An attempt at deep West Eurasian phylogeny

[PopGenist82]

What was the Z-value for your graph  ? 


I'm not entirely convinced about a separate, pre-ZK "Basal" clade.

Also interesting that your hypothetical maps imply that Europe was a kind of 'secondary homeland' for dispersal of Eurasian clades. But some might argue for different possible routes. Hard to say for sure

[Chad]

My advice would be to start with only the pre 40kya and work your way up, adding one at a time according to its date.

Working up step-by-step in chronological manner isn't exactly a methdological novelty.

Any more clarifications on your odd claims Caucasus MUP have no relation to later west asians ?

I'm debating whether to discuss it here or float it into part of my book, setting up farmer origins.

[PopGenist82]

My advice would be to start with only the pre 40kya and work your way up, adding one at a time according to its date.

Working up step-by-step in chronological manner isn't exactly a methdological novelty.

Any more clarifications on your odd claims Caucasus MUP have no relation to later west asians ?

I'm debating whether to discuss it here or float it into part of my book, setting up farmer origins.

Well, perhaps keep the 'secret punch' for your Book, but for now you might attempt to clarify whether you mean
 
1. Kotias _Up and Dzudzuana specficially were reproductively unsuccesful and became locally extinct, which is perfecty reasonable, but groups like them from the broader region contributed to later West Asians, WHG, EHG, etc

2. The entire proposal that groups like them ^ are not in fact relevant for later West Asians, despite suggestions from literature & from independent qpGraphs such as Kales (and what Ive seen from mine)

[Chad]

If I share this with you, it is going to be through email and not on here. How about that? You will then see what I'm talking about. I have Dstats, qpAdm and qpGraph showing the exact same thing.

[Kale]

Do you still avoid using Afontova Gora- 3 ? Its coverage is not too horrid

AG3 is 265k snps, the next lowest coverage groups are MA1, Pinarbasi, and Georgia_UP all at around 800k.
I just don't think AG3 really has anything to add. In graphs I've run before it just occupies ~the ANE3 node.

What was the Z-value for your graph  ? 
I'm not entirely convinced about a separate, pre-ZK "Basal" clade.
Also interesting that your hypothetical maps imply that Europe was a kind of 'secondary homeland' for dispersal of Eurasian clades. But some might argue for different possible routes. Hard to say for sure

I'm assuming by z-value you mean the 'worst f4 residual' which is how graphs in Admixtools1 were generally evaluated.
Admixtools2 by default utilizes F3 stats and evaluates the 'best' graph by a more general 'score' (more like a chisq I suspect).
I've included the F3 residuals in the zip file 'residuals' or 'residuals.xls', the worst of which was Mbuti Pinarbasi China_UP = Z -2.01, meaning relative to the graph, in actuality Pinbarbasi is less related to China_UP by a factor of Z = 2.01

1. Kotias _Up and Dzudzuana specficially were reproductively unsuccesful and became locally extinct, which is perfecty reasonable, but groups like them from the broader region contributed to later West Asians, WHG, EHG, etc

2. The entire proposal that groups like them ^ are not in fact relevant for later West Asians, despite suggestions from literature & from independent qpGraphs such as Kales (and what Ive seen from mine)

Kotias_UP/Dzudzuana really look like near-direct ancestors of WHG. From the graph, Italy_WHG are modeled as ~32.4% Gravettian, 22.5% Goyet-related, 38.2% Georgia_UP, 7% ANE.
So looking at analogous F3 stats...
F3: Mbuti Gravettian_KremsVestonice Kotias_UP = 0.2657
F3: Mbuti GoyetQ116_1 Kotias_UP = 0.2597
F3: Mbuti AG3 Kotias_UP = 0.2528
F3: Mbuti AreneCandide = 0.2803

So if WHG has 38.2% Kotias-UP like ancestry, what would the F3 value of that to Kotias-UP actually be to produce the increase in affinity? The answer 0.3107
F3: Mbuti Dzudzuana Kotias_UP = 0.3222
That's pretty close, suggesting Kotias_UP and Dzudzuana only have a limited history together after their near-relative mixed into WHG.

I don't think they are directly relevant for anyone else though. Pinarbasi has a ton of similar ancestry, but it branched off a ways before, and it's Pinarbasi related stuff that is more relevant for other Near-Eastern (and North African) populations.

[TanTin]

This is my minimum distance tree for some of these paleolitic populations.
How do I build my tree?  I use PCA/ MDS data for all the individuals.  Once I have the MDS matrix, I run cluster analysis only for a small set of selected individuals to build the tree (min distance tree).
As you may notice my tree is basicaly the same as the most reliable qpWave structure. 
How deep Wes

[Quint]

Oase-1 may be contaminated, but why was Salkhit left out? Both may be relevant to the origins of ANE.

[TanTin]

Oase-1 may be contaminated, but why was Salkhit left out? Both may be relevant to the origins of ANE.

Thanks for asking this question.

