The GenArchivist Forum
Basal Eurasian discussion - Printable Version

+- The GenArchivist Forum (https://genarchivist.com)
+-- Forum: Human Population Genetics (https://genarchivist.com/forumdisplay.php?fid=21)
+--- Forum: Ancient (aDNA) (https://genarchivist.com/forumdisplay.php?fid=181)
+--- Thread: Basal Eurasian discussion (/showthread.php?tid=644)

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

RE: Basal Eurasian discussion - Kale - 04-08-2024

I've been thinking again about Basal Eurasian and the associated deep structure therein. Typically I run giant qpgraphs on the matter consisting of 24 populations to cover every base and make sure no false positives get through that would be shot down immediately with the introduction of one of these key populations.
Show Content
However, changes can be made to the deep structure without effecting certain other areas of the graph, that is to say some areas are self-contained or 'modular' as I've been referring to it in my own head as.
For example, whatever you do in the deeper areas of the graph will not change the very straightforward solution that Native Americans are East-Asian + ANE.
Having Native Americans in the full graph just helps constrain and focus the ANE component insofar as it contributes to other populations.
I've repeatedly asked for others to sketch graphs that illustrate their ideas, but have gotten 0 returns on that.
If we are discussing at the time depth of Basal Eurasian and initial peopling of Eurasia though, many of these 'modules' are largely irrelevant, and the list of 24 populations can be trimmed significantly, hopefully encouraging people to give their hand at working on this matter. These should be the basic 10 populations. Just make sure the MA1-Iran_N relationship is addressed by MA1 into Iran, not the other way around.

Mbuti
Iran_N
ZlatyKun
Ust_Ishim
BachoKiro_IUP
China_UP (Tianyuan & AR33k)
Andaman
Goyet/Fournol
Yana
MA1

RE: Basal Eurasian discussion - kolompar - 04-08-2024

(04-08-2024, 02:29 PM)Kale Wrote: I've been thinking again about Basal Eurasian and the associated deep structure therein. Typically I run giant qpgraphs on the matter consisting of 24 populations to cover every base and make sure no false positives get through that would be shot down immediately with the introduction of one of these key populations.
Show Content
However, changes can be made to the deep structure without effecting certain other areas of the graph, that is to say some areas are self-contained or 'modular' as I've been referring to it in my own head as.
For example, whatever you do in the deeper areas of the graph will not change the very straightforward solution that Native Americans are East-Asian + ANE.
Having Native Americans in the full graph just helps constrain and focus the ANE component insofar as it contributes to other populations.
I've repeatedly asked for others to sketch graphs that illustrate their ideas, but have gotten 0 returns on that.
If we are discussing at the time depth of Basal Eurasian and initial peopling of Eurasia though, many of these 'modules' are largely irrelevant, and the list of 24 populations can be trimmed significantly, hopefully encouraging people to give their hand at working on this matter. These should be the basic 10 populations. Just make sure the MA1-Iran_N relationship is addressed by MA1 into Iran, not the other way around.

Mbuti
Iran_N
ZlatyKun
Ust_Ishim
BachoKiro_IUP
China_UP (Tianyuan & AR33k)
Andaman
Goyet/Fournol
Yana
MA1

Do you think Anatolia/Caucasus and Kostenki/Gravettian are fully settled or why aren't they in the basic pops?
I've posted a graphlet like this before, what format do you want? This is mostly with your populations except with Kostenki instead of Goyet, surely Goyet should be the same thing with just some extra BachoKiro.
Show Content
The Iran/ANE relationship is a hard part because I think it might not be a simple one-way admixture, with more recent haplogroups X, R1a and R1b (mtDNA) shared between them.

Also, there's a plot_graph_map function in admixtools2, have you tried that? Maybe we could see big things with it, like the Persian plateau hub Big Grin

RE: Basal Eurasian discussion - Kale - 04-08-2024

Pinarbasi/Caucasus_UP/Kostenki/Sunghir/Muierii/Gravettian, fall into a number of those 'modules'. You can do a number of things with them, but what will not change is that they all function well as something intermediate between Goyet and Iran_N (minus ANE). I know Chad is in agreement about the idea of Iran_N representing a significant pole in the Near-East/Basal/West arena of things. Not sure if he's also on board with Goyet being the opposite end of that spectrum.

