanish. But Basque_Spanish again only takes AG3. Then as you move South East, MA1 increases and AG3 decreases, where Greek2 basically only takes MA1.

The pattern doesn't look completely random to me. And the interesting part is that it goes back to Yamnaya and CW-BB. Here a sample of the populations used:

Corded_Ware_Germany
"Hungary_HG" 40.1
"AfontovaGora3" 30.9
"Basal_Eurasian" 23.35
"Loschbour" 4.15
"Bichon" 1.25
"Atayal" 0.25
"Esan_Nigeria" 0
"MA1" 0
"Villabruna" 0
distance=0.018342

Bell_Beaker_Germany
"Hungary_HG" 43.9
"Basal_Eurasian" 26.7
"AfontovaGora3" 13.25
"MA1" 9.75
"Bichon" 6.3
"Esan_Nigeria" 0.1
"Atayal" 0
"Villabruna" 0
"Loschbour" 0
distance=0.014126

(Note: I've been running these tests with the ghost Basal Eurasian. I'll have to double check with Anatolia Neolithic instead, though I think it shouldn't change).

Matt said...

Re: the D-stats sheet, can I ask if you only have the one Chimp sample in your dataset? If you have more than one, I'd be interested to see the same sheet with D (Chimp Pop1 Chimp Pop2) in place of D(Chimp, Test)(Mbuti, X). Or if not, is it possible to do the sheet with D(Chimp, X)(Mbuti, Test)?

I think either of those might change some behaviour with the Yoruba2, BedouinB2, Druze2 as X in ways that might be interesting to plot and model.
Also, would it be possible to do the same datasheet with El Miron, Vestonice16, Kostenki14, GoyetQ116-1 as columns and with Villabruna as column in place of LaBrana?

Alberto said...

Removing the Basal Eurasian ghost and including Anatolia_Neolithic and Kotias, the difference still stands in CW vs. BB. So it seems that MA1 was not a dead lineage after all? And it actually seems to be useful to track a population that spread it?

Corded_Ware_Germany
"Anatolia_Neolithic" 34.85
"Hungary_HG" 24.9
"AfontovaGora3" 24.2
"Kotias" 15.9
"Atayal" 0.15
"Esan_Nigeria" 0
"MA1" 0
"Villabruna" 0
"Loschbour" 0
"Bichon" 0
distance=0.012459

Bell_Beaker_Germany
"Anatolia_Neolithic" 43.8
"Hungary_HG" 24.7
"Kotias" 14.15
"MA1" 10.4
"AfontovaGora3" 6.95
"Atayal" 0
"Esan_Nigeria" 0
"Villabruna" 0
"Loschbour" 0
"Bichon" 0
distance=0.007875

huijbregts said...

@ Alberto
I think a more parsimonious model is:
Bell_Beaker_Germany:

"Corded_Ware_Germany" 65.05
"Germany_MN" 30.35
"Hungary_HG" 2.6
"Motala_HG" 1
"Esan_Nigeria" 0.6
"Atayal" 0.4
"AfontovaGora3" 0
"Anatolia_Neolithic" 0
"Bichon" 0
"Iberia_EN" 0
"Iberia_MN" 0
"Loschbour" 0
"MA1" 0
"Villabruna" 0
distance%=0.3689

Rob said...

Alberto

Im not sure what to make of the differential affinities to MA-1 and AF-G, if Ive even understood completely.

However, it does remind me of something you say back in November last year, after the Jones paper, that there are somewhat different affinities to hypothetical subtypes within CHG also ?

Chad said...

Why would Bell Beaker be 65% CW with a 100% Y and over 50-75% mt turnover? That's not parsimonious at all.

Matt said...

Reversing the rows and columns in the sheet for AfontovaGora3 vs MA1:

AfontovaGora3 MA1 Difference
Samara_HG 0.4704 0.4345 0.0359
Satsurblia 0.3811 0.3661 0.015
Denisovan -0.5599 -0.5742 0.0143
Karitiana2 0.4172 0.404 0.0132
Mansi2 0.4013 0.3896 0.0117
Motala_HG2 0.4152 0.4045 0.0107
Papuan2 0.3265 0.3184 0.0081
Iberia_Chalcolithic 0.3696 0.3637 0.0059
South_Indian2 0.3701 0.3646 0.0055
Druze2 0.3631 0.3613 0.0018
Yoruba 0.0937 0.0934 0.0003
Han2 0.3526 0.3534 -0.0008
Dai2 0.3537 0.3545 -0.0008
Iranian_Jew2 0.3658 0.3667 -0.0009
Munda2 0.354 0.355 -0.001
Cypriot2 0.3659 0.3674 -0.0015
BedouinB2 0.3448 0.3467 -0.0019
LaBrana1 0.3779 0.3835 -0.0056
Anatolia_Neolithic2 0.3539 0.3604 -0.0065

AfontovaGora3 exhibits more "ANE" behaviour here in a sense (plus Denisovan for some reason and apparently Papuan). MA1 is mildly more linked to the EEF associated West Eurasians.

Possibly this relates to AG3 being later in history.

Comparison for ElMiron vs Goyet116-1:

ElMiron GoyetQ116-1 Difference
LaBrana1 0.4752 0.4094 0.0658
Motala_HG2 0.4276 0.3913 0.0363
Iberia_Chalcolithic 0.405 0.3783 0.0267
Denisovan -0.5548 -0.5803 0.0255
Samara_HG 0.404 0.3872 0.0168
Anatolia_Neolithic2 0.3795 0.3631 0.0164
Mansi2 0.3714 0.3577 0.0137
Iranian_Jew2 0.3673 0.356 0.0113
Druze2 0.3656 0.3548 0.0108
Cypriot2 0.3713 0.3605 0.0108
BedouinB2 0.3537 0.344 0.0097
Satsurblia 0.3581 0.3497 0.0084
Karitiana2 0.3559 0.3521 0.0038
Han2 0.3462 0.3437 0.0025
South_Indian2 0.3521 0.3499 0.0022
Dai2 0.3443 0.3426 0.0017
Munda2 0.3438 0.3426 0.0012
Papuan2 0.3148 0.3143 0.0005
Yoruba 0.091 0.0932 -0.0022

El Miron vs Villabruna:

ElMiron Villabruna Difference
Motala_HG2 0.4276 0.4663 0.0387
Samara_HG 0.404 0.4305 0.0265
Iberia_Chalcolithic 0.405 0.4198 0.0148
Satsurblia 0.3581 0.3706 0.0125
Druze2 0.3656 0.3777 0.0121
Anatolia_Neolithic2 0.3795 0.3894 0.0099
Iranian_Jew2 0.3673 0.3759 0.0086
Cypriot2 0.3713 0.3798 0.0085
BedouinB2 0.3537 0.3617 0.008
LaBrana1 0.4752 0.4828 0.0076
Mansi2 0.3714 0.3786 0.0072
Yoruba 0.091 0.093 0.002
South_Indian2 0.3521 0.3509 -0.0012
Karitiana2 0.3559 0.3532 -0.0027
Munda2 0.3438 0.3396 -0.0042
Papuan2 0.3148 0.3067 -0.0081
Han2 0.3462 0.3377 -0.0085
Dai2 0.3443 0.3351 -0.0092
Denisovan -0.5548 -0.5649 -0.0101

Matt said...

Using the mass D-stats in D-stats1 to try and find finer structure among the 4 WHGs / Villabruna cluster:

http://i.imgur.com/fFqyWME.png

PC1 between 4 WHGs is a strong tendency for greater La Brana relatedness among the Western WHG (Bichon and Loschbour) and lesser among Villabruna and Hungary_HG, vs relatively greater relatedness to Anatolia_Neolithic and Samara_HG among Villabruna and Hungary_HG.

PC2 is greater relatedness to an array of populations for more recent Loschbour and Hungary_HG, vs lesser for less recent Bichon and Villabruna.

Plus ElMiron:

http://i.imgur.com/sK8nvcT.png

Plus Vestonice16:

http://i.imgur.com/vIXAJHT.png

Particularly Vestonice16 recapitulates the same distinction between the Villabruna cluster, with the HungaryHG pole more related to Samara, Anatolia Neolithic and Motala and the Bichon/Loschbour pole related to La Brana.

Ryukendo K said...
This comment has been removed by the author.
Krefter said...

@Alberto,

How did you create your Basal Eurasian ghost?

There's a good chance Basal Eurasian will turn out to be equally related to all Crown Eurasians. This is because all East Asians are equally related to West Eurasians.

I've realized the model(Basal+WHG) doesn't work with D-stats for EEF. Unless Basal Eurasian had EHG and East Asian shared drift and was strangely distant from WHG, the model can't work.

It looks like your Basal Eurasian adapted to the problem of Anatolia_Neolithic not being WHG+Basal Eurasian, by being extra close to EHG and East Asians and extra distant from WHG.

Ryukendo K said...
This comment has been removed by the author.
Alberto said...

@Krefter

Yes, as RK said, I made it based on the assumption that Anatolia_Neolithic is a mix of Hungary_HG (KO1) and Basal Eurasian. So I calculate the distance between both in each column and multiply it by 2.1 (in this case, since AN should be almost 50% Basal Eurasian. Changing that number depending on the percentage you estimate). You can see the formula here (selecting each cell of Basal Eurasian):

https://docs.google.com/spreadsheets/d/1HDpm_lBPozmgg9pngBcOH0UXfJJvhZKVstElVKjkd7Q/edit?usp=sharing


@RK

Yes, I suspect also that EHG has some CHG admixture, though I'm not sure if both of them or only Samara_HG (have you seen stats with both in this sense?).


@Rob

I guess that in the same way that we're seeing now that some WHGs are better references for later populations than others, we'll see the same with CHG when we get more samples. We'll see then if Kotias is the best reference or not.

What I expected re: ANE, was that AG3 would be the better reference for Europe, as it was for Native Americans and Karelia_HG. But then I started to see a lot of MA1 in Southern Europe. Going back to LN/BA and even back to Yamnaya the difference stands, so it looks like it means something.

Speculating, could be that a first wave of early CHG into the steppe were quite Kotias-like, but that later a new wave arrived carrying this MA1 admixture. But that's just a guess. If I have time I'll try to make a spreadsheet so others can check the pattern and see if they find any meaning to it.

Olympus Mons said...

Alberto.
Steppe..steppe ... steppe. Well really there is something starting to stink.


*Are still saying that bell beakers had anything to do with steppe? No, right. That is becoming just weird.
*Bell beaker is Ma1 because its stock Never really crossed back the caucasus to the north, except a small amount that after ubaid(5000bc) had to cross and were part of the later yamnaya (as some J2).

So what is this steppe pounding here?






Olympus Mons said...

Cam someone point me a table with which populations have Ma1? Does it exist?

Olympus Mons said...

And again Bell beaker having a sniff of sub saharan (esa nigeria). Where do you guys think they got that? :) hummm mtdna l3f, l3f my dear friends.

Chad said...

KO1 and EHG are significantly further from Ust Ishim, to the level of MN pops. Unlike other WHG samples. Remove EHG and KO1 from your runs and see how much stuff changes. Use Villabruna and AG3.

Olympus Mons said...

And again Bell beaker having a sniff of sub saharan (esa nigeria). Where do you guys think they got that? :) hummm mtdna l3f, l3f my dear friends.

Chad said...

Ko1 is also significantly closer to EHG and near significant to AG3. QpAdm has no issue making KO1 a mix of WHG, AN, and ANE.

Chad said...

Olympus,
That Sub Saharan is deamination, from not being UDG treated. There is no SSA in quality samples. Your ideas have no basis in fact.

Olympus Mons said...

Chad. Respectfully you have no ideia what my ideia are:)
And regarding fact, finding some genes on a fluke finding in some cave it doesnt make fact. Makes a nice grounds for inference. Not fact.
Everybody has a pet story, that is it. Dont fool yourself. So What is yours?

Matt said...

Re: Basal Eurasian, I tried modeling it by using a regression equation on the populations (Anatolia_Neolithic, Germany_MN, Hungary_EN, Hungary_HG, Iberia_EN, LBK_EN, Villabruna) with a regression of the asymmetry between LaBrana1 and Samara_HG.

My assumption was that where relatedness to LaBrana1 and Samara_HG converged to 0, would be Basal Eurasian.

In case anyone wants to try this in nMonte, outputs as follows for 0:

Variable Anatolia_Neolithic2 BedouinB2 Cypriot2 Dai2 Denisovan Druze2 Han2 Iberia_Chalcolithic Iranian_Jew2 Karitiana2 LaBrana1 Mansi2 Motala_HG2 Munda2 Papuan2 Samara_HG Satsurblia South_Indian2 Yoruba

Basal_Equal_Samara_LaBrana 0.4294 0.3879 0.4065 0.3293 -0.5858 0.3988 0.3289 0.4205 0.3973 0.3385 0.3793 0.3598 0.3875 0.3336 0.2959 0.3793 0.3772 0.3473 0.0953

(0.3973 is the stat for LaBrana and Samara, and for some reason Motala_HG2 is higher than the average of these)

The stats the ancient UP Europeans get here make me skeptical of Basal Eurasian more though.

Anonymous said...

@David

O, just for the heel of it. Genetiker published a ADMIXTURE run and Paglicci108 showed, I think, some ENA affinity.

Mbuti Paglicci108 Villabruna Vestonice16
Mbuti Paglicci108 Villabruna Muerii1
Mbuti Paglicci108 Villabruna Kostenki14
Mbuti Paglicci108 Villabruna Paglicci133
Mbuti Paglicci108 Villabruna Ostuna2

Mbuti Paglicci108 Villabruna AfontovaGora3
Mbuti Paglicci108 Villabruna Malta

Mbuti Paglicci108 Loschbour Villabruna
Mbuti Paglicci108 Loschbour ElMiron
Mbuti Paglicci108 Loschbour GoyetQ166-1
Mbuti Paglicci108 Loschbour GoyetQ-2

Anonymous said...

Sorry, ANE affinity, that was

Matt said...

Just after posting that, using a multiple regression equation method and the same set of populations (Anatolia_Neolithic, Germany_MN, Hungary_EN, Hungary_HG, Iberia_EN, LBK_EN, Villabruna), here's simulated stats for:

1) "Basal Eurasian" (equally related to MotalaHG, LaBrana1 and SamaraHG):
Anatolia_Neolithic2 BedouinB2 Cypriot2 Dai2 Denisovan Druze2 Han2 Iberia_Chalcolithic Iranian_Jew2 Karitiana2 LaBrana1 Mansi2 Motala_HG2 Munda2 Papuan2 Samara_HG Satsurblia South_Indian2 Yoruba

Basal_All_Equal 0.4298 0.3863 0.4053 0.3229 -0.5718 0.401 0.3227 0.4181 0.3948 0.3302 0.3637 0.3505 0.3637 0.3272 0.2917 0.3637 0.3744 0.3423 0.0932

2) "Basal Eurasian" (equally related to MotalaHG, LaBrana1 and SamaraHG and also separately equally related to Karitiana and Han):

(column names as above)

Basal_All_Equal_2 0.4311 0.3844 0.3966 0.3278 -0.5495 0.3915 0.3272 0.4479 0.3796 0.3272 0.3615 0.3369 0.3615 0.3239 0.2936 0.3615 0.3755 0.3473 0.0978

3) "Basal Eurasian" (equally related to all of MotalaHG, LaBrana1, SamaraHG, Karitiana, Han):

(column names as above)

Basal_All_Equal_3 0.4304 0.4057 0.4156 0.3311 -0.559 0.4 0.3321 0.4553 0.3883 0.3321 0.3321 0.3386 0.3321 0.3262 0.3057 0.3321 0.3853 0.3528 0.0996

Out of these, 3) would seem to best satisfy what Basal Eurasian should be, theoretically. I think 1) may end up being closest to reality, though.

