Basal Eurasian discussion

[kolompar]

The affinity for Zlaty Kun in Iranian HG's is likely from their Baradostian ancestry, which we don't have samples of, and which is not Zlaty Kun, just some early branch of Eurasians. Perhaps this is what those early researchers called "Basal Eurasian": a mixture of Baradostian (admixture calculators go for Zlaty Kun as a proxy in the absence of actual Baradostian samples) and "ANA" (quite possibly Emiran). Zlaty Kun had Neanderthal ancestry though, and probably so did Baradostian.

If I recall, even Iberomaurusians had higher than average Neanderthal ancestry, which doesn't make a lot of sense if operating under the assumption that ANA is really some sort of genuinely African type ancestry. I'm skeptical that the ANA-side of IBM is really some deeply rooted local Aterian relict and not just some very old Egyptian or Levantine (Emiran as you mentioned) reflux.

They clearly have European HG ancestry, maybe all of their non-ANA is simply WHG-like, so that's where their Neanderthal comes from. I also think Aterian is too old, haplogroup E probably came from East Africa less than 40 thousand years ago, and that should be most of the African part.

Aren't Nilote's y-hg A mostly/all A-M13? Y-full and FtDNA both have the TRMCA of M13 at ~10KBC. Probably a later incursion from the South.
Y-hg B seems to have most of it's diversity in Central Africa.
At some point in the past, it is possible East Africa may not have had y-hg A or B, just E (and D0).

Could be just a bottleneck? I used to think CT is Eurasian because it would be unlikely that C and F (and D?) all went extinct in Africa, but there really isn't all that much diversity in Africa, something like 10 branches at the OoA time depth with E and D0 both there from CT. So likely a lot went extinct, plus sampling is also way behind, maybe something could turn up eventually.

What kind of lame Basal Eurasian is that if it's not more related to populations with Basal Eurasian?

What I was trying to illustrate regarding point 3 was this.
Congo_Mbuti.DG Sunghir.SG      ZlatyKun.SG Ust_Ishim.DG  0.00278  4.44  1127764
Congo_Mbuti.DG MA1.SG          ZlatyKun.SG Ust_Ishim.DG  0.00285  3.89  791393
Congo_Mbuti.DG Andaman_100BP.SG ZlatyKun.SG Ust_Ishim.DG  0.00289  4.22  1111692
Congo_Mbuti.DG Iran_Wezmeh_N.SG ZlatyKun.SG Ust_Ishim.DG -0.00048 -0.75 1126449
If Iran_N is ~60% of it's ancestry from something in the neighborhood of Sunghir/MA1/Andaman, and there is no connection between ZlatyKun and Iran_N, shouldn't that last stat have 60% of the d-value of the first 3?

Well, what could be wrong there...

Code:

Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_Ust_Ishim.DG -8.73E-05 0.000547693 -0.159441747 0.87332085 1024958
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG China_Tianyuan 0.001100443 0.000547264 2.010810096 0.044345519 841640
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_Kostenki14 0.004344356 0.000534625 8.125992513 4.44E-16 1003557
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_MA1_HG.SG 0.006491183 0.000560707 11.57679177 5.40E-31 729232
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Indian_GreatAndaman_100BP.SG 0.001585569 0.000522755 3.033100144 0.002420553 1016556

It might be something in Ust-Ishim, the same thing that also makes him less related to East Eurasians than he should be.
The other option is this is proof there was Iran/Basal gene flow into West Eurasians.

Let me add something about this Z- requirement.
It is valid only when we do statistics for some group of individuals or for some population.
However we can run F-statistics for individuals only.
In our case: ZK - is one individual only.
BK are 4 or 5, but we may run Fstats for for whichever we like. 
Laos_Hoabinhian.SG  - is pretty much the same..
So to avoid this inconvenience with Z-requirement - the simple step to bypass it:  just select one individual  and run the F-test with it.
In such case F-stats will just count the number of shared alleles between these 4 individuals and it will provide results in the form of simple count and percentage.. 
Z-number doesn't matter in such case. More especially in the case where est  is near 0 .  
Dividing near 0 value with some small number is not much recommended.
If our test populations are a group with significant numbers, then we must take into account Z-value. however when working with individuals we may just ignore this.

Z is only a requirement to show there was gene flow. Maybe you can't prove a negative but if Z is close to zero you have no evidence of gene flow, that's how f-stats are usually used.
Don't really know about bias but f4 should be less affected by it if I go by this.
https://uqrmaie1.github.io/admixtools/ar...stats.html

F4( TEST, NA19012.SG ,  Vindija_snpAD.DG, Loschbour.DG )

For Africans that will be positive because Japan and Loschbour are both OoA bottlenecked. Will a simple PCA not work? First component should separate modern and archaic.

In fact, until now I still believe that BE split from other Eurasians before 55-60 kya. But if there is not much Neanderthal in Oceanians and Onge, then the Neanderthal pulse shared by all Eurasians at around 55 kya may be the similar size. The extra Neanderthal in Oase, Bacho Kiro and East Asians is secondary. Furthermore, BE may not be very basal, while Crown Eurasian may be more derived than expected.

It was Hoabinhian, Jomon, and maybe Longlin and AASI. For Oceanians I don't know because I can't really calculate