In regards to sketching out ideas, I was more talking more about people that don't/can't run qpgraph, so I meant sketch literally if need be.
Hmm didn't know about plot_graph_map, I'll have to see what that looks like.

EDIT: Don't know what's going wrong with plot_graph_map
Error in plot_graph_map(graph, leafcoords):
all(c("iid", "group", "lat", "lon") %in% names(leafcoords)) is not TRUE

Maybe I misunderstand how this works? Do I have to supply coordinates for the mixture events or just the populations in the leafcoords file?

RE: Basal Eurasian discussion - kolompar - 04-08-2024

(04-08-2024, 05:31 PM)Kale Wrote: Pinarbasi/Caucasus_UP/Kostenki/Sunghir/Muierii/Gravettian, fall into a number of those 'modules'. You can do a number of things with them, but what will not change is that they all function well as something intermediate between Goyet and Iran_N (minus ANE). I know Chad is in agreement about the idea of Iran_N representing a significant pole in the Near-East/Basal/West arena of things. Not sure if he's also on board with Goyet being the opposite end of that spectrum.

In regards to sketching out ideas, I was more talking more about people that don't/can't run qpgraph, so I meant sketch literally if need be.
Hmm didn't know about plot_graph_map, I'll have to see what that looks like.

EDIT: Don't know what's going wrong with plot_graph_map
Error in plot_graph_map(graph, leafcoords):
all(c("iid", "group", "lat", "lon") %in% names(leafcoords)) is not TRUE

Maybe I misunderstand how this works? Do I have to supply coordinates for the mixture events or just the populations in the leafcoords file?
There's an example_anno object that shows what the format should be, I think it's best if you make one from the real anno file for just your populations with that four columns. But maybe not worth it as the result isn't really spectacular.
   

(04-08-2024, 12:28 AM)Desdonas Wrote: The significant bottleneck (over 10k) of C, F and D1 relative to E and D2 does indicate that the OoA population has a relatively isolated history (likely geographically). Of course, it is impossible to have an Arabian Standstill over 30k and an OoA of 80kya, but I believe it is reasonable for them to list these alleles. Perhaps this is a part of the OoA bottleneck itself, and if the population size is not large, the selection may complete quickly (during 10k?). Or we need more genome from East Africa to check whether pre-C and pre-F exist.

Second, Vallini's "Tianyuan-K14" binary model may indeed appear more "traditional" in 2024. But in terms of climate, the Iranian Plateau is indeed feasible as a hub for Eurasians, and the allele selection may also occur during the migration from East Africa to Iran (65-55kya).

Finally, Eurasians may include multiple components, including Crown Eurasians, ZK/para ZK, and a deeper BE. The third may come from the Arabian Peninsula, while the first two may both have roots in Iran (47kya and 50kya?). Para-ZK may be represented by y-hg G (also IJ?), while deep BE may be represented by some mt-N's primary branches. They have different split depth but may both contribute to West Eurasian populations.
We are being diplomatic but that study is almost useless. Why not try to fit a tree onto their map like I just did, or at least calculate the allele sharing for a few more populations. What were they even thinking? "Let's gather the greatest scientists, we are calculating allele sharing with Kostenki and Tianyuan" Big Grin
They don't show that Iran was feasible as a hub, it was barely habitable, it's just that other regions like Arabia were even less so and they just wanted to choose a hub somewhere there. They actually suggest a decreased suitability for Iran in the period for their standstill.
At least they show that North India had a much higher sustainable population, makes it the right place for an East Eurasian hub, while the small West/Basal population was waiting for their big Heinrich event and Campanian eruption. Iran and the Central Asian desert was more like the divider between them, just like how it is today too. Occam says this two-way model is better than East+West+Basal+ZK+etc.
[Image: 41467_2024_46161_Fig4_HTML.png?as=webp]

RE: Basal Eurasian discussion - Jerome - 04-09-2024

(04-07-2024, 04:44 PM)TanTin Wrote: Next, I will give you the same results for  Denisova3-Neanderthal , but will use a different arrangement to correspond the time - space location for the above groups.
The arrangement is made by using another F4 statistics, that I am not showing here.
[Image: Neand-Denisova-all-3.png]

Thank you.
Btw could you add those various east Eurasian populations that you did with on page 1?