Alberto said...

I made a run with many populations to see the AfontovaGora3 vs. MA1 pattern because this looked like the most interesting thing from the new datasheet. I included Kotias and Anatolia_Neolithic instead of using a Basal Eurasian ghost. This would be more realistic, though since Kotias has a good amount of ANE (and preference for AG3, see model above), this makes things less clear than when using the Basal Eurasian ghost.

Anyhow, here it is:

https://docs.google.com/spreadsheets/d/1KNPbAczNTVu_shiMcpx-ZCexceHoBAPndLSuy6YsmVw/edit?usp=sharing

I included Ust-Ishim to be able to run more populations. I guess Ust-Ishim there just means "undefined South Asian component". I also run with 2 WHGs (Loschbour and Hungary_HG) in case that would give any other clue.

For what I can see, I'd say that AG3 represents a "northern" source of ANE, while MA1 might represent a "southern" source of ANE. Or who knows.

Alberto said...

@Olympus Mons

I'm actually not much of a steppe guy. I think, like you, that R1b came from the south. But in my own pet theory it would be the ones who came after Shulaveri. I think that R1b-L23 could be found in Transcaucasia in the early 4th millennium BCE, giving rise to cultures like Kura-Araxes and Maykop. But this is something that only time will tell.

Alberto said...

@Matt

Modeling Anatolia_Neolithic and Kotias with the 3 Basal ghosts, it prefers the first one, but the amount of Basal Eurasian looks too high:

Anatolia_Neolithic
"Basal_All_Equal" 80.45
"Hungary_HG" 19.5
"AfontovaGora3" 0.05
"Basal_All_Equal_2" 0
"Basal_All_Equal_3" 0
"Loschbour" 0
"Kostenki14" 0
distance=0.016905

Kotias
"Basal_All_Equal" 42.3
"Kostenki14" 28.6
"AfontovaGora3" 21.75
"Basal_All_Equal_3" 7.35
"Basal_All_Equal_2" 0
"Hungary_HG" 0
"Loschbour" 0
distance=0.08142

The third one looks closer to the expected values:

Anatolia_Neolithic
"Basal_All_Equal_3" 56.65
"Hungary_HG" 36.15
"Kostenki14" 7.2
"Loschbour" 0
"AfontovaGora3" 0
distance=0.039272

Kotias
"Kostenki14" 38.95
"Basal_All_Equal_3" 36.55
"AfontovaGora3" 23.95
"Hungary_HG" 0.55
"Loschbour" 0
distance=0.084122

Matt said...

@ Alberto, thanks for testing that in nMonte. You're right that it comes out high for the first one. That ghost of "Basal_All_Equal" is effectively not really Basal Eurasian as we would think of it (a minority contributor equally related to all "Crown Eurasians" that contributes to an essentially mostly WHG mix to create Anatolia_Neolithic) but something more like a ghost of a "Levantine Hunter Gatherer" (equally related to Samara_HG, Motala_HG and LaBrana but decidedly West Eurasian in its affinities) that would contribute the vast majority of Anatolia_Neolithic's ancestry. "Basal_All_Equal_3" is much more like the traditional Basal Eurasian construct (equally related to all Crown Eurasians).

Rob said...

Alberto

I probably did not express myself clearly.

Remember last year you and ? Ryu thought there might be different "kinds" of CHG in Europe. One type is found in the southern Europe and southern Asia through to north western Europe. A second type found in Yamnaya and Northern Europe, and the two types interact in north western Europe.

Maybe that difference in CHG subtype affinity was actually reflecting what you've now found- actually a different in ANE type

Olympus Mons said...

@Alberto,
Pet theories. We grow attached to them... I meant it. :)

However you are wrong. Kura-araxes (or previous Sioni) or Maikop are too late to allow R1b to be dispersed by Bell Beakers.
And those Kura will be mainly J2/J1 (naturally some R1b, those that stood and didn't leave). Maikop will have r1b (those that flee and crossed the mountains to flee the serpents but were a minority to whom ever already was in the region (maybe R1b but not L23).

My thesis is in chapters and suppl. see suppl 1 "The serpents are coming".
http://blogs.sapo.pt/cloud/file/eb6b52b82097d41dfa0e5797a2fa7945/olympusmons/2016/From%20Shulaveri%20to%20Bell%20beaker.pdf

Ubaid was a leveling event. One that did a turnover of population. It was something of biblical proportions. from 4.800 bc onwards it was a whole new people. -- J2/J1 from the south and southeast.

Chad said...

Olympus,

Do you even know the R1b tree? If you think M269 expanded from Iberia in the Chalcolithic, that is where Yamnaya would have to start. That makes no sense. L23 is needed before you go to L51, which you don't seem to understand. At least, that is what it seems like from your words...

" And the spread of that cultural and genetic trait started in the Iberia peninsula, because after the immediate ending of the SSC not millennia but centuries later pure r1b (M269) inhabit the peninsula making the downstream clades that populate western world (L11 and M51).”

Not a single bit of that makes sense.... It goes M269> L23> L51. Do you really think Yamnaya came from Iberia? Come on man!

Davidski said...

Matt,

Here's the datasheet with Villabruna in the columns. Btw, I don't think using two chimps would work, even if it were possible. I tried Gorilla + Chimp before I settled on Chimp + Mbuti. The results weren't informative.

https://drive.google.com/file/d/0B9o3EYTdM8lQeEl2VHlzNHZxSEE/view?usp=sharing

rk,

Mbuti Bichon Altaian Karasuk_outlier 0.0235 8.328 29748 28383 593516
Mbuti Loschbour Altaian Karasuk_outlier 0.0271 9.509 25156 23830 503131
Mbuti LaBrana1 Altaian Karasuk_outlier 0.0232 8.343 28262 26983 563991
Mbuti MA1 Altaian Karasuk_outlier 0.032 10.788 21692 20349 427046
Mbuti Karelia_HG Altaian Karasuk_outlier 0.0301 11.493 28317 26665 556333
Mbuti AfontovaGora2 Altaian Karasuk_outlier 0.0343 6.343 2476 2312 48564
Mbuti Karitiana Altaian Karasuk_outlier -0.0133 -5.555 29314 30108 594705
Mbuti Zapotec Altaian Karasuk_outlier -0.013 -6.305 29226 29996 594705
Mbuti Daur Altaian Okunevo -0.0511 -15.697 13783 15267 295293
Mbuti Oroqen Altaian Okunevo -0.0518 -15.712 13807 15314 295293
Mbuti Japanese Altaian Okunevo -0.0504 -15.726 13779 15242 295293
Mbuti Korean Altaian Okunevo -0.0526 -15.663 13748 15274 295293
Mbuti Han_NChina Altaian Okunevo -0.0517 -15.919 13739 15236 295293
Mbuti Han Altaian Okunevo -0.0521 -16.622 13738 15247 295293
Mbuti Yi Altaian Okunevo -0.0481 -14.603 13791 15185 295293
Mbuti She Altaian Okunevo -0.0519 -15.766 13743 15248 295293
Mbuti Tu Altaian Okunevo -0.0467 -14.696 13800 15153 295293
Mbuti Tujia Altaian Okunevo -0.0505 -15.53 13762 15225 295293
Mbuti Miao Altaian Okunevo -0.0501 -15.383 13764 15217 295293
Mbuti Dai Altaian Okunevo -0.048 -15.103 13763 15151 295293
Mbuti Kinh_Vietnam Altaian Okunevo -0.0484 -15.041 13761 15162 295293
Mbuti Lahu Altaian Okunevo -0.0482 -14.825 13770 15164 295293
Mbuti Thai Altaian Okunevo -0.0422 -13.494 13792 15007 295293
Mbuti Cambodian Altaian Okunevo -0.0433 -13.862 13771 15017 295293
Mbuti Ami Altaian Okunevo -0.0472 -13.906 13789 15156 295293
Mbuti Atayal Altaian Okunevo -0.0438 -12.54 13839 15108 295293
Mbuti Uzbek Altaian Okunevo -0.0101 -3.665 14159 14449 295293
Mbuti Uygur Altaian Okunevo -0.0179 -6.249 14107 14621 295293
Mbuti Turkmen Altaian Okunevo -0.0059 -1.972 14194 14362 295293
Mbuti Turkish Altaian Okunevo 0.007 2.472 14230 14031 295293

epoch,

I can combine them, although not sure if that'll help. Which stats do you want? I can only run up to 30 per request though.

Krefter said...

@RK,
"Krefter, the pattern with EHG being especially weird in Crown-Basal terms is repeated in stats with Ust-Ishim as well
----------------------------------------------------------------------
Its likely that EHG contains a small quantity of Basal Eurasian ancestry via a small fraction of CHG, which is what Chad's previous qpADm shows as well."

This can't explain Basal Eurasian being closer to East Asians than to WHG.

Krefter said...

@Alberto,
"Seems to be at odds with what we know. Anatolia_Neolithic shares significantly more drift with Villabruna than with Vestonice16, and also a lot more than with AfontovaGora3."

The model doesn't claim West Asians are 40-50% Vestonice. It claims they're 40-50% from a brother clade of Vestonice(such as WHG), plus some ANE and WHG. I have no idea whether this is accurate or not and still investigating.

Chad said...

Still not sold on Basal Eurasian here...

result: Mbuti Ust_Ishim GoyetQ116-1 LaBrana1 -0.0209 -3.108 32326 33705 630038
result: Mbuti Ust_Ishim GoyetQ116-1 Loschbour -0.0159 -2.620 33755 34842 649774
result: Mbuti Ust_Ishim GoyetQ116-1 Hungary_HG -0.0288 -4.373 28876 30591 552722
result: Mbuti Ust_Ishim GoyetQ116-1 Motala_HG -0.0201 -3.661 33512 34885 635312
result: Mbuti Ust_Ishim GoyetQ116-1 Karelia_HG -0.0272 -4.276 32819 34655 620864
result: Mbuti Ust_Ishim GoyetQ116-1 AfontovaGora3 -0.0265 -3.482 10033 10580 195115
result: Mbuti Ust_Ishim GoyetQ116-1 MA1 -0.0102 -1.510 25537 26064 475460
result: Mbuti Ust_Ishim GoyetQ116-1 Iberia_EN -0.0376 -6.967 33507 36123 635354
result: Mbuti Ust_Ishim GoyetQ116-1 Anatolia_Neolithic -0.0395 -7.521 35091 37978 656598
result: Mbuti Ust_Ishim GoyetQ116-1 Satsurblia -0.0365 -5.409 24585 26448 452806
result: Mbuti Ust_Ishim GoyetQ116-1 Karitiana -0.0124 -2.068 36834 37757 656924
result: Mbuti Ust_Ishim GoyetQ116-1 Han -0.0128 -2.261 37098 38063 656924
result: Mbuti Ust_Ishim GoyetQ116-1 Papuan -0.0188 -2.747 37079 38503 656923
result: Mbuti Ust_Ishim LaBrana1 Loschbour 0.0055 0.996 41078 40632 889528
result: Mbuti Ust_Ishim LaBrana1 Hungary_HG -0.0103 -1.756 30707 31344 650517
result: Mbuti Ust_Ishim LaBrana1 Motala_HG -0.0006 -0.137 41281 41333 833856
result: Mbuti Ust_Ishim LaBrana1 Karelia_HG -0.0091 -1.586 41427 42190 808272
result: Mbuti Ust_Ishim LaBrana1 AfontovaGora3 -0.0033 -0.484 11664 11741 224720
result: Mbuti Ust_Ishim LaBrana1 MA1 0.0102 1.756 35061 34353 638962
result: Mbuti Ust_Ishim LaBrana1 Iberia_EN -0.0199 -4.406 42784 44521 825963
result: Mbuti Ust_Ishim LaBrana1 Anatolia_Neolithic -0.0212 -5.075 47375 49424 893801
result: Mbuti Ust_Ishim LaBrana1 Satsurblia -0.0173 -2.993 33952 35147 625214
result: Mbuti Ust_Ishim LaBrana1 Karitiana 0.0064 1.229 50718 50074 898226
result: Mbuti Ust_Ishim LaBrana1 Han 0.0062 1.267 51618 50986 898226
result: Mbuti Ust_Ishim LaBrana1 Papuan 0.0002 0.025 51910 51894 898225
result: Mbuti Ust_Ishim Loschbour Hungary_HG -0.0162 -2.937 30097 31088 667060
result: Mbuti Ust_Ishim Loschbour Motala_HG -0.0063 -1.502 41961 42490 878313
result: Mbuti Ust_Ishim Loschbour Karelia_HG -0.0136 -2.423 42992 44178 850945
result: Mbuti Ust_Ishim Loschbour AfontovaGora3 -0.0059 -0.926 11849 11990 229366
result: Mbuti Ust_Ishim Loschbour MA1 0.0064 1.122 37100 36627 677413
result: Mbuti Ust_Ishim Loschbour Iberia_EN -0.0248 -5.662 44290 46543 864624
result: Mbuti Ust_Ishim Loschbour Anatolia_Neolithic -0.0256 -6.458 50189 52830 952100
result: Mbuti Ust_Ishim Loschbour Satsurblia -0.0254 -4.594 35910 37782 670480
result: Mbuti Ust_Ishim Loschbour Karitiana 0.0022 0.440 54265 54027 961123
result: Mbuti Ust_Ishim Loschbour Han 0.0019 0.415 55225 55013 961123
result: Mbuti Ust_Ishim Loschbour Papuan -0.0036 -0.613 55516 55920 961122

Chad said...