I have Salkhit from another dataset, it is in V52 .  And this projection was generated by using more principal components ( I used 80).
Salkhit is on the Easterns side,  however it is connecting directly to ZK. I was expecting it will connect to Tianyuan.   The algorithm found it closer to ZK. 

Similar is the picture on the qpGraph:

File (abc1166_massilani_sm.pdf)

https://www.science.org/action/downloadS...ani_sm.pdf

[Kale]

Oase-1 may be contaminated, but why was Salkhit left out? Both may be relevant to the origins of ANE.

Oase1 (1240k version) looks fine contamination wise, but the coverage isn't great so it's noisy. The .SG version definitely is contaminated.
Salkhit (from reichlab v50) looks heavily contaminated.

[Norfern-Ostrobothnian]

What do you make of these stats
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_UP_Gravettian 0.00439 0.000649 6.77 1.31e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_Krems1_1 0.00596 0.000636 9.37 7.60e-21
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_KremsWA3 0.00534 0.000843 6.34 2.30e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice5 0.00436 0.000652 6.69 2.26e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice6 0.00427 0.000668 6.40 1.60e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice16 0.00514 0.000662 7.77 8.07e-15
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_South_HG_Ostuni1 0.00708 0.000736 9.62 6.75e-22

Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_UP_Gravettian 0.000723 0.000800 0.903 0.367
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_Krems1_1 0.000226 0.000763 0.296 0.767
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_KremsWA3 -0.0000640 0.000954 -0.0671 0.947
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice5 -0.000134 0.000774 -0.173 0.863
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice6 0.000387 0.000798 0.485 0.628
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice16 0.000154 0.000783 0.197 0.844
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_South_HG_Ostuni1 0.000736 0.000855 0.861 0.389

Chimp.REF Italy_UP_Gravettian Czech_Vestonice16 Italy_South_HG_Ostuni1 0.0115 0.00104 11.0 2.57e-28
Chimp.REF Austria_Krems1_1 Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00998 0.00108 -9.27 1.82e-20
Chimp.REF Austria_KremsWA3 Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00854 0.00129 -6.63 3.35e-11
Chimp.REF Czech_Vestonice5 Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0142 0.00125 -11.3 1.10e-29
Chimp.REF Czech_Vestonice6 Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0155 0.00114 -13.7 1.99e-42

[Kale]

What do you make of these stats
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_UP_Gravettian    0.00439 0.000649  6.77 1.31e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_Krems1_1      0.00596 0.000636  9.37 7.60e-21
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_KremsWA3      0.00534 0.000843  6.34 2.30e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice5      0.00436 0.000652  6.69 2.26e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice6      0.00427 0.000668  6.40 1.60e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice16      0.00514 0.000662  7.77 8.07e-15
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_South_HG_Ostuni1 0.00708 0.000736  9.62 6.75e-22

Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_UP_Gravettian    0.000723  0.000800  0.903  0.367
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_Krems1_1        0.000226  0.000763  0.296  0.767
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_KremsWA3      -0.0000640 0.000954 -0.0671 0.947
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice5      -0.000134  0.000774 -0.173  0.863
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice6        0.000387  0.000798  0.485  0.628
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice16      0.000154  0.000783  0.197  0.844
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_South_HG_Ostuni1  0.000736  0.000855  0.861  0.389

Chimp.REF Italy_UP_Gravettian Czech_Vestonice16 Italy_South_HG_Ostuni1  0.0115  0.00104  11.0  2.57e-28
Chimp.REF Austria_Krems1_1    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00998 0.00108  -9.27 1.82e-20
Chimp.REF Austria_KremsWA3    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00854 0.00129  -6.63 3.35e-11
Chimp.REF Czech_Vestonice5    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0142  0.00125 -11.3  1.10e-29
Chimp.REF Czech_Vestonice6    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0155  0.00114 -13.7  1.99e-42

Nothing relevatory or unexpected.
- BK1653's affinity for Italian and Central European Gravettians is a known quantity. BK1653 + Sunghir = Gravettians.
- AR33k stats = ~0, no reason to suspect them not to.
- Gravettians cluster (with internal drift) by geography, sounds reasonable.

[Norfern-Ostrobothnian]

What do you make of these stats
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_UP_Gravettian    0.00439 0.000649  6.77 1.31e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_Krems1_1      0.00596 0.000636  9.37 7.60e-21
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Austria_KremsWA3      0.00534 0.000843  6.34 2.30e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice5      0.00436 0.000652  6.69 2.26e-11
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice6      0.00427 0.000668  6.40 1.60e-10
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Czech_Vestonice16      0.00514 0.000662  7.77 8.07e-15
Chimp.REF Bulgaria_BachoKiro_LatePleistocene Russia_Kostenki14 Italy_South_HG_Ostuni1 0.00708 0.000736  9.62 6.75e-22

Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_UP_Gravettian    0.000723  0.000800  0.903  0.367
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_Krems1_1        0.000226  0.000763  0.296  0.767
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Austria_KremsWA3      -0.0000640 0.000954 -0.0671 0.947
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice5      -0.000134  0.000774 -0.173  0.863
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice6        0.000387  0.000798  0.485  0.628
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Czech_Vestonice16      0.000154  0.000783  0.197  0.844
Chimp.REF China_AmurRiver_Paleolithic Russia_Kostenki14 Italy_South_HG_Ostuni1  0.000736  0.000855  0.861  0.389

Chimp.REF Italy_UP_Gravettian Czech_Vestonice16 Italy_South_HG_Ostuni1  0.0115  0.00104  11.0  2.57e-28
Chimp.REF Austria_Krems1_1    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00998 0.00108  -9.27 1.82e-20
Chimp.REF Austria_KremsWA3    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.00854 0.00129  -6.63 3.35e-11
Chimp.REF Czech_Vestonice5    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0142  0.00125 -11.3  1.10e-29
Chimp.REF Czech_Vestonice6    Czech_Vestonice16 Italy_South_HG_Ostuni1 -0.0155  0.00114 -13.7  1.99e-42

Nothing relevatory or unexpected.
- BK1653's affinity for Italian and Central European Gravettians is a known quantity. BK1653 + Sunghir = Gravettians.
- AR33k stats = ~0, no reason to suspect them not to.
- Gravettians cluster (with internal drift) by geography, sounds reasonable.

Kale that's not BK1653 that's the IUP samples having an affinity for Gravettians without East Eurasians having any affinity for them. And we all know that Aurignacians have East Eurasian affinity too on top of Bacho Kiro IUP affinity.

[PopGenist82]

And we all know that Aurignacians have East Eurasian affinity too on top of Bacho Kiro IUP affinity.

In qpGraph, it manifests as 'extra East Eurasian' but I wonder if it represents 'non-Neaderthal admixed IUP groups' rather than actual waves of East Eurasian. 
This is supported by the observation that East Eurasian as absent in contemporaneous Eastern Europe and West Asia.

[Song]

user Enki's post on genarchivist.freeforums pointed this out. Copy-pasting it here in case it was missed. 

The only issue I can see with this is the archaic affinities not lining up. Given how early the admixture between neanderthals and Bacho Kiro was, you would certainly see this in Ust-Ishim/Tianyuan had he been ancestral to them. However, the neanderthal affinities in BK are much closer to Vindija neanderthals than Altai neanderthals which was present in Tianyuan, Ust-Ishim having only "basal" neanderthal from the near east that we all have. While these Vindija affinities are absent in early east eurasians where you see altai neanderthal and denisova alleles popping up in ANE and Tianyuan, this would signify BK/Oase being a dead branch of IUP in Europe rather than ancestral to east eurasians. There was a common ancestor likely somewhere in central asia/iran who knows that contributed to the "northern" part of tianyuan and to Ust-Ishim. Now given the fact that Vindija was a relatively recently derived branch of neanderthal, these affinities would be very unique. I will make a larger post on the genarchivist proper site for the thread you made.


We find that Bacho Kiro individuals share significantly more derived alleles with Neandertals than with the Denisovan individual (Z-score < -26.27, Fig. S7.15 and S7.16). Furthermore, the three IUP Bacho Kiro individuals and BK1653 share significantly more derived alleles with Chagyrskaya and Vindija than they do with the Altai Neandertal (-14225.08 < Z-score < -2.77, Fig. S7.15, and -5.03 < Z-score < -3.01, Fig. S7.16), similar to what has been previously observed for other ancient and present-day humans,11,17.1424

www.nature.com/articles/s41586-021-03335-3

www.nature.com/articles/s41586-022-05283-y

[Norfern-Ostrobothnian]

One thing I find interesting is that Onge seemingly has ZERO Neanderthal.
Could you verify this?
Also the affinity for Gravettian only appears when Bacho Kiro IUP is in a single population not when looked at individually. I think this might be some weird artefact so I suggest separating the samples perhaps.
Chimp.REF Altai_Neanderthal.DG Mbuti.DG ONG.SG 0.000244 0.000268 0.910 0.363
Chimp.REF Altai_Neanderthal.DG Czechia_Bohemia_UP_HG ONG.SG -0.000910 0.000370 -2.46 0.0139
Chimp.REF Altai_Neanderthal.DG YRI.SG ONG.SG 0.000719 0.000242 2.97 0.00294
Chimp.REF Bulgaria_BachoKiro_LatePleistocene_F6620 Russia_Sunghir3.SG Czech_Vestonice16 0.000581 0.000769 0.755 0.450
Chimp.REF Bulgaria_BachoKiro_LatePleistocene_CC7335 Russia_Sunghir3.SG Italy_South_HG_Ostuni1 -0.0000543 0.000810 -0.0670 0.947
The Yoruba stat can be explained by the Paleo African in Yoruba