RE: Basal Eurasian discussion - Jerome - 04-09-2024

(04-07-2024, 03:23 PM)TanTin Wrote:
(04-07-2024, 11:28 AM)Jerome Wrote: Could you add Kostenki,afontovaGora,Tutkaul and Tyumen_HG here, dosent this show WHG as having more Neanderthal than MA1?
And iran_n has more Neanderthal affinity(2.16) than ust ishim(1.87)?

Dosent this refute lazaridis conception of basal eurasian as needed to account for lower Neanderthal Ancestry?
Maybe deselection against Neanderthal alleles could explain that??

Why else is basal eurasian even needed now?

There's also the fact that iran and central Asia was full of Neanderthals(warsawi,Dodekatym-2) so the area being full of basal eurasians dosent make sense.
Any basal eurasian living there would end up with lots of Neanderthal Ancestry which defeats the purpose of basal itself.


I did the rerun.

Show Content

[Image: F4-Kostenki.png]

Are these scores based on z value or the weights?

RE: Basal Eurasian discussion - Jerome - 04-09-2024

Does anyone know if deselection against Neanderthal alleles can lower the apparent allele sharing?

RE: Basal Eurasian discussion - Desdonas - 04-09-2024

@Kale @kolompar

What do you think of India as a hub for early Eurasian population? I have seen that many haplogroup migration maps locate the origin of F in India, and India seems to have some unresolved F*. If this would be true, the main OoA bottleneck could be in there. IJ and G (west/basal) headed west first, then C, D1, F2, H and K stayed here and shared extra bottleneck for about 3,000 years. India is on a similar temperature zone of the Arabian peninsula, and the OoA allele selection may be possible.

However, if India is not the hub, then there may be some issues on the "Arabian Standstill". Do you agree that this selection/OoA bottleneck had already happened within Africa, and there was no real standstill after the exodus out of Africa (55-50kya)? If so, maybe all OoA related elements (pre-C, pre-F, pre-D1, etc.) were taken over by the E-population later.

RE: Basal Eurasian discussion - Jerome - 04-09-2024

I think it was more likely the regions around Caspian sea and western central Asia.
From there you can easily head to Caucasus and Anatolia on one way and central Asia - Siberia on other,giving rise to early IUP industries like kulbalkian.

South Asia is less likely, because of its late IUP/EUP industries in centre.
They seem to be mostly present in south or in NW,where they are connected with central Asian industries.

RE: Basal Eurasian discussion - Kale - 04-09-2024

(04-09-2024, 03:19 AM)Desdonas Wrote: @Kale @kolompar

What do you think of India as a hub for early Eurasian population? I have seen that many haplogroup migration maps locate the origin of F in India, and India seems to have some unresolved F*. If this would be true, the main OoA bottleneck could be in there. IJ and G (west/basal) headed west first, then C, D1, F2, H and K stayed here and shared extra bottleneck for about 3,000 years. India is on a similar temperature zone of the Arabian peninsula, and the OoA allele selection may be possible.

However, if India is not the hub, then there may be some issues on the "Arabian Standstill". Do you agree that this selection/OoA bottleneck had already happened within Africa, and there was no real standstill after the exodus out of Africa (55-50kya)? If so, maybe all OoA related elements (pre-C, pre-F, pre-D1, etc.) were taken over by the E-population later.

Modern distribution seems to mean less and less the more ancient samples we get. Imagine if BachoKiro's pre-F2'4, Kostenki's C1b, Oase's K2a-M2308 and mt-pre-N, Goyet & Ostuni's Mt-M, etc. all survived? I think India and Southeast Asia somehow just preserved more early diversity, not that they are the source of it. 
In regards to rapid OOA vs Arabian standstill, I haven't looked into the idea enough to really form a perception.