result: Mbuti Ust_Ishim Motala_HG Karelia_HG -0.0086 -1.837 40195 40890 820110
result: Mbuti Ust_Ishim Motala_HG AfontovaGora3 0.0010 0.174 11265 11242 227665
result: Mbuti Ust_Ishim Motala_HG MA1 0.0124 2.490 33656 32830 632149
result: Mbuti Ust_Ishim Motala_HG Iberia_EN -0.0188 -5.662 42987 44630 832787
result: Mbuti Ust_Ishim Motala_HG Anatolia_Neolithic -0.0198 -6.786 46541 48416 885195
result: Mbuti Ust_Ishim Motala_HG Satsurblia -0.0194 -4.038 32958 34264 617481
result: Mbuti Ust_Ishim Motala_HG Karitiana 0.0082 1.946 49089 48289 886806
result: Mbuti Ust_Ishim Motala_HG Han 0.0074 1.901 50700 49958 886806
result: Mbuti Ust_Ishim Motala_HG Papuan 0.0016 0.306 51015 50851 886805
result: Mbuti Ust_Ishim Karelia_HG AfontovaGora3 0.0088 1.301 10522 10339 223655
result: Mbuti Ust_Ishim Karelia_HG MA1 0.0182 3.049 31795 30660 612850
result: Mbuti Ust_Ishim Karelia_HG Iberia_EN -0.0112 -2.358 42916 43885 810669
result: Mbuti Ust_Ishim Karelia_HG Anatolia_Neolithic -0.0123 -2.890 45975 47124 857439
result: Mbuti Ust_Ishim Karelia_HG Satsurblia -0.0114 -1.923 31874 32610 598661
result: Mbuti Ust_Ishim Karelia_HG Karitiana 0.0161 3.122 46379 44911 859121
result: Mbuti Ust_Ishim Karelia_HG Han 0.0145 3.014 48860 47465 859121
result: Mbuti Ust_Ishim Karelia_HG Papuan 0.0086 1.475 49589 48740 859120
result: Mbuti Ust_Ishim AfontovaGora3 MA1 0.0177 2.357 7774 7503 169948
result: Mbuti Ust_Ishim AfontovaGora3 Iberia_EN -0.0176 -3.114 11940 12369 228224
result: Mbuti Ust_Ishim AfontovaGora3 Anatolia_Neolithic -0.0194 -3.662 12119 12598 231975
result: Mbuti Ust_Ishim AfontovaGora3 Satsurblia -0.0166 -2.283 8127 8401 159389
result: Mbuti Ust_Ishim AfontovaGora3 Karitiana 0.0053 0.919 11634 11510 232022
result: Mbuti Ust_Ishim AfontovaGora3 Han 0.0061 1.136 12635 12482 232022
result: Mbuti Ust_Ishim AfontovaGora3 Papuan -0.0002 -0.027 12663 12667 232021
result: Mbuti Ust_Ishim MA1 Iberia_EN -0.0282 -5.714 33539 35484 624535
result: Mbuti Ust_Ishim MA1 Anatolia_Neolithic -0.0303 -6.608 36888 39197 679336
result: Mbuti Ust_Ishim MA1 Satsurblia -0.0287 -4.782 25607 27118 476924
result: Mbuti Ust_Ishim MA1 Karitiana -0.0031 -0.593 36390 36619 683873
result: Mbuti Ust_Ishim MA1 Han -0.0036 -0.699 38616 38893 683873
result: Mbuti Ust_Ishim MA1 Papuan -0.0088 -1.466 38793 39481 683872
result: Mbuti Ust_Ishim Iberia_EN Anatolia_Neolithic -0.0017 -0.677 44273 44423 872150
result: Mbuti Ust_Ishim Iberia_EN Satsurblia -0.0014 -0.287 32457 32545 606261
result: Mbuti Ust_Ishim Iberia_EN Karitiana 0.0250 5.713 50591 48126 873267
result: Mbuti Ust_Ishim Iberia_EN Han 0.0244 6.035 51137 48700 873267
result: Mbuti Ust_Ishim Iberia_EN Papuan 0.0185 3.605 51294 49426 873266

Chad said...

result: Mbuti Ust_Ishim Anatolia_Neolithic Satsurblia 0.0009 0.211 36435 36370 669937
result: Mbuti Ust_Ishim Anatolia_Neolithic Karitiana 0.0266 6.841 56628 53690 961098
result: Mbuti Ust_Ishim Anatolia_Neolithic Han 0.0260 7.300 57231 54325 961098
result: Mbuti Ust_Ishim Anatolia_Neolithic Papuan 0.0203 4.190 57468 55182 961097
result: Mbuti Ust_Ishim Satsurblia Karitiana 0.0256 4.937 39241 37284 676358
result: Mbuti Ust_Ishim Satsurblia Han 0.0253 5.178 39851 37882 676358
result: Mbuti Ust_Ishim Satsurblia Papuan 0.0203 3.384 40043 38452 676357
result: Mbuti Ust_Ishim Karitiana Han -0.0004 -0.112 50726 50764 970156
result: Mbuti Ust_Ishim Karitiana Papuan -0.0061 -1.281 53794 54450 970155
result: Mbuti Ust_Ishim Han Papuan -0.0058 -1.283 53323 53941 970155

Chad said...

Either everyone has it since the UP, or possibly Ust-Ishim ancestry was drained out of Northern Eurasian for the most part, just as those "goofy" qpAdm models showed. Something odd...

Karl_K said...

@Olympus Mons

"Kura-araxes (or previous Sioni) or Maikop are too late"

Not at all. Probably too early for being Bell Beaker ancestors.

Imagine for just a moment that some groups in the early Bronze Age developed both a useful technology, and methods of travel that were much quicker than previous methods. These could be things similar to horses or boats.

In that case, perhaps it would take only weeks or months to get from the steppe to Iberia, if that had been their goal and they had a map. It could even happen without an accurate goal or map if they had been totally misinformed.

The ability to move quickly changed everything. It's like determining the genetics of an ancient 'word of mouth' gold rush.

Krefter said...

@Davidski,

I got one quick D-stat. I'm making a column for MA1 via stats you did a while go that don't include Anatolia_Neolithic. I think inorder to help understand what Basal Eurasian is we'll have to use lots of East Asians and all Stone age North Eurasians as columns.

(Chimp Anatolia_Neolithic)(Mbuti MA1)

Anonymous said...

@David

"I can combine them, although not sure if that'll help. Which stats do you want? I can only run up to 30 per request though."

Sorry to be so obsessive about it, but the mtDNA of Paglicci71 makes it such an attractive candidate for a proto-Villabruna. I realize that it is indeed a silly longshot. Furthermore, I hope I am doing the D-stat requests right.

Anyway, what would we want to know? If it is especially related to Villabruna. How it is related to older genomes. And if it is especially related to the Magdalenians.

Mbuti Kostenki14 Paglicci_combined Vestonice16
Mbuti Kostenki14 Paglicci_combined KremsWA3
Mbuti Kostenki14 Paglicci_combined Ostuni1
Mbuti Kostenki14 Paglicci_combined Ostuni2
Mbuti Kostenki14 Paglicci_combined Muierii2
Mbuti Kostenki14 Paglicci_combined GoyetQ116-1

Gravettian with mtDNA U:

Mbuti Kostenki14 Paglicci_combined Goyet2878-21

The Magdalenia.

Mbuti Paglicci_combined ElMiron Vestonice16
Mbuti Paglicci_combined ElMiron Villabruna
Mbuti Paglicci_combined ElMiron GoyetQ116-1
Mbuti Paglicci_combined ElMiron KremsWA3
Mbuti Paglicci_combined ElMiron Paglicci133

Mbuti Paglicci_combined GoyetQ-2 Vestonice16
Mbuti Paglicci_combined GoyetQ-2 Villabruna
Mbuti Paglicci_combined GoyetQ-2 GoyetQ116-1
Mbuti Paglicci_combined GoyetQ-2 KremsWA3
Mbuti Paglicci_combined GoyetQ-2 Paglicci133

WHG

Mbuti Paglicci_combined Villabruna Vestonice16
Mbuti Paglicci_combined Villabruna ElMiron
Mbuti Paglicci_combined Villabruna GoyetQ116-1

Mbuti Paglicci_combined Villabruna AfontovaGora3
Mbuti Paglicci_combined Hungary_HG AfontovaGora3

With the explicit disclaimer that this is an experiment which possibly lumps together two non-related genomes.

Ryukendo K said...
This comment has been removed by the author.
Krefter said...

@RK,
"If they get more ancestry from a brother clade of Vestonice than Villabruna, then they may have increased drift with Vestonice than with Villabrun"

We don't know yet what the tree relationship between Ice age Europeans is, because we don't have enough genomes. With the data we have available Vestonice and WHG are brother clades. I'm not suggesting Middle Easterners have Vestonice-specific ancestry, I'm suggesting they have ancestry from an unsampled third brother in the Vestonice/WHG clade.

I tend to think Anatolia Neolithic is (Member of Ice age European Family)+Minor WHG.

Look at these stats. They prove Anatolia Neolithic can't be Villabruna+Basal Eurasian.

[(Chimp, Loschbour)(Mbuti, Villabruna)-(Chimp, Anatolia Neolithic)(Mbuti, Villabruna)]
Subtracted by
[(Chimp, Loschbour)(Mbuti, Han)-(Chimp, Anatolia Neolithic)(Mbuti, Han)]

Result: 0.1182.

The non_Basal Eurasian side of Anatolia Neolithic can't even be as close to WHG as ElMiron is. Here's the result when you replace ElMiron with Villabruna.

Result: 0.57

When I replace Loschbour with Vestonice.

Result: 0.0114

If we assume Basal Eurasian exists and it has as great of a reduction in affinity to West Eurasians as to East Asians, then Anatolia_Neolithic's Non-Basal Eurasian side is about as close to WHG as Vestonice is.

Considering Anatolia_Neolithic certainly has WHG admixture, this means a population maybe that separated from WHG 40,000 years ago, is the main non-Basal Eurasian ancestor of Middle Easterners.

Last thing. We have to take into consideration Crown Eurasian ancestry that isn't West Eurasian in the Middle East. This and Basal Eurasian could be what's reducing Middle Easterner's affinity to Stone age North Eurasians. This type of ancestry would be impossible to detect. If it exists, it would mean Middle Easterner's West Eurasian ancestry is more similar to WHG than Kostinki or Vestonice.

Krefter said...

-----------------

It may be hard to believe but mtDNA UK might be the ONLY maternal lineage of the Paleo-European family. There is lots of mtDNA U native to the Middle East: U1, U3, U7, K(K8b2), all of which haven't been found in Paleo Europeans. Plus there's Y DNA J in the Middle East, all of this is consistent with them having decent from a relative of Paleo Europeans(that wasn't WHG who was mtDNA U5b and Y DNA I2).

The rest of the maternal/paternal lineages(IJ, RQ, C1) of West Eurasia might be of Basal Eurasian, unknown Crown Eurasian, and unknwon Basal West Eurasians(very distant relatives of Paleo European clade or Paleo European+ANE clade).

Anonymous said...

@David

I made some mistakes. Mbuti Kostenki14 should be Mbuti Villabruna. And as we don't have much we should probably use genomes with the highest coverage. However, I think more than one ElMiron sample because Magdalenian is WHG related.

So:
Mbuti Villabruna Paglicci_combined Vestonice16
Mbuti Villabruna Paglicci_combined KremsWA3
Mbuti Villabruna Paglicci_combined Ostuni1
Mbuti Villabruna Paglicci_combined Ostuni2
Mbuti Villabruna Paglicci_combined Kostenki14
Mbuti Villabruna Paglicci_combined GoyetQ116-1

Mbuti Paglicci_combined ElMiron Vestonice16
Mbuti Paglicci_combined ElMiron Villabruna
Mbuti Paglicci_combined ElMiron GoyetQ116-1
Mbuti Paglicci_combined ElMiron KremsWA3
Mbuti Paglicci_combined ElMiron Paglicci133

Mbuti Paglicci_combined GoyetQ-2 Vestonice16
Mbuti Paglicci_combined GoyetQ-2 Villabruna
Mbuti Paglicci_combined GoyetQ-2 GoyetQ116-1
Mbuti Paglicci_combined GoyetQ-2 KremsWA3
Mbuti Paglicci_combined GoyetQ-2 Paglicci133

Mbuti Paglicci_combined Villabruna Vestonice16
Mbuti Paglicci_combined Villabruna ElMiron
Mbuti Paglicci_combined Villabruna GoyetQ116-1

Mbuti Paglicci_combined Villabruna AfontovaGora3
Mbuti Paglicci_combined Hungary_HG AfontovaGora3

Davidski said...

rk,

Lack of markers might be an issue here.

https://drive.google.com/file/d/0B9o3EYTdM8lQYUdOQXhtMGVGaFk/view?usp=sharing

Krefter,

result: Chimp Anatolia_Neolithic Mbuti MA1 0.3575 96.219 61028 28883 655483

epoch,

result: Mbuti Kostenki14 Paglicci Vestonice16 0.0330 2.796 2651 2482 48914
result: Mbuti Kostenki14 Paglicci KremsWA3 0.0429 1.963 879 806 15526
result: Mbuti Kostenki14 Paglicci Ostuni1 0.0022 0.116 1117 1112 22549
result: Mbuti Kostenki14 Paglicci Ostuni2 0.0126 0.158 51 50 1144
result: Mbuti Kostenki14 Paglicci Muierii2 0.0622 1.835 317 280 5721
result: Mbuti Kostenki14 Paglicci GoyetQ116-1 0.0076 0.634 2839 2796 51302
result: Mbuti Paglicci ElMiron Vestonice16 0.0459 3.342 2210 2016 39819
result: Mbuti Paglicci ElMiron Villabruna 0.0237 1.757 2221 2118 43942
result: Mbuti Paglicci ElMiron GoyetQ116-1 0.0130 0.887 1997 1946 39907
result: Mbuti Paglicci ElMiron KremsWA3 0.0575 2.417 707 630 12528
result: Mbuti Paglicci GoyetQ-2 Vestonice16 0.0627 1.605 248 219 4317
result: Mbuti Paglicci GoyetQ-2 Villabruna -0.0451 -1.228 239 262 4849
result: Mbuti Paglicci GoyetQ-2 GoyetQ116-1 -0.0198 -0.501 219 228 4520
result: Mbuti Paglicci GoyetQ-2 KremsWA3 0.0436 0.667 78 71 1441
result: Mbuti Paglicci Villabruna Vestonice16 0.0322 2.604 2714 2545 48869
result: Mbuti Paglicci Villabruna ElMiron -0.0237 -1.757 2118 2221 43942
result: Mbuti Paglicci Villabruna GoyetQ116-1 -0.0134 -1.066 2732 2806 50380
result: Mbuti Paglicci Villabruna AfontovaGora3 -0.0734 -3.695 934 1082 19399
result: Mbuti Paglicci Hungary_HG AfontovaGora3 -0.0489 -2.517 902 995 18460

Davidski said...

epoch,

How am I supposed to run this?

Mbuti Paglicci_combined ElMiron Paglicci133

Ryukendo K said...
This comment has been removed by the author.
Anonymous said...

@David

Mbuti Paglicci ElMiron Paglicci133

Paglicci133 is according to the paper an outlier not very easily associated with any of their clusters. Or did you add Paglicci133 as well? If it doesn't work, never mind.

Anyway, it does look like there is a tiny affinity with Vestonice16. I'd say nothing surprising. :-/

Anyway, would it be asking too much to do this as well?

Mbuti Villabruna Paglicci_combined Vestonice16
Mbuti Villabruna Paglicci_combined KremsWA3
Mbuti Villabruna Paglicci_combined Ostuni1
Mbuti Villabruna Paglicci_combined Ostuni2
Mbuti Villabruna Paglicci_combined Kostenki14
Mbuti Villabruna Paglicci_combined GoyetQ116-1

Many thanks for the effort, anyway. Really appreciated.

Olympus Mons said...

@Chad Rohlfsen

Your comment at May 15, 2016 at 4:30 PM is just brilliant. If you want o be prick, I mean. you a take out of my 50 page all text thesis a reference to a post which I target as my first stance on this issue and the first digital record liking Shulaveri to the stock of R1b that made the later Bell beaker.

Oh, and the Yamnaya from Iberia, its just brilliant. If you want to be disingenuous. Its like me saying: Chad, really UFO carrying R1b from Yammanya into IBeria. UFOs, really man?

It would really represent your opinion or any of the Steppe fans, would it?

Congrats.

Davidski said...

epoch,

There's no Paglicci71 in the dataset. The Paglicci sample in the D-stats above is made of Paglicci 108 and 133. These are the only Paglicci samples in the dataset.

Please look at the ID sheets more carefully.

Ryukendo K said...
This comment has been removed by the author.
Olympus Mons said...

@ Karl_K

We get attached to little letters of Str mutations here, and don’t really look out the “window” any more. And out there is raining! :- )

Sioni, Kura or Maykop really have nothing Bell beaker (or proto bell beaker). Nothing, Not architecture, no lifestyle, nor burials… it’s just not them. Now… The shulaveri, one needs to be blind no to see the similarities. Let me tell you, out of dozens, the latest one (ht to Bellbeakerblogger, see is post).

Spelt is a specific wheat that is first “seen” in Europe with BB in Germany. Actually not true because I have shown papers that it showed up also in Northern Portugal (which I think is Club wheat) by the same time. But that is not the point. Spelt is an Hybriazation of club wheat (Hexaploid) and Emmer (tetraploid) made to survive harsh environments, but different in Europe as the one in Asia.
It only happened twice that Hybridization. One found at the heartland of the Shulaveri in Arukhlo by 5k bc) and the second in Europe with Bell beakers. And you know what. Genetic analysis points to the fact that the Club wheat part was common to both, just the Emmer (Neolithic trough out Europe) was different.

This is one, one of several of this connections. Just the lasted

If you have the time, read my thesis. no hurry.

http://blogs.sapo.pt/cloud/file/eb6b52b82097d41dfa0e5797a2fa7945/olympusmons/2016/From%20Shulaveri%20to%20Bell%20beaker.pdf

Olympus Mons said...

Karl_K,
Just to clarify something. Maykop will have R1b, just that will not be a predominant and dominant one. Someone else was in the driving seat.
Sioni or Kura will have lots and lots of J2/J1. R1b will be remnant. All this to clarify that, in my opinion, only sampling Shulaveri and Aratashen will you find R1b (M269) and on the later stages (L23).
Sampling Merimde/EL-omari you will find r1b with L51 (probably) and Sampling late neolithic IBeria (lets see what Max Plank samples shows) but I bet P310 is the mark of Iberia R1b by 3000 BC, so R1b bell beakers will have it.

Davidski said...

rk,

I didn't save the numbers of markers, but almost all of the stats are based on well over 200K SNPs, which means they should be fine. The only stats that are often based on less than 200K are those involving AfontovaGora3, so they may not be very reliable.

The Oase1 stats are probably not worth including in the datasheet due to the low numbers of markers used.

Alberto said...

@Rob

Yes, exactly, I think he differences in AG3 vs. MA1 must be mediated by different CHG populations. So for example if taking Kotias' model earlier in the thread it would be something like Basal_Eurasian + Kostenki14 + AfontovaGora3, then maybe another type of CHG will be Basal_Eurasian + Kostenki14 + MA1. The second type would be a yet unsampled one that seems to represent a more SC Asian type, that probably entered the steppe later than the ones represented by Kotias and made it into southern and western Europe with Bell Beakers, while it didn't make it into CWC or most of northern Europe.

Or who knows what exactly. But I think that the MA1 presence does mean something. Otherwise I don't know why would it show up instead of AG3, which is more modern and a better reference for both Kotias and Karelia_HG.

Davidski said...

Olympus,

Not sure if you're aware of this, but the Middle Bronze Age Armenians that belong to R1b carry a shit load of steppe admixture. In fact, they have more steppe ancestry than any modern population in the region, and they just happen to carry R1b-M269, and indeed M269 lineages related to those of Yamnaya.

That's not to say that there won't be any R1b in pre-Iron Age West Asia not linked to steppe ancestry so clearly, but I'm betting that there won't be a lot of it, and it won't be M269 anyway.

So I reckon if you want to posit the origins of Bell Beakers in West Asia, then it has to be the steppe first, then West Asia, and then across the Med to Iberia.

I entertained this type of scenario back in the day when we had very little ancient data to work will. But now it all seems very far fetched and unnecessary to me. Really, if we're honest and objective, then it looks like an explosion from the steppe to the west, east and south during the Bronze Age, and this is where most West Asian M269 comes from.

No doubt that Bell Beakers are closely related to the above mentioned Middle Bronze Age Armenians, but in all likelihood they're cousins stemming from parallel lines from the steppe.

Alberto said...

@RK

Here are the diffs for the Basal_All_Equal runs above for Anatolia_Neolithic and Kotias:

https://docs.google.com/spreadsheets/d/1IsxRk-Bb3HHzbMoKxqPSqn3zmwSDRIlrGFsp4BQx0qw/edit?usp=sharing

Olympus Mons said...

@Davidski.
If you read my thesis that is clear - Yes, to me (and everyone else I suppose) R1b Shulaveri were exogenous to the southern Caucasus. And most probably from the steppe, for sure. I dwell endlessly if Volga or Dnieper basin and so forth. And yes R1b current Armenia is “shulaveri”. And also those part of northern Iran R1b I think. And also Maykop (at least part)… but those are the ones that stood behind. So M269 and probably L23 with buck loads of Steppe. Makes sense?

Let me try to tell the whole story in a Paragraph:
8k bc they were in steppe, 7K bc arrive at South Caucasus and were Shulaveri-Shomu, by end 5k bc “kicked out dogs” by Ubaid J2/j1 that made Kura-araxes culture and south Mesopotamia big civilizations, and by 4.9k bc were at Tell Tsaf for a while, and also in Shiqmin/ Saharonim, by 4.7K bc as Merimde and El-omari, by 4K bc disappeared again and those who stood were part of Maadi (and some part of Later Badarian), by 3.7k bc roaming in tassilli and related places as Bovidean phase II (white people pastoral phase), by 3.5k Bc pushed way by 5.9K kiloyear event and sahara desert birth pushed into south Iberia, by 3K bc “huge” cities of porto torrão (500ha!!) , by 2.9k bc at Zambujal and Vnsp and part of Perdigoes site shrine, by 2,7K bc roaming into rest of Europe (they never stopped actually).
So, M269, going L23, going L51, going P310 by these route.

What is funny is,
By 1K bc in Portugal, as the Oestrimni (people from farwest), they were telling everyone who asked (like periplus the greek mariner): “we are the people that have been living here for long, but coming from a faraway land which we had to flee because of an attack of serpents” (actually the Ubaid/Uruk). That is why Lisbon was called Ophiusa like the Ubaids, by the greeks.

See, simple. Yes, they come from the steppe. But the route to become Bell beakers and R1b Europe was the one I described. :-) - Wanna bet?

Anonymous said...

@David

"There's no Paglicci71 in the dataset. The Paglicci sample in the D-stats above is made of Paglicci 108 and 133. These are the only Paglicci samples in the dataset.

Please look at the ID sheets more carefully."

The paper has Paglicci71 with the same mtDNA as El Miron but with less than 4000 SNPs. While they mentioned not using it I thought they were up for download.

Sorry for the bother.

Davidski said...

Olympus,

You seen this?

https://drive.google.com/file/d/0B9o3EYTdM8lQODA2N19MUGZ5YU0/view?usp=sharing

If your theory was correct, we'd expect a migration edge from the ancient Armenians to Corded Ware. But it runs from (or between) Copper Age Europe and something close to Afanasievo. The Armenians are irrelevant.

So for you everything hinges on those Iberian Bell Beakers. But admittedly, things, as they currently stand, aren't looking good.

Full blog entry here...

http://eurogenes.blogspot.com.au/2015/07/population-genomics-of-early-bronze-age.html

Unknown said...

"Really, if we're honest and objective, then it looks like an explosion from the steppe to the west, east and south during the Bronze Age, and this is where most West Asian M269 comes from."

R-M269 is found in Ethiopia. The Eurasian haplogroup J lineages in Ethiopia are probably Neolithic but don't know about the R-M269.

Davidski said...

Yeah, R-M269 is native to the Americas and Africa. And Yamnaya is from South Asia.

Anyone care to come up with something more off the wall than what we've seen in recent weeks in the comments here?

Unknown said...

I don't think that R-M269 is native to Africa. Clearly Ethiopia got it from West Asia, and you may well be correct that West Asia in turn got it from the Steppe, I don't know. There are various dates given for Eurasian admixture in the Horn of Africa, the R-M269 there could have entered <3kya, I really don't know.

André de Vasconcelos said...

@OM
Your interpretation of the Ora Maritima is not what is defended by conventional historians, it does not point to Uruk but rather NW Iberia. And Lisbon wasn't called Ophiusa.

I suggest you read professor Mattoso's take on the subject.

Nirjhar007 said...

Will you guys stop acting and proposing comments like kindergarten kids?.

We simply don't have enough aDNA coverage from Eurasia to come up with anything!...

Davidski said...

We simply don't have enough aDNA coverage from Eurasia to come up with anything!

Wishful thinking.

Olympus Mons said...

@Andre de Vasconcelos,
Historians know shit about that story. Its assumed that there is no factual and historical grounds to work it out. Thatis what i am giving now.
Two things André: matosso knows shit about it, like anyone else. A historical ground is whati am giving now.
Second you jave this habit of undermine me with a phrase or two that only represents a anterior cingulate cortex "error detection" that is meaningless for the subject. If you have something alternative to say just say it.
What ora maritima interpretation are you talking about regarding an issue nobody could find an explantion for...

Davidski said...

I've got a new sheet. Instead of running D(Chimp,X)(Mbuti,Outgroup), I ran D(Chimp,Outgroup)(Mbuti,X). The results are usually very similar, but never identical.

https://drive.google.com/file/d/0B9o3EYTdM8lQc2pVdVZ1MTlsQ0k/view?usp=sharing

Olympus Mons said...

@André,
You know that anyone can just google "Lisbon Ophiusa" and there are tens of references, right?

Nirjhar007 said...

Wishful thinking

Yes exactly, that what you mostly suffer from , thanks for speaking it out.

Olympus Mons said...

@Davidski,
That is because you don’t consider bell beaker as I think they were.
Bell beaker is the spread of R1b, not the Caucasus Adna or anything steppe or another adna for that matter. Bell beaker is a male spread. So when you start looking at bell beakers in Iberia you will see a whole lot of EEF in them. But as they picked up women along the way, it really gets diverse in terms of aDdna. Because Bell beaker were the R1b new wolves in the prairie and they pick up women like sex addicts and those women had fathers that were very local ir even not (roaming others from other places). Yamnaya had R1b Shulvari because a lot of them jump the moutains to flee Ubaid (4500bc). Eastern Anatolia should also have it, as Northern western Iran and so forth.
To be clear:
*Bell beaker are from Zambujal Portugal but not R1b.
* R1b Bell Beakerish didn’t really stop in Portugal for long. At the same time they were coalescing in South Iberia and then in Tagus basin area, some were crossing into south France, North Italy, Switzerland. So finding R1b ( M269/P310) in say south France by 3300bc would be all normal.
*R1b/bell Beaker was a male thing, and they took some women but mostly the get women from others whenever they could.

Note: Do you also consider that some samples showing traces of SSA is a software error or do you think there are grounds to think that it is remnant of they taking women that had some residues of SSA?

Olympus Mons said...

shit...
"*Bell beaker are from Zambujal Portugal but not R1b."
what I meant was.
*Bell beaker are from Zambujal Portugal but not the point of origin of the R1b as a stock of people for Europe."

Same stock that stooped in Portugal were on their way to cross the Pyrenees.

Olympus Mons said...

Nirjhar007
"Wishful thinking"

you know that its basically what applies to us all, right? Or do you have the idea that is anything else?
In this issues addressed here we are that specific point in science where its proven that 90% of what is postulated will turn out completely wrong or highly incorrect. There is no rebuttal or conclusive presumption in this little tiny bits of DNA apart from a lot of inferences.


Anonymous said...

@David

Well, it sure shows one thing. We know Paglicci133 clustered with the Vestonice cluster. Adding Paglicci108 didn't change that. Which means that the Gravettians in Italy remained Vestonice-like and may not be the source of the later WHG. Gravettian until 28.000 years ago.

Mbuti Paglicci ElMiron Vestonice16 0.0459 3.342 2210 2016 39819
From the Paper (ElMiron, Vestonice; Paglicci133, Mbuti) Z=-3.5

So maybe Paglicci71, based *only* on mtDNA, represents a new population, popping up around the same time as El Miron. They share the same mtDNA. Allthough D(Villabruna, Vestonice16; Paglicci133, Mbuti) in the paper is Z=-2.7, which may point to a slight admixture either way, as paper also models (10%).

huijbregts said...

Bell Beaker admixture

@ Chad
Yesterday I commented on a post of Alberto.
IMO his models were not parsimonious and to illustrate that I posted a model with Bell_Beaker_Germany <- (Corded_Ware_Germany,Germany_MN,....).
Next you threw cold water on my supposed claim that bell_Beakers descended from Corded_Ware.
That was not what I had stated, but sure, I might have expressed myself more carefully
Alberto has a private theory and to explore it, he has used an atypical selection from the available genomes.
My point was that his interpretation was not parsimonious, not even within this atypical selection (I just added German_MN for an obvious geographical reason).

If my intention had been to explore the possible ancestry of the Bell_Beakers, I would have processed a different selection.
As I got curious, I did this exploration and here is my result:
____Bell_Beaker_Germany____:
"Yamnaya_Kalmykia" 37.45
"Hungary_EN" 27.75
"Germany_MN" 15.75
"Bichon" 5.65
"Motala_HG" 5
"Yamnaya_Samara" 4.6
"Saami" 2.3
"Kotias" 1.5
"AfontovaGora3" 0
"Corded_Ware_Germany" 0
"MA1" 0
distance%=0.2194

Olympus Mons said...

@huijbregts
"That was not what I had stated, " ...

Maybe we can both start by asking Chad not to do that, because its rude?

Alberto said...

My atypical selection could never really be parsimonious, because it's clear that Bell Beakers never got direct input from a 24,000 year old genome from Lake Baikal. But it was intentional, to explore some genetic difference.

Your model is a confirmation of mine. Bell Beakers are better modeled with Yamnaya_Kalmykia, that has a signal of MA1-related (?) ancestry, while Corded Ware is better modeled with Yamnaya_Samara, that lacks such signal. We're seeing the same phenomenon (whatever it really means, if anything at all).

Matt said...

@ Davidski, thanks for doing that new sheet with D(Chimp,Outgroup)(Mbuti,X). Using PC on the differences between equivalent columns in the two spreadsheets, the main outliers look like Spanish_Cataluna, Turkmen, Turkish, Spanish_Castilla_y_Leon, Srubnaya, Tajik_Ishkashim, Tajik_Rushan and Tubalar. (These ones suggest some degree of error in the difference.) The others more or less fall along a single vector, although the Syrian sample is an outlier on this vector.

The populations that tend to be higher on the D(Chimp,Outgroup)(Mbuti,X) tend to be recent Mediterranean populations (with some exceptions - Andronovo is high for some reason), while the opposite case, higher on the D(Chimp,X)(Mbuti,Outgroup) tend to be ENA and Upper Paleolithic.

Comparing specific stats for the columns of BedouinB and other Near Eastern populations for the top matches:

http://i.imgur.com/uUC00fm.png

There aren't huge differences, but you can see that in the stats with (Chimp,Outgroup) compared to (Mbuti,Outgroup), the Early European Farmers (and their European descendents) tend to drop down, while the Middle Eastern populations tend to move up. I think this is probably because the use of Mbuti next to the outgroup tends to isolate the non-African like ancestry of the Near Eastern populations more, bringing them relatively closer to EEF (and also generally increasing the stat). However, the EEF are still high matches. The changes are not as drastic as I expected they might have been.

Olympus Mons said...

@Alberto,
to me is clear (I mean, as inference of course. :- ) )
So, CW (as R1a) stock was up there in volga as Samarra and the source of "my shulaveri" was actually Kalmykia (ish). What a bummer, they just cross the mountains? how disappointing. Expected they had come form afar. :-)
Yamnaya is more a problem than a solution. Lots of even "shulaveri" stock od R1b crossed back and were part of Yamanaya and Maykop. so not the best samples for clearing origins of bell beaker.

However I think the end game for me will not be long - See, Bertille Lyonnet colected samples from Mentesh tepe (5800bc) in late 2015 and sent to "Ancient DNA studies (under the direction
of E. Heyer and C. Bon, Museum National d'Histoire Naturelle,
Paris)"...

So I am waiting on them for a while now. I assume they will go for Y and Mtdna.
All references to that expedition are: "Currently the results of the archaeological campaign are being studied and are not communiquables."


So will see.

Matt said...