RE: Basal Eurasian discussion - Merriku - 04-09-2024

(04-09-2024, 03:19 AM)Desdonas Wrote: @Kale @kolompar

What do you think of India as a hub for early Eurasian population? I have seen that many haplogroup migration maps locate the origin of F in India, and India seems to have some unresolved F*. If this would be true, the main OoA bottleneck could be in there. IJ and G (west/basal) headed west first, then C, D1, F2, H and K stayed here and shared extra bottleneck for about 3,000 years. India is on a similar temperature zone of the Arabian peninsula, and the OoA allele selection may be possible.

However, if India is not the hub, then there may be some issues on the "Arabian Standstill". Do you agree that this selection/OoA bottleneck had already happened within Africa, and there was no real standstill after the exodus out of Africa (55-50kya)? If so, maybe all OoA related elements (pre-C, pre-F, pre-D1, etc.) were taken over by the E-population later.

I am skeptical of those Indian Fs. Considering the amount of Indian private samples I think F should have turned up by now.
I think those Indian F*s are from old (STR based?) studies. If I am not mistaken, when these studies were published haplogroup H3 wasn't known & many of those F* may actually be H3.
I don't know about FTDNA but both theytree and yfull don't have any Indian (F ×GHIJK) samples.
 The only clue we have is 2 F2 samples from the user Tipirneni's 23&me matches , but then again it could just be the case of older nomenclature.

RE: Basal Eurasian discussion - Norfern-Ostrobothnian - 04-09-2024

That Ju-Hoan stat is very interesting
I have three propositions
Either the South Africans have genuine Neanderthal ancestry, they have some deeper Denisovan related ancestry or the Human ancestry in Neanderthals is closest to South Africans in particular.

Today this wouldn't make sense, but there's always a chance South Africans used to reside closer to Neanderthals than the ancestors of all other Humans did.
The low stat for West Africans can be explained with archaic ancestry, but South Africans are higher than many Eurasians are.

RE: Basal Eurasian discussion - TanTin - 04-09-2024

(04-09-2024, 01:40 AM)Jerome Wrote:
(04-07-2024, 03:23 PM)TanTin Wrote:
(04-07-2024, 11:28 AM)Jerome Wrote: Could you add Kostenki,afontovaGora,Tutkaul and Tyumen_HG here, dosent this show WHG as having more Neanderthal than MA1?
And iran_n has more Neanderthal affinity(2.16) than ust ishim(1.87)?

Dosent this refute lazaridis conception of basal eurasian as needed to account for lower Neanderthal Ancestry?
Maybe deselection against Neanderthal alleles could explain that??

Why else is basal eurasian even needed now?

There's also the fact that iran and central Asia was full of Neanderthals(warsawi,Dodekatym-2) so the area being full of basal eurasians dosent make sense.
Any basal eurasian living there would end up with lots of Neanderthal Ancestry which defeats the purpose of basal itself.


I did the rerun.

Show Content

[Image: F4-Kostenki.png]

Are these scores based on z value or the weights?

These are the F4. The first column - for estimation.
The vertical intervals are for the error.
Z is not shown on the picture, but it is in the table in the 3-rd column.
Z = est/se . se is the standard error.

RE: Basal Eurasian discussion - TanTin - 04-09-2024

(04-09-2024, 11:05 AM)Norfern-Ostrobothnian Wrote: That Ju-Hoan stat is very interesting
I have three propositions
Either the South Africans have genuine Neanderthal ancestry, they have some deeper Denisovan related ancestry or the Human ancestry in Neanderthals is closest to South Africans in particular.

Today this wouldn't make sense, but there's always a chance South Africans used to reside closer to Neanderthals than the ancestors of all other Humans did.
The low stat for West Africans can be explained with archaic ancestry, but South Africans are higher than many Eurasians are.

I did additional tests yesterday to find where this Neanderthal component came for South Africans.
It is not genuine Neanderthal.
They admixed with some North African archaic group which had already the Neanderthal components.
And that's why they score high .
Before to expand to Europe and Asia, the Neanderthals were obviously in North Africa.

RE: Basal Eurasian discussion - Norfern-Ostrobothnian - 04-09-2024

The level is so high that it'd require the likely East African Nilotic or Cushitic ancestry to be a very large portion of their ancestry. Either this is artefacting, which it might be, or this is genuine ancestry from Neanderthal related populations.