@ Davidski, also, thanks for the Villabruna column. Comparing to La Brana:

http://i.imgur.com/7r8yE6P.png

They're mostly correlated. There are some logical outliers - El Miron, GoyetQ116-1 on the La Brana side and a slight tendency for East Asian to prefer La Brana and modern European to prefer Villabruna - but also some odd outliers which I can't explain - Turkmen, Turkish, Srubnaya, Tajik Rushan, Syrian and some of the Spanish, who also tended to come out as outliers in the D(Chimp,X)(Mbuti,Outgroup) vs D(Mbuti,X)(Chimp,Outgroup) comparisons I was doing.

Matt said...

Re: the multiple regression method I was using above to simulate Basal Eurasian ghost, I was messing around and realized another thing you could do with it is use PC data on the d-stats to form another multiple regression, so I could effectively pick a place in the PCA space that no real population we have occupies, and then transform it back into a set of stats. (This is probably an unnecessarily awkward way of doing it though).

So, in case anyone wants to test, using that, I simulated up a few populations that we like the WHG-EHG on PC1 (that separates WHG-SHG-EHG-Siberian HG together from Near East), but like Anatolia Neolithic on PC2 (that splits off Anatolia Neolithic related from everyone else), to simulate a theoretical ME UHG population:

Columns Anatolia_Neolithic2 BedouinB2 Cypriot2 Dai2 Denisovan Druze2 Han2 Iberia_Chalcolithic Iranian_Jew2 Karitiana2 LaBrana1 Mansi2 Motala_HG2 Munda2 Papuan2 Samara_HG Satsurblia South_Indian2 Yoruba

SimulatedUHG1 0.4223 0.3844 0.4083 0.3431 -0.5814 0.4021 0.3443 0.4367 0.4005 0.3605 0.4544 0.3875 0.4537 0.3459 0.3082 0.427 0.3854 0.3594 0.0956

SimulatedUHG2 0.4165 0.3791 0.3995 0.3367 -0.5785 0.3931 0.3382 0.4377 0.3907 0.349 0.4731 0.3774 0.4582 0.3398 0.3057 0.4178 0.371 0.3515 0.0955

SimulatedUHG3 0.4117 0.3748 0.3922 0.3313 -0.5761 0.3857 0.3331 0.4385 0.3826 0.3396 0.4885 0.3691 0.4618 0.3348 0.3038 0.4102 0.359 0.345 0.0955

The difference between three simulated is that for first one I didn't control for the PC3, which relates to a shared factor loading towards Kotias and Samara and away from UP Europeans, while for the latter two, I controlled to make the simulated UHG more like Mesolithic and UP Europeans.

Also, a simulated Basal Eurasian using the same regression on PCA method, lining up against BedouinB on PC1 and Anatolia_Neolithic on PC2:

(Columns as above)

Basal Sim 0.4156 0.3927 0.4057 0.3188 -0.595 0.4003 0.3188 0.399 0.3997 0.3163 0.3483 0.343 0.351 0.325 0.2829 0.3442 0.378 0.3404 0.0975)

Obviously these are "fantasy populations" and just based on picking points on the PCA, with no real basis behind then, so definitely don't overinterpret them as real, but in case anyone wanted to try these in nMonte, Alberto and others.

Krefter said...

@Davidski,

Can you run these D-stats.

https://docs.google.com/spreadsheets/d/1HzxBkIrlUFhbbA8Z_eiN8vjpMpe0klNxaQdE12UQlys/edit#gid=0

These are vital to determine what the Basal Eurasian signal is. I don't have enough D-stats on my disposal to answer important questions.

With what I do have it looks there's definitely Crown(xEast, West) Eurasian in EEF and CHG. If we are to assume Basal Eurasian reduces East Asian affinity as much as it does West Eurasian affinity, then EEF's non-basal Eurasian side is more distant from Kostinki14 than Ma'lta boy is and is a very very distant relative of WHG. The only way to explain this IMO is Crown(xEast, West) Eurasian ancestry.

Davidski said...

Here...

https://drive.google.com/file/d/0B9o3EYTdM8lQTWRqT2U0LVRkZk0/view?usp=sharing

No Onge though.

Ryukendo K said...
This comment has been removed by the author.
Davidski said...

There's no perfect method. Double outgroup D-stats are the best input into nMonte that I can think of. F3-stats wouldn't work at all unless I got rid of all the missing markers, which wouldn't be practical.

It's not like qpGraph is perfect. Not even close. So we have to work with the limitations in mind.

Ryukendo K said...
This comment has been removed by the author.
Davidski said...

I've updated the post.

rk,

Which stats would you like to see run on the same number of markers?

Ryukendo K said...
This comment has been removed by the author.
Davidski said...

OK, here. But I had to drop AG3.

https://drive.google.com/file/d/0B9o3EYTdM8lQY0s0anJJZWRKbEE/view?usp=sharing

Alberto said...

@Matt

I tried the latest Basal Eurasian simulation. It still seems to show too high Basal Eurasian, though that's only if we accept the estimates for EEF and Kotias in the literature as correct.


Anatolia_Neolithic
"Basal_Sim" 69.85
"Loschbour" 25.35
"AfontovaGora3" 4.8
"Kostenki14" 0
"MA1" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Masai_Kinyawa" 0
"Kotias" 0

BedouinB
"Basal_Sim" 71.7
"MA1" 14.1
"Masai_Kinyawa" 6.7
"Loschbour" 4.85
"AfontovaGora3" 2.65
"Kostenki14" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Kotias" 0

Alberto said...

I checked many populations with Yamnaya_Samara and Yamnaya_Kalmykia, and they correlate perfectly with the AG3 and MA1 results. It seems that there is some admixture in Yamnaya_Kalmykia (and Afanasievo) that is missing in Yamnaya_Samara. And such admixture is present in Bell Beakers, South Europeans and Asians, but not in CWC, North Europeans and Siberians.

Aram said...

Some archaeological news
""Rare 5,000-year-old kurgan-type tumulus from the Bronze Age unearthed in Istanbul""

http://www.dailysabah.com/istanbul/2016/05/16/rare-5000-year-old-kurgan-type-tumulus-from-the-bronze-age-unearthed-in-istanbul

Olympus Mons said...

@Alberto,
Asians... which populations have you used?

Davidski said...

I'd hate to be the fly in the ointment here, but the German Bell Beakers are just a mixture of local Middle Neolithic farmers and steppe invaders.

Using the new sheet that I posted, I basically got perfect results for Corded Ware and Afanasievo using the most plausible reference samples, plus as few odd choices just to see what would happen.

Afanasievo
"Yamnaya_Samara" 94.65
"Kotias" 2.65
"Ulchi" 1.8
"Germany_MN" 0.5
"Armenia_BA" 0.4
"Anatolia_Neolithic" 0
"Iberia_MN" 0
"Karelia_HG" 0
"Loschbour" 0
"MA1" 0
"Motala_HG" 0

"distance%=0.6803 / distance=0.006803"

Corded_Ware_Germany
"Yamnaya_Samara" 64.4
"Germany_MN" 29.25
"Motala_HG" 5.1
"Ulchi" 0.85
"Loschbour" 0.4
"Anatolia_Neolithic" 0
"Armenia_BA" 0
"Iberia_MN" 0
"Karelia_HG" 0
"Kotias" 0
"MA1" 0

"distance%=0.4393 / distance=0.004393"

We can all agree that these are essentially perfect estimates that line up with what we've seen before, both here and in the various recent papers. So this result for the German Bell Beakers should also be correct. The fit isn't quite as good, probably because of deamination (which often shows up as Sub-Saharan ancestry), but it's still pretty damn good. And the algorithm has a lot to choose from, including ancient Iberians and Armenians.

So where's the Iberian and Near Eastern influence?

Bell_Beaker_Germany
"Germany_MN" 50.05
"Yamnaya_Samara" 41.6
"Motala_HG" 6.85
"Loschbour" 0.75
"Ulchi" 0.75
"Anatolia_Neolithic" 0
"Armenia_BA" 0
"Iberia_MN" 0
"Karelia_HG" 0
"Kotias" 0
"MA1" 0

"distance%=0.9254 / distance=0.009254"

Rob said...

Aram

Very interesting.
Looks like an stone cist burial. That Jug looks very "Anatolian Chalcolithic".
I wonder if any parallels can be drawn Arslan tepe ? (Id imagine that the direction one woul dhave to look)

Olympus Mons said...

@Davidsi,
To me it just tell me that the CW spread with a lot more women then the BB. So, BB will always show more local substratum and less Yamnaya (Just wait until there are South Caucasus neolithic DNA!) and CW will carry more steppe.
Does it make sense?

Kristiina said...

Alberto, is that Anatolia Neolithic Barcin 6500-6200 BC? Does that mean that MA1 ancestry entered Nort Africa after Anatolia Neolithic? I would expect that if it arrived before 6000 BC it would be present also in Anatolia Neolithic and not only in modern Bedouins. If MA1 ancestry came after 6000 BC, I would somehow connect it with Southern Central Asia and Iran. In any case, there was a cultural continuum from North Africa to Central Asia and India at least in Arab and Turkic times and contacts may have started very much earlier. For example, Sumerian language may have arrived from an area with considerable MA1 type ancestry percentage as this language has been connected with language families such as Kartvelian, Munda, Dravidian, Uralic, Altaic and Nostratic but not with any southern languages families. Sumerian culture surely radiated to the area where Bedouins formed.

Olympus Mons said...

@Davidski,
"probably because of deamination (which often shows up as Sub-Saharan ancestry)".

Is it really, or is just "probably" because it does not fit your narrative? Honestly. How sure are you of the deamination effect?

Because it all seems like a "fudge factor", or tunning that fits your narrative. And we all know how dangerous that is.

Davidski said...

It doesn't make any sense to put the origins of these Bell Beakers anywhere but North-Central Europe and the Eastern European steppe.

There are no other signals from anywhere else. It is what it is.

Olympus Mons said...

@Kriistina,
"MA1 ancestry entered Nort Africa after Anatolia Neolithic"

Yes, exactly at 4.500 up until 3800 BC. Does it make sense?
:-)

Olympus Mons said...

@davidsi.
I dont get it. If we assume that from the beginning that stock was full of what you call steppe Dna (Which I think part is a spill back into steppe from Caucasus by 4.900bc) so you can not assume that it was a route (the only contention we have here) from Steppe to Europe, of which we actually have no traces (and by that I mean, Architecture, ways of life, etc) , at all.

Davidski said...

If the Bell Beaker package came to North-Central Europe from the Near East via Iberia, then it was just a cultural transmission, which is possible.

It's not parsimonious to suggest that a people basically identical to Yamnaya migrated from the steppe into the Near East, and then via North Africa to Germany, and remained identical to Yamnaya until they got to Germany.

The Corded Ware people obviously moved very quickly into North-Central Europe from north of the Black and/or Caspian Sea. So this looks like the most plausible route into Germany for the steppe ancestors of the German Bell Beakers.

Some of you guys are thinking zebras when hearing horses.

Rob said...

both of you have points there
David is correct in that the steppe ancestry entered BB around the Rhine, via the same route as CWC - ie north European plain

Did a cultural impetus for BB come from Iberia - definitely, and beyond ? Why not

huijbregts said...

@ Davidsky
The fit of the Bell_Beakers is very good once you permit them to meet some farmers up in the Carpathian plain:
____Bell_Beaker_Germany____
"Yamnaya_Kalmykia" 35.15
"Hungary_EN" 24.9
"Germany_MN" 19.6
"Yamnaya_Samara" 7.55
"Motala_HG" 5.2
"Bichon" 4.4
"Kotias" 1.6
"Loschbour" 1
"Ulchi" 0.6
distance%=0.2202

Davidski said...

Ah, I see, I forgot about the Hungarian samples.

Olympus Mons said...

@Davidsi,
its like talking to guys that only see shadows (DNA) regarding colors and shapes. :-)
If not anything else, SPELT, that bellbeakerblogger posted. That Hybridization occur in south Caucasus (Shulaveri heartland) at 5k bc and them found with Bell beaker, in Switzerland... but also in Northern Portugal even earlier.
How do you figure that happen? UFOs :-)

Karl_K said...

"If the Bell Beaker package came to North-Central Europe from the Near East via Iberia, then it was just a cultural transmission, which is possible.

It's not parsimonious to suggest that a people basically identical to Yamnaya migrated from the steppe into the Near East, and then via North Africa to Germany, and remained identical to Yamnaya until they got to Germany."

I am not going to eliminate the possibility that this was just a Yamnaya-like land migration from the east that met up with a Bell Beaker culture somewhere in Western or Central Europe.

But honestly, I think it will be shown to not be anywhere near Germany, and probably within Iberia. And it won't be purely a cultural transmission.

It is probably related to the lactose tolerance allele being picked up somewhere (likely in Basque country).

Why there? Because if the Bell Beaker movement was fueled by lactose tolerance picked up from some smaller Neolithic group, then it would mean that it could outcompete Neolithic groups anywhere else, but much less so at the source where it picked up the mutation.

From Iberia, the culture/R1b/lactose_allele could have easily spread down to Africa (explaining the R1b and lactose tolerance in Chadic speakic groups). And it could spread east, meeting up with the R1a groups around Germany.



Olympus Mons said...

@Karl_K,

There are Pottery sherds with milk fat (indication consumption and lactose tolerance) in Acacus mountains north Africa by about 5000BC. See...?

Karl_K said...

Maybe I should be more clear.

I think R1b-rich steppe people arrived in Iberia in some manner that avoided the R1a migration in North-East Europe. This could have been a purely coastal/sea migration, or an unobserved early land migration just ahead of the R1a-rich migration.

In either case, a cattle herding culture met up with a lactose tolerant culture, and it was synergistic.

We know that the Irish population turned over rapidly. And we know what the genetic elements were

No matter what, this didn't just spread from the east in a diffuse way. There was some kind of counter-pulse from the west with R1b and lqctose tolerance.

Olympus Mons said...

@Karl_K,
Yes, Iberia was pretty much "wasteland" by 4000BC. then by 3300bc had more hundreds of thousands of people (San blas, La pijotilla, Porto torrão, Perdigões, etc) .. all in SOUTH Iberia. Not north or east. South and west, just near the Gibraltar!
By 2800 bc at the most western part of Iberia (literally) Bell beaker were starting to pop up as R1b. Warrior culture was already part of Zambujal (birth place of Bell beakers) in western part of Portugal.

So it was people lots of people. And they left trails. Tell Tsaf, Merimde, Bovidean pastoral, Iberia. -- Where are the scars in Eastern Europe? Where are the thousands of boats needed to transport these people? If by land where are the mark in the landscape these hundreds of thousand would obviously have left?

John Thomas said...

.....everyone is Polish - but only they don't know it.

Olympus Mons said...

@mickeydodds1
Ahahah. No, Everyone would be polish, if it weren't for the bell beakers stopping the R1a on their tracks by 2000bc.

Anonymous said...

@Karl_K

"We know that the Irish population turned over rapidly. And we know what the genetic elements were"

And turned the population *less* Iberian-like. The BB's from Ireland show affinity with people in Germany, around the Elbe.

Alberto said...

@Kristiina

Yes, Anatolia_Neolithic are those early neolithic samples from Barcin. It's probably correct that MA1 related ancestry arrived to the Levant after the early neolithic, probably with a CHG-like population coming from further east.

However, I'd like to point out that those latest numbers for Anatolia_Neolithic and BedouinB were done using a simulation of a Basal Eurasian sample that Matt made and they were for purely testing purposes. But it doesn't seem to affect any of your comments, though, so it's just a general warning about them.

Alberto said...

@Aram

Thanks, that looks like a very interesting find. I hope we'll get DNA from it next year or so. For more detailed pictures:

http://arkeofili.com/?p=14300

Alberto said...

@Olympus Mons

Basically all Asia populations (excluding Siberians) prefer MA1 over AG3, and also Yamnaya Kalmykia over Samara. You can check which populations in the spreadshhet I linked to above.

But I think Yamnaya is too late for your theory, so it shouldn't matter much.

Alberto said...

@Davidski et al.

Still only few tests with the latest datasheet that swaps Chimp and Mbuti's position. My first impresion is that it gives cleaner results (kind of like PCA based datasheet in that sense), but I'm not sure if more accurate or not. For example, the SSA seems to go away in Spaniards, even in Extremadura where it should be present. Here with the new datasheet:

Spanish_Castilla_la_Mancha
"Anatolia_Neolithic" 51.1
"Yamnaya_Kalmykia" 38.8
"Loschbour" 10.1
"Yamnaya_Samara" 0
"Motala_HG" 0
"Hungary_HG" 0
"Karelia_HG" 0
"Kotias" 0
"Masai_Kinyawa" 0
"Atayal" 0
"Esan_Nigeria" 0

Spanish_Extremadura
"Anatolia_Neolithic" 54.45
"Yamnaya_Kalmykia" 33.7
"Loschbour" 10.55
"Atayal" 1.3
"Yamnaya_Samara" 0
"Motala_HG" 0
"Hungary_HG" 0
"Karelia_HG" 0
"Kotias" 0
"Masai_Kinyawa" 0
"Esan_Nigeria" 0

And now with the previous one:

Spanish_Castilla_la_Mancha
"Anatolia_Neolithic" 55.9
"Yamnaya_Kalmykia" 30.9
"Loschbour" 9.5
"Kotias" 1.6
"Atayal" 1.4
"Esan_Nigeria" 0.7
"Hungary_HG" 0
"Yamnaya_Samara" 0
"Karelia_HG" 0

Spanish_Extremadura
"Anatolia_Neolithic" 56.8
"Yamnaya_Kalmykia" 27.7
"Loschbour" 10.9
"Atayal" 1.9
"Esan_Nigeria" 1.65
"Kotias" 1.05
"Hungary_HG" 0
"Yamnaya_Samara" 0
"Karelia_HG" 0

The new one looks too clean? I'll keep testing to see if i find further differences.

Alberto said...

Also for Spanish_Castilla_la_Mancha (more or less average Spanish), ~39% Yamnaya (vs. ~31% with previous one) is quite significant difference (and IMO looks a bit too high).

Olympus Mons said...

@Alberto,
I will,
Yamnaya is starting to irritate me. Just read a small chapter from the book "The Horse, the Wheel, and Language: How Bronze-Age Riders from the Eurasian" and there is a short chapter about Nalchik (Nalchik and north caucasians cultures), just a few mile west of Kalmykia and its all about how it was people from Shulaveri shomu, with inhumations dated to... 4.900 bc. 147 burials were exactly the same as the shulaveri burials in contrast to the ones up in samara completely different.
Go figure the bell beakers had lots of steppe DNA similar to Yamnaya a thousand years later. The shulaveri-Shomu spill by the buck loads in the 5th millennium into Kalmykia area.
Check Nalchik inhumations with Merimde in Nile delta at exactly the same time and will be the same people.

Karl_K said...

@epoch2013

"And turned the population *less* Iberian-like. The BB's from Ireland show affinity with people in Germany, around the Elbe."

Of course. Those are the sampled populations. It does not imply the direction of movement. I agree they came from the steppe initially. But it seems like there was a secondary boost from western europe. I believe this was a lactose toletance allels boost.

Matt said...

@ Alberto, thanks. To try and show how these populations fit on the PCA - http://i.imgur.com/PgSd7vW.png.

I basically chose the PCA positions beforehand, them I used the regression transforming PCA dimensions to the stats to plot them.

If you're looking for a simulant / ghost that could get anything like the estimates in the literature (IRC around 50:50 UHG:Basal Eurasian for Anatolia Neolithic or 60:40 for LBK?), I'd guess I'd need to pick a PCA point about as far away from Anatolia_Neolithic as my UHG estimate (and therefore would combine in a 50:50). So doing so:

Anatolia_Neolithic2 BedouinB2 Cypriot2 Dai2 Denisovan Druze2 Han2 Iberia_Chalcolithic Iranian_Jew2 Karitiana2 LaBrana1 Mansi2 Motala_HG2 Munda2 Papuan2 Samara_HG Satsurblia South_Indian2 Yoruba

Basal_Sim2 0.413 0.396 0.4046 0.3091 -0.6004 0.3997 0.3088 0.3841 0.3993 0.2988 0.3063 0.3254 0.3103 0.3167 0.2729 0.3114 0.375 0.3328 0.0982

That shucks off the last of the West Eurasian affinity to be roughly equally related to HG and ENA (distance to Han: 0.3088, distance to La_Brana1 0.3063).

On a PCA:

http://i.imgur.com/k3vdQKS.png

Alberto said...

@Matt

With this second simulation, for Anatolia Neolithic, with UHG1 and UHG3:

Anatolia_Neolithic
"SimulatedUHG1" 53.45
"Basal_Sim2" 44.2
"AfontovaGora3" 2.35
"Loschbour" 0
"Kostenki14" 0
"MA1" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Masai_Kinyawa" 0
"SimulatedUHG3" 0

And without the UGHs:

Anatolia_Neolithic
"Basal_Sim2" 52.1
"Loschbour" 41.9
"AfontovaGora3" 6
"Kostenki14" 0
"MA1" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Masai_Kinyawa" 0

So the HGs position clearly affect the amount of Basal Eurasian too (not unexpected, anyway).

BedouinB and Kotias:

BedouinB
"Basal_Sim2" 58.25
"Loschbour" 20.35
"AfontovaGora3" 12.95
"Masai_Kinyawa" 5.3
"MA1" 3.15
"Kostenki14" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"SimulatedUHG1" 0
"SimulatedUHG3" 0

Kotias
"Basal_Sim2" 48.7
"AfontovaGora3" 39.2
"Loschbour" 12.1
"Kostenki14" 0
"MA1" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Masai_Kinyawa" 0
"SimulatedUHG1" 0
"SimulatedUHG3" 0

Which I would say that generally looks good, though in the case of Kotias we have this stat for Satsurblia:

Mbuti Satsurblia Villabruna MA1 -0.0024 -0.373 381469

Which doesn't agree so much with the output above (though the 3 samples are different, but that shouldn't change too much in a stat with Kotias Loschbour and AG3, I think).

With the Basal Eurasian I made based on Anatolia_Neolithic and KO1, for Kotias I get:

Kotias
"Basal_Eurasian" 32.25
"Kostenki14" 31.55
"AfontovaGora3" 26.05
"Loschbour" 10.15
"MA1" 0
"Atayal" 0
"Esan_Nigeria" 0
"Ust_Ishim" 0
"Masai_Kinyawa" 0

Which probably agrees more with the stat above, while also agreeing with the Basal Eurasian estimates of the last paper and with stats showing strong affinity to Kostenki14 (IIRC). But then it fails to explain why Ust-Ishim would be equally related to Anatolia_Neolithic and Kotias (since for Anatolia_Neolithic I get about 48% Basal). So I guess it's just difficult to place a ghost in the right place to fit with all the stats. And maybe, as someone suggested, there are more than one Basal Eurasian populations that could even be quite divergent between them.

FrankN said...

A few unrelated comments:

@OM: Spelt
The earliest European evidence of spelt stems from the early Neolithic Köros Culture in Transsylvania, and as such predates the Ubaid expansion. Reference (acc. to German Wikipedia):
Beatrice Ciută: Plant Species within the prehistoric communities from Transylvania. Editura Mega, Cluj 2012, p. 60

Dating MA1-like influx into Africa: The Horn of Africa was latest by the early 3rd mBC linked tradewise to the Persian Gulf / India. To get a better idea of when MA1-like genetic material reached East Africa, a nMonte analysis of Mota might be helpful. Anybody having the genome available and wanting to give it a try (Alberto, Matt, Dave, rk, etc)?

Spread of BB: Given that the "BB package" shows obvious Iberian influence, I still find it difficult to believe that there had been no respective gene flow into the CE plains. However, that's what all aDNA so far is showing.
A little regarded element of the BB package is the widespread use of Baltic Amber beads, which latest by the Bronze Age became a Trans-European currency (among others, e.g. Tyrolean/Slovakian ring-shaped copper ingots), and strongly resemble Christian and Muslim prayer chains. They make their first, localised appearance in the Northern (Jutlandic/ Holstein) FBC, and become more widesspread by GAC (a/o distributed into W. Ukraine; access to Sambian amber is seen as a key motive of GAC intrusion into the Narva culture).
https://repozytorium.amu.edu.pl/bitstream/10593/9792/1/Amber%20on%20the%20Threshold%20of%20a%20World%20Career.pdf

On the Iberian peninsula, whioh so far had used local or Sicilian amber (unstandardised), the first Baltic amber appears during the BB period Basque country). Latest by the Middle BA, Baltic amber (beads) had in Iberia more or less completely replaced other (including local) sources, and served a/o to finance Scandinavian copper import from NW Iberia. The distribution is quite peculiar, concentrating along the coasts (including Andalucia/ Catalonia), with hardly any inland finds.
https://www.academia.edu/1877438/Amber_Sources_and_Trade_in_the_Prehistory_of_the_Iberian_Peninsula

So, we might be dealing here with a maritime BB colonisation of the Atlantic/ British coasts out of the CE coastal plain; the "Iberian" influence would then primarily reflect imports financed by amber.

Note that the eponymous Bell Beakers have apparently evolved from earlier Beakers (CW, FB), following a tradition of the CE plain. As concerns the BB chronology, the usual caveats (reservoir effects, "old wood" charcoal dating) apply, especially when it comes to coastal sites (e.g. early Portugese dates). The Western (Rhinish) Beaker Group represents a CW-BB transition that appeared more or less contemporarily with Iberian finds (28th cBC). The Schönfelder Group on the Lower Middle Elbe, N of Magdeburg/ the Harz, successor to the Elbe-Havel-Culture, also displays a swift FB/GAC-BB transition during the same time. This group had little CW influence (only after 2500 BC). They cremated their deads, and are generally held responsible for the Eulau Massacre where a whole CW settlement was killed.

Further reading:

https://www.academia.edu/19633909/Bell_Beakers_-_Chronology_Innovation_and_Memory_a_Multivariate_Approach

https://www.academia.edu/11073974/New_radiocarbon_evidence_for_European_Bell_Beakers_and_the_consequences_for_the_diffusion_of_the_Bell_Beaker_phenomenon

Chad said...

CHG is also significantly closer to WHG and ANE than UP Europeans.

Chad said...

Being more Kostenki than WHG or ANE shouldn't be the case.

Krefter said...

@Alberto,

These are personal notes and so are sloppy, but maybe it is comprehensible.

Here are the stats.
https://docs.google.com/spreadsheets/d/1HzxBkIrlUFhbbA8Z_eiN8vjpMpe0klNxaQdE12UQlys/edit#gid=99313968

Here are the notes.

https://docs.google.com/document/d/1Tw_68H6opWqS18YROo-zek30bM0XHjG7s93aVoGIM_c/edit

Here vocab I use.
Paleo=Pre_Neolithic Europe and MA1.
Paleo Euro=Pre-Neolithic Europe.
NE=Kotais/CHG and Anatolia Neolithic/EEF.

Here are the highlights.'
>Paleos and NE are closer to Han than to Ust_Ishim.

>Extra Han affinity in MA1, EHG, and EEF.

>If the Basal/Paleo model is correct none of our Paleo genomes is a good reference for the Paleo side of our NEs.

>CHG isn't ANE/Basal. It has variable relation to Paleo Euros unlike MA1; closest to WHG and El Miron. Also, it has a much smaller ANE/Han ratio than MA1 does, while if it was ANE+Basal it would have the same ratio.

>EEF and CHG have Crown Eurasian ancestry that isn't Western or Eastern.

>CHG and EEF share a significant amount of ancestry that Paleos lack. Paleos share a significant amount of ancestry CHG and EEF lack.

__________________________________________________________________________________

To me the model that makes the most sense is; Paleo Euro+Basal Eurasian+WHG+Crown Eurasian as the foundational population of West Asia. Then in Turkey they mixed with WHG and in the Caucasus with ANE.

Krefter said...

Maybe I'm wrong about Crown Eurasian(xEast and West) in CHG and EEF. I've used (Chimp, Colmn)(Mbuti, Row), where Row are Paleo Europeans, MA1, Han, and Ust-Ishim to estimate ancestry. I added Basal and Crown Eurasian as possible ancestors.

Here are results for Anatolia Neolithic, MA1, ElMiron, La Brana-1, and Villabruna.
***: I had to do some modifying. We don't have two Goyets, so i assumed he has 0.50 closeness to himself. I can't have the same row and colmn population, so for some individuals there were differnt column populations.

"distance%=0.8978 / distance=0.008978"


Anatolia_Neolithic
"Basal" 41.7
"Vestonice16" 28.35
"Loschbour" 23.15
"MA1" 6.8
"Crown" 0

"distance%=0.6805 / distance=0.006805"


Kotias
"Basal" 43.7
"MA1" 22.15
"Loschbour" 15.35
"Vestonice16" 10.9
"Crown" 7.9

"distance%=0.6034 / distance=0.006034"

"distance%=1.8073 / distance=0.018073"


ElMiron
"Loschbour" 54.35
"GoyetQ116-1" 44.2
"Basal" 1.45
"Vestonice16" 0
"MA1" 0
"Crown" 0


LaBrana1
"Loschbour" 63.2
"ElMiron" 23.15
"Vestonice16" 4.1
"Basal" 4
"MA1" 3.65
"Crown" 1.9

"distance%=0.9786 / distance=0.009786"


Villabruna
"Loschbour" 88.9
"Basal" 9.5
"Vestonice16" 1.5
"MA1" 0.1
"Crown" 0
"ElMiron" 0

Krefter said...

Look at these results for Karelia_HG.

"distance%=2.3018 / distance=0.023018"


Karelia_HG
"MA1" 56.45
"Loschbour" 31.25
"Crown" 10.95
"Basal" 1.35
"Vestonice16" 0

Then when I add Kotais as a possible ancestor.

"distance%=1.4462 / distance=0.014462"

Karelia_HG
"MA1" 51.7
"Loschbour" 28.55
"Kotias" 18.95
"Basal" 0.8
"Vestonice16" 0
"Crown" 0

Then when I add Basal East Asian(closeness to Han is 0.4) as possible ancestor.

"distance%=1.4017 / distance=0.014017"


Karelia_HG
"MA1" 50.45
"Loschbour" 28.8
"Kotias" 17.55
"Basal East" 3.2
"Vestonice16" 0
"Basal" 0
"Crown" 0

Matt said...

@ Alberto: Not to say either is better for sure, to try and compare the Basal Eurasian ghost you simulated from upthread (Basal_Eurasian,0.46679,0.4057,0.42572,0.31455,-0.58369,0.41674,0.31372,0.41227,0.4107,0.31501,0.27507,0.33202,0.29588,0.32064,0.28032,0.32174,0.37897,0.33901,0.09475), I stuck it through PCA along with the other Basal Eurasian and UHG ghosts I simulated from PCA. I've labelled it as "Alberto_Basal_Eurasian".

http://i.imgur.com/1OliD9Y.png

Visually, you can see why the Basal_Sim2 tends to get more of a straight mix with AfontovaGora3 to generate Kotias in nMonte, as it mostly fits a cline fairly well between the two in dimensions 1-3 (then less well in the much smaller dimension 4 showing drift more specific to Kotias and Satsurblia). Your Basal Eurasian ghost tends to work really well to model Kotias as a mix with AfontovaGora3 in dimension, but doesn't really fit that in dimension 3, where it loads strongly towards the AfontovaGora3 pole of the dimension, and so needs extra admixture from the European-WHG side to fit.

That all makes sense in light of that Basal_Eurasian being less related to WHG compared to Samara, and Han, e.g. distance to La_Brana1 0.27507 and Motala_HG 0.29588, vs to Samara 0.32174, Han 0.31372, Karitiana 0.31501. It's kind of weakly an anti-WHG component with a relatively high Samara affinity (which makes sense from the method of generating it, specifically pushing Anatolia_Neolithic away from WHG). So it needs to pick up more WHG in higher dimensions to fit with West Eurasians.

Where the Basal_Sim2 sim seems a little bit more of a neutral thing to all the "Crown Eurasian" groups (LaBrana1 0.3063, Han 0.3088, Samara_HG 0.3114, Karitiana 0.2988, Motala_HG 0.3103) although not perfectly so.

Krefter said...

We can test the theory of East Asian admixture in Paleos instead of Basal Eurasian admixture in the Middle East with D-stats. What we'll need are a bunch of Middle Easterners and East Asians as the rows.

Here are attempts with Anatolia Neolithic, Kotais, Ust_Ishim, and Han as the row populations.

"distance%=2.5721 / distance=0.025721"


Vestonice16
"Anatolia_Neolithic" 44.1
"Basal East" 40.35
"Kotias" 15.55

"distance%=2.837 / distance=0.02837"


GoyetQ116-1
"Basal East" 49.45
"Anatolia_Neolithic" 33.8
"Kotias" 16.75

One hole in this theory is Paleos are closer to CHG/EEF than CHG/EEF are to Paleos. Besides that though I don't see how this is less plausible than the Basal Eurasian theory.

Olympus Mons said...

@FrankN,
Thanks. Will look into It.
However careful with details. ;) - See Koros is 5800bc-4500bc. So, Shulaveri were on the run for over 500 years when Koros ended. And actually the first spelt is from well the beginning of the 7th millennium in ARUKHLO , the Biggest settlement of the Shulaveri. So much earlier than koros.

Then there is an issue with spelt because there some confusion of Club wheat and spelt (because of the Hexaploid) but not really the spelt. But I dont know. Will look into it. Thanks.

Krefter said...

Edit:

Another whole in the East Asian admixture in Paleos theory, is mtDNA/Y DNA. About 100% of Paleoshad mtDNA U and Y DNA C1 or I or R. If they were part East Asian we would probably see mtDNA and Y DNA lineages closely related to East Asian ones. I guess you could argue Y DNA R and C1 are, but the relationship is very distant is basically goes back to the Eurasian root.

Middle Easterners on the other hand have a huge diversity in mtDNA and Y DNA lineages going back to the Eurasian roots. Plus they have lots of mtDNA U and Y DNA IJ, making them look like a mixture of Paleo and other types of Eurasian. You don't see mtDNA R0 or JT or Y DNA G or E1b in Paleos, like you see mtDNA U and Y DNA IJ in the Middle East.

Matt said...

Alberto, also re the stat

Mbuti Satsurblia Villabruna MA1 -0.0024 -0.373 381469

while I don't doubt that, for these stats, we have

Villabruna LaBrana1 Motala_HG2 Samara_HG Han Karitiana
Kotias 0.3632 0.3643 0.3724 0.3828 0.3313 0.3433

which is consistent with Kotias being from a neutral basal origin then picking up AG3 or Karelia related admixture. I can't really explain why there is a difference in these stats from that "direct" D-stat though! Unless Satsurblia is strongly variant from Kotias here (and in which case how?).

Olympus Mons said...

@Alberto,
Man... the further you talk about Yamnaya_Kalmykia/Bell Beaker the wider my eyes open.

We know that Kalmykia, via Nalchik findings, is actually extremely connected to shulaveri-Shomu. So, would I be correct, in inference off course, that your results are showing that genome from inhumations from Yamnaya Kalmykia (ie Shulaveri traits) show relatedness to Bell beakers by a large amount, in contrast with genome found in Yamnaya farther north in Samara that are more related to CW and do not have that relatedness to BB?
or put in another way - If Kalmykia is shulaveri so shulaveri show relateness to Bell beaker, Right?

Davidski said...

Olympus,

These Central European Bell Beakers don't have any Near Eastern admixture that wasn't mediated via Middle Neolithic Central European farmers or Bronze Age steppe populations, so their ancestors did not travel through the Near East or North Africa. You'll have to accept this at some point and include it in your model for Bell Beaker origins.

Kristiina & Frank,

There's no perceptible MA1-related (aka ANE) ancestry in Africa. I've looked at this issue closely.

Seinundzeit said...

David,

I'm rather late to all of this, but I just wanted to thank you for creating these nMonte sheets, great stuff!

Matt, Alberto, Krefter, and others,

I think Matt's simulated populations allow for models which make good sense, and which (to my mind) seem to be tracking real affinities/mixtures.

Some models for different West Eurasian groups (same reference populations for all of them, any reference population not shown was at 0%):

Ancient Anatolia/Ancient Europe/Ancient West Asia

Anatolia_Neolithic
53.45% UHG
44.20% Basal Eurasian
2.35% ANE
(consistent with CHG-related admixture in these Neolithic Anatolians)

Iberia_MN
61.60% UHG
27.75% Basal Eurasian
7.90% Villabruna
1.65% GoyetQ116-1
1.10% ENA
(Nice to see GoyetQ116-1-related admixture here, the El Miron connection comes to mind)

Loschbour
84.45% Villabruna
15.55% UHG

Kotias
47.65% Basal Eurasian
39.35% ANE
6.85% Villabruna
6.15% UHG
(the high level of Basal Eurasian admixture ties into other work, like how CHG has appeared to be quite basal in various analyses David tried. High ANE is as expected, and some UHG/WHG also appears, just as we expected)

Europe

Basque_French
57.15% UHG
26.55% Basal Eurasian
14.10% ANE
2.15% ENA

Lithuanian
54.85% UHG
26.75% ANE
16.85% Basal Eurasian
1.55% ENA

Finnish
50.30% UHG
27.55% ANE
15.10% Basal Eurasian
7.05% ENA

Chuvash
34.00% UHG
32.7% ANE
19.65% ENA
13.65% Basal Eurasian

Saami
48.8% UHG
22.2% ANE
21.8% ENA
7.2% Basal Eurasian

Steppe

Karelia_Hunter
59.35% ANE
23.95% UHG
16.70% Villabruna

Yamnaya_Samara
47.30% ANE
36.05% UHG
16.65% Basal Eurasian

Andronovo
41.25% ANE
35.50% UHG
20.35% Basal Eurasian
2.30% Villabruna
0.55% ENA
0.05% Kostenki14

South Central Asia

Kalash
40.15% ANE
28.60% Basal Eurasian
13.10% ENA
10.90% UHG
7.25% Kostenki14

Pashtun
39.55% ANE
27.80% Basal Eurasian
13.90% ENA
9.90% UHG
8.85% Kostenki14

Tajik_Ishkashim
38.05% ANE
27.15% Basal Eurasian
20.65% UHG
14.15% ENA

Tajik_Rushan
37.7% ANE
27.70% Basal Eurasian
23.70% UHG
10.40% ANE
0.50% Kostenki14

Brahui
34.9% ANE
33.00% Basal Eurasian
22.85% Kostenki14
7.30% ENA
1.95% UHG
(In every nMonte analysis, the Brahui/Baloch turn out quite strange. Same with ADMIXTURE results. I'm pretty sure that the Brahui/Baloch are living remnants of something very ancient)

West Asia/Caucasus

Iranian
39.50% Basal Eurasian
30.4% ANE
16.05% UHG
9.65% Kostenki14
4.15% ENA
0.25% Villabruna

Georgian
43.50% Basal Eurasian
29.30% UHG
24.65% ANE
2.55% ENA

Lezgin
35.60% Basal Eurasian
29.95% UHG
29.85% ANE
4.60% ENA

Note: in the models I tried, most populations scored substantial AG3 and MA1 together, no real differentiation, with some exceptions.

Krefter said...

@Davidski,

I have more D-stats to include some modern West Eurasians into the model. It's a lot but they're all needed. Using all 6000 BC+ old genomes as row populations would be a good idea for nMonte.

https://docs.google.com/spreadsheets/d/12spvBc3PkdaTQOxoxK9txjRJj8cf3FB_TRGa6Kpfv7I/edit#gid=0

Davidski said...

There's at least one problem with that list. Kostenki is spelled Kostinki.

Krefter said...

@Matt,

We can't assume the UHG+Basal Eurasian model by Laz 2013 is correct. We need to test every possibility. Your Basal Eurasian ghosts adopts to the problem of the UHG+Basal Eurasian model not agreeing with D-stats. It'll give good nMonte results because it was forced to, not because the Basal+UHG model works.

These results for your Basal ghost don't make sense.
Basal-Motala HG=0.351
Basal-Samara HG=0.3442
Basal-LaBrana1=0.3483

30,000-40,000 years ago WHG had many relatives. It proves there were many people who were related to WHG but not WHG. WHG was a close knit group and the West Eurasians who lived in the Middle East didn't have to apart of that close knit group. When a realistic Basal Eurasian is used, all near Easterners use more Vestonice(distant relative of WHG) than WHG.

Krefter said...

"There's at least one problem with that list. Kostenki is spelled Kostinki."

Fixed it.

Seinundzeit said...

For the fun of it, Albanians and the Iron Age Scythian:

Albanian
45.0% UHG
35.1% Basal Eurasian
16.4% ANE
3.5% ENA

Scythian
37.10% ANE
24.55% UHG
19.95% Basal Eurasian
9.35% ENA
9.05% Villabruna

Matt said...

@ Krefter, I wouldn't say forced to fit really. I picked the points on the PCA made of the D-stats, and then made a few test populations that have the statistical properties those points should have, and then, so long as the PC dimensions generated capture a lot of the variance, obviously they're then going to combine in ways that can approximate the stats for the populations. I'm pretty open about there being completely theoretical at the moment and that we don't have evidence for them.

I don't know if that's less realistic than using a Basal Eurasian model that is just arbitrarily declared to have 0.3 in each stat from every Eurasian and 0.27 from Papuans. Also I was wondering what you did with the Anatolia_Neolithic, BedouinB2, Satsurblia columns - did they get 0.3 as well? Which would be Basal Eurasian being equally related to them and to other Eurasians. Those columns and including them in the model are crucial to getting any kind of Basal Eurasian that gives even semi-plausible fits.

@ Sein, thanks for testing with those, I guess you used the Basal Eurasian Sim 2?

Also separately from Basal Eurasian and the results being tested there I wanted to say to you and Alberto it might be worth testing Yamnaya now with Anatolia_Neolithic, Kotias, Villabruna and AfontovaGora3 (and whatever other plausible ancestors you could add) as AfontovaGora3 has some quite different stat properties from MA1 and is later in history. We could also try with Karelia_HG+AfontovaGora3 as well.

I'm interested if the model of Kotias+KareliaHG to the exclusion of Anatolia_Neolithic, etc. still stands up exactly the same as it did with adding AfontovaGora3 ancestry as an option. Even though for reasons of geography Kotias+KareliaHG is more compelling, it would be good to test if nMonte comes down decidely for Kotias+KareliaHG still.

Ryukendo K said...
This comment has been removed by the author.
Krefter said...

@Davidski,

Do you think making literally almost every human population a row population would make the best D-stat spreadsheet for nMonte? Or at least we can do most West Eurasians.

@Ryuk,
"Ust_Ishim."

True. I expect Basal Eurasian to be real, but who knows.

@Matt,

We don't know if Basal Eurasian had the same relationship to all modern East Asians and Stone age North Eurasians. We currently don't have the data needed inorder to get an answer. So, i think it's best to model Basal Eurasian as being equally close to modern East Asians and Stone age North Eurasians.

That's why I gave Basal Eurasian a 0.3 stat with all of them. I deleted any row populations with Basal Eurasian ancestry from the row when I estimated Basal Eurasian. This is because we don't have the ancient genomes needed to know how close Basal Eurasian would score to modern populations with Basal Eurasian ancestry. To model Anatolia_Neolithic as Basal Eurasian+UHG with Anatolia_Neolithic as a row population, Basal Eurasian would have to be closer to Anatolia_Neolithic than Anatolia_Neolithic is to itself.

Alberto said...

@Matt

Thanks for those plots. Yes, my Basal_Eurasian is a simple straight line from Hungary_HG passing through Anatolia_Neolithic, so that AN will always be almost in the middle between the Basal_Eurasian and Hungary_HG. The most simplistic approach. Good to use different ones to test how the results agree/disagree with the stats we have.

Re: those stats about Kotias-WHG-ANE relationship, I think that Samara_HG quite likely has Kotias admixture. I don't know if you saw the qpAdm model, but there it showed close to 25% Kotias. Which looks too high to me, unless Karelia_HG also has some amount of it (which didn't work in qpAdm, but Karelia_HG is closer to Ust-Ishim that other Crown Eurasians). I guess if someone could try:

Mbuti Ust_Ishim MA1 Loschbour
Mbuti Ust_Ishim MA1 AfontovaGora3
Mbuti Ust_Ishim MA1 Karelia_HG
Mbuti Ust_Ishim MA1 Samara_HG
Mbuti Oase1 MA1 Loschbour
Mbuti Oase1 MA1 AfontovaGora3
Mbuti Oase1 MA1 Karelia_HG
Mbuti Oase1 MA1 Samara_HG
Mbuti Kotias Samara_HG Karelia_HG

Alberto said...

@Matt

I run some nMonte models with AG3 and MA1, without Karelia_HG. For Yamnaya the Anatolia_Neolithic doesn't change. The only interesting thing is that Yamnaya_Kalmykia takes MA1 while Samara doesn't:

https://docs.google.com/spreadsheets/d/1KNPbAczNTVu_shiMcpx-ZCexceHoBAPndLSuy6YsmVw/edit?usp=sharing

Here with Karelia_HG:

Yamnaya_Samara
"Karelia_HG" 53.45
"Kotias" 28.4
"Anatolia_Neolithic" 17.15
"Atayal" 0.65
"AfontovaGora3" 0.35
"Esan_Nigeria" 0
"Hungary_HG" 0
"Loschbour" 0
"MA1" 0

Yamnaya_Kalmykia
"Karelia_HG" 34.5
"Kotias" 31.2
"Anatolia_Neolithic" 18.05
"MA1" 14.25
"Motala_HG" 1.65
"Esan_Nigeria" 0.35
"Atayal" 0
"Hungary_HG" 0
"Loschbour" 0
"AfontovaGora3" 0

Alberto said...

Ah, and what should the UHG represent? I thought it should be a Near Eastern HG that mixed with Basal Eurasian, before then mixing further with WHG to produce Anatolia_Neolithic/EEF? But from the models it looks more like an Iberian_HG?

And what would be a good theoretical model for Kotias? One where it would be something like 40% Basal_Eurasian, 30% Villabruna and 30% AG3? Though that looks difficult. I think that Kotias will always be closer (on a PCA) to AG3 than to Villabruna, so the stat making them equally related would be mediated by a West Eurasian that is closer to Villabruna than to AG3, more or less like Kostenki14 (though not Kostenki14 itself, since stats show he shares less drift with Kotias than Villabruna or AG3 do? So we'd need an UHG that is somehow similar to Kostenki14 to represent this Near Eastern lineage?

Ryukendo K said...
This comment has been removed by the author.
Olympus Mons said...

@Davidski,
Actually no.
The only interaction that stock of people had while traveling trough North Africa was with proto-berber populations. For which I don’t think we have any DNA to match against. As far as we know probably all they (stock becoming bell beakers) would have picked up there, while picking up Mtdna H1, was an increase of BASAL EURASIAN compared to YAMNAYA_KALMYKIA. Is that not the case? Can it be tested? – This would be a good indication.
Second I would expect to have a whiff , a Whiff of SSA into BB, because they could have picked up some women while at Fayum (Merimde phase) that were Mtdna L3f and those should have had a bit of SSA and I suppose some BB later Iberia and beyond could show minor traces of SSA. That’s why I keep on asking about the whiff of SSA that seem to pop up with some breakdowns of BB.
So, until we test Proto-berber (which I think is shit loads of Basal Euraasian) and see how close those were to Middle Neolithic Farmers... I figure pretty close. I am sure if one testes Merimde inhumations those will be pretty close to say Kalmykia (truth is test Nalchik a bit south!!).
Well see.

Olympus Mons said...

@ALberto,
would it be possible to see percentages of Basal Eurasian ( or Anatolia Neolithic if better) in Yamnaya_Kalmykia and percentages of Basal Eurasian in bell beakers, still with MA1 in the mix?

Alberto said...

@OM

We can't really measure Basal Eurasian accurately for now, we're just doing experiments with ghosts. So more realistic is using Anatolia_Neolithic and Kotias. For that, I posted it earlier here:

https://docs.google.com/spreadsheets/d/1KNPbAczNTVu_shiMcpx-ZCexceHoBAPndLSuy6YsmVw/edit?usp=sharing

But keep in mind that Yamnaya_Kalmykia is a population that didn't exist in the steppe 6000 BC. At that time we don't know who could be in the steppe, maybe something like Karelia_HG, or maybe no one at all? A better option for the origin of those Shulaveri guys that you're tracking might be east of the Caspian Sea, like Djeitun in South Turkmenistan.

Kristiina said...

Davidski, I am not able to make any calculations myself, but I still think that there should be some traces of ANE admixture in countries like Egypt and areas like the Horn of Africa due to this cultural continuum from North Africa to southern Central Asia that has very early roots. In any case, Krefter finds 6.8% MA1 in Anatolian Neolithic and Alberto 14.1% MA1 in Bedouins and 4.8% AfontovaGora3 in Anatolia Neolithic. There are so many different ways of calculating the admixtures and getting different results.

Olympus Mons said...

@alberto,
Thanks.
No, no.
-6000BC where Shulaveri.
-4.500 BC crossed the Mountains fleeing and where, say Nalchik (147 burials) just south of Kalmykia, after what I call Ubaid nuclear bomb event (only it took almost 1000 year exploding :- ) )
- by 4000BC probably already in the Kalmykia region
-By 3500 and later in all those regions and lots of them in the south Ukraine. Actually where you see a lot of Corded ware dead guys struck by what I would call Yamnaya arrowheads that resemble Shulaveri time one and at regions where inhumation lay on contracted for instance have Shulaveri Obsidian knifes…. But that is a long story.

So thanks. Kalmykia is the best fit at this point and that is what I bet my chips in. To me Kalmykia will have already a bit more than Shulaveri of whatever existed in the region already by 4500 BC, but if one samples, pretty much alike guys in Merimde at Egypt Nile.

Alberto said...

@Kristiina

I think that BedouinB is from somewhere around the border between Israel and Egypt? I don't know where BedouinA is from.

Anyway that calculation that you mention was with a simulated Basal Eurasian sample, so here with only real samples:

BedouinB
"Anatolia_Neolithic" 73.8
"Kotias" 15.55
"Esan_Nigeria" 7.45
"Atayal" 3.2
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"MA1" 0
"Motala_HG" 0

BedouinA
"Anatolia_Neolithic" 63.65
"Kotias" 20.1
"Esan_Nigeria" 10.55
"Atayal" 5.25
"MA1" 0.45
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"Motala_HG" 0

So it seems they do have some MA1 related ancestry via Kotias. If I add Yamnaya_Kalmykia, they even prefer that over Kotias:

BedouinB
"Anatolia_Neolithic" 82.25
"Esan_Nigeria" 7.85
"Yamnaya_Kalmykia" 5.65
"Atayal" 4.25
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"MA1" 0
"Motala_HG" 0

BedouinA
"Anatolia_Neolithic" 67.65
"Yamnaya_Kalmykia" 16.2
"Esan_Nigeria" 11.15
"Atayal" 5
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"MA1" 0
"Motala_HG" 0

Though what they really prefer is Armenia Bronze Age:

BedouinB
"Anatolia_Neolithic" 52.35
"Armenia_BA" 40.6
"Esan_Nigeria" 7.05
"Atayal" 0
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"MA1" 0
"Motala_HG" 0
"Yamnaya_Kalmykia" 0

BedouinA
"Armenia_BA" 49.6
"Anatolia_Neolithic" 38.65
"Esan_Nigeria" 10.1
"Atayal" 1.65
"Hungary_HG" 0
"Loschbour" 0
"Karelia_HG" 0
"AfontovaGora3" 0
"MA1" 0
"Motala_HG" 0
"Yamnaya_Kalmykia" 0

Krefter said...

@Alberto,

Middle Eastern diversity can't be explained by ancient DNA we have so far. We shouldn't be modelling them as a mixture of our current ancient genomes. With the row populations Davidksi's spreadsheets have, Middle Easterners fit best being mostly Cypriot aka Anatolia_Neolithic without high EEF and WHG affinity. Anatolia_Neolithic can take the place of Cypriot but the fits won't be as good.

A weird thing about Middle Easterners is EEF is closer to them as opposed to Mbuti than they are to each other. It's very weird. I guess that shows a weakness of D-stats.

Ryukendo K said...
This comment has been removed by the author.
Kristiina said...

Alberto, yes I agree, but I see that in Seinundzeit's recent calculations Kotias is 39.35% ANE and if in your calculations Bedouins are between 15-20% Kotias, it should mean that Bedouins carry c. 10% ANE. I also think that it is understandable that Bedouins prefer Armenia Bronze Age which is closer in time to them over more ancient genome. How do the Bedouins relate to the Scythian genome which is more recent? Of course, Scythian has probably much more all kinds of ENA to be a good reference point.

Olympus Mons said...

@Alberto,

Just to be imprinted for future reference (who knows, right)!


SHULAVERI SHOMU
Anatolia_Neolit 25
Kotias 30
AfontovaGora3 15
MA1 15
Hungary_HG 15
Loschbour 0
Ust-Ishim 0
Atayal 0
Esan_Nigeria 0

BELL BEAKER IBERIA
Anatolia_Neolith 50 (by H1 mtdna)
Kotias 15
AfontovaGora3 12
MA1 12
Hungary_HG 10
Loschbour 0
Ust-Ishim 0
Atayal 0
Esan_Nigeria 1

Kristiina said...

Ruykendo, Nubian complex is probably the earliest culture uniting Near East and Northeast Africa. Kebaran culture (21-16 kya) is restricted to the Levant but it is in Levant where Africa meats Asia. Mushabian culture seems to have covered Nile Valley and Levant. Natufian culture covers approximately the Kebaran area. During the period from c. 20 kya to 10 kya, archaeologists seem to argue for migrations from Africa to Levant and from Levant to Africa. The well known case is African yDNA R1b and Neareastern yDNA E and a less known case is mtDNA M1 and U6. The big empires came later. Egyptian Culture and Mesopotamian Cultures (Sumer, Akkad, Assyria) meet each other in Palestine and Sinai, and there surely were trade contacts and movements of people from an early date. At its highest glory, Egyptian Empire extended to Syria and Iraq. Persian Empire at its greatest extant (c. 500 BC) was huge: it covered Egypt, Levant, Turkey, Caucasus, Iraq, Pakistan, Afghanistan and Iran. Then, as you mentioned yourself, there was this Islamic Caliphate which extended from Spain to Pakistan. After the Arabs, Turks took over the caliphate and ruled over many areas from Egypt/Levant (Ottoman Empire) to India (Mughal Empire).

Kristiina said...

By the way, it seems that Tutankhamun's yDNA may be R1b. For a reason or other, the results have never been published.

Davidski said...

You guys make this blog very interesting, but half of the time many of you are out and about with the fairies in la la land.

I'll show you later today what Yamnaya really is, and it has nothing to do with Anatolia.

Grey said...

Karl_K

"No matter what, this didn't just spread from the east in a diffuse way. There was some kind of counter-pulse from the west with R1b and lqctose tolerance."

Yes, if there's difficulty fitting all the elements of BB to a single template then maybe it's because there are two?

One model that might explain it is a single source population taking two separate routes to the west hence cultural similarities between the two but genetic dissimilarity due to mixing with different populations en route so maybe 1) river BB: steppe + german neolithic and 2) maritime BB: steppe + Iberian neolithic.

epoch

"And turned the population *less* Iberian-like. The BB's from Ireland show affinity with people in Germany, around the Elbe."

Hare and tortoise?

If there were two BBs then maybe a faster maritime Iberian hare got to Ireland first only to be displaced later by a slower German tortoise.

Grey said...

"it seems that Tutankhamun's yDNA may be R1b"

Yes, it's annoying they are hiding it as it might throw light on the V88 story.

It wouldn't necessarily mean much overall as I assume lots of the many dynasties were the result of temporary takeovers by mercenaries etc.

Kristiina said...

Should Alberto be correct with his model(Ma1 ancestry in (south) Levant only after Neolithic), then ANE gene flow to Levant and Northeast Africa may have started with Mesopotamian Cultures and intensified during the Persian (c. 500 BC), Hellenic (c. 300 BC) and Turkic empires (c. 1200 AD). Hittite empire (c. 1600 BC–1200 BC) was restricted to Anatolia and Syria and did not extend to Africa or Iran. The Roman Empire extended to Africa, Levant, Caucasus and Mesopotamia but not to Iran. Trade and cultural contacts surely have existed from the early Egyptian and Mesopotamian eras to the present day.

Nirjhar007 said...

You guys make this blog very interesting, but half of the time many of you are out and about with the fairies in la la land.

Instead of focusing on common and regular commentators, you should focus on the scientists and contact them to correct their silly and childish mistakes!.

Olympus Mons said...

@Grey,
Apparently the TUT was R1b M269 and not V88! amazing. And its M269 without any farther mutation so very old.
Here is a part of my thesis, where I just address this issue as a passing trough

"So, there should not be a surprise if Tutankhamun was really an R1b.
I don’t really want to go there… but. It’s not even difficult to insulate this. The top North Africa areas with R1b are Tuaregs from Niger (33%) and Siwa berbers from Egypt (29%). The first ones, are the dead-end of the R1b-V88 going south. The second ones are the ones that stood behind and stick around the Siwa oasis (big archeological site and home of clearly outcast people of ancient Egypt for millennia). Siwa oasis means the protectors of the sun god AMUM… go figure why the boy king R1b was called Tutankh-AMUM."



Davidski said...

Krefter & Alberto,

https://drive.google.com/file/d/0B9o3EYTdM8lQQ1NKdUJHU2hPNlk/view?usp=sharing

Karl_K said...

"By the way, it seems that Tutankhamun's yDNA may be R1b. For a reason or other, the results have never been published."


That is usually a sign that the result was just contamination and couldn't be confirmed.

Olympus Mons said...

@Karl_K
Yes, usually...
However, having in mind that results showed tjat the Kid was R1b M269 it would really trigger a shit storm of biblical proportions for having an "European" as the pharaoh. Even if it would be erroneous to state such a think because there were be no "europeans" at that point in Europe.
:-) well who knows. They never said up until today that it was contamination, didnt they? Yet what showed up in that National Geographic (I think) was really an R1b M269.


Matt said...

@ Alberto: Thanks for testing if AG3 changes anything, looks like it doesn't.

Ah, and what should the UHG represent? I thought it should be a Near Eastern HG that mixed with Basal Eurasian, before then mixing further with WHG to produce Anatolia_Neolithic/EEF? But from the models it looks more like an Iberian_HG?

If we're talking about the stats for the ghost UHG I ran up, I think one way to look at this is to look at the difference from their closest real samples, Loschbour, Hungary_HG, Villabruna and Bichon:
http://i.imgur.com/A5Jx8a0.png

Comparing UHG-1 as it was the simplest construct, we might have expected it to share a very low affinity to Samara, *but* actually statistically what it is, is pretty similar to WHG in affinities to all the outgroups, and to Samara_HG, but drops lots of the affinity to La Brana (especially in comparison to Loschbour and Bichon) in favour of increased affinity to the Near East.

It gets pulled away from Karelia_HG because it has high affinity to Anatolia_Neolithic, which Karelia_HG has low affinity to, not because the UHG sim itself has low affinity to Samara_HG.

So, as a ghost, it seems like it does function like we would expect a population would function if it was a effectively clade with the Villabruna cluster, but contributed more to the ME and Anatolia_Neolithic (while another member of the clade went off to the Eastern European steppe to mix with AfontovaGora3 relatives there and that gives us EHG and SHG).

(The UHG-2 and UHG-3 are more "anti-Samara" because I targetted them at positions close to the European Upper Paleolithic HG in Dimension 3 of my PCA, which distinguished EUP from Kotias and AG3).

As for the question about Kotias, not sure.

Matt said...

@ Krefter: To model Anatolia_Neolithic as Basal Eurasian+UHG with Anatolia_Neolithic as a row population, Basal Eurasian would have to be closer to Anatolia_Neolithic than Anatolia_Neolithic is to itself.

That's a really reasonable objection; you can't get a perfect fit through nMonte unless you have this kind of strange outcome of a population being closer to Anatolia_Neolithic2 (column) than Anatolia_Neolithic1 (row) is.

But on the other hand, without the specific Near Eastern columns, you are going to get outcomes like Vestonice contributing a high share, because you have a lack of any specific affinity to the Near East being measured, and the UP Europeans are similar to Basal Eurasian admixed populations in relatively low affinity to ENA.

(Though please note that none of the Basal or UHG sims I created get more affinity to Anatolia_Neolithic2 than Anatolia_Neolithic 1 has, because the PC plus regression method is not going to be exaggerating everything about Anatolia_Neolithic relative to WHG, and all the columns have a unique vector in the PC space being regressed on. They do still get appropriate fits in nMonte, because the other columns are statistically close.)

Krefter said...

@Matt,
"But on the other hand, without the specific Near Eastern columns, you are going to get outcomes like Vestonice contributing a high share, because you have a lack of any specific affinity to the Near East being measured, and the UP Europeans are similar to Basal Eurasian admixed populations in relatively low affinity to ENA."

EF's Near Eastern affinity probably mostly comes from after UHG+Basal mixed with each other, so there's no point in making its Near Eastern affinity a mixture of UHG and Basal ones. EEF has too low WHG affinity to be Basal+WHG or Basal+UHG.

This is why UP Europeans are used more than WHG in nMonte. It's not arguing EEF is part Vestonice or Kosteinki, it's arguing a Near Eastern population with lots of affinity to modern near Easterners but the same affinity to my row populations, contributed a lot to EEF.

Grey said...

Olympus Mons

"The top North Africa areas with R1b are Tuaregs from Niger (33%) and Siwa berbers from Egypt (29%)."

Just googled Siwa oasis a bit. Used to have a tradition of silversmiths due to it being on a silver trade route to Egypt from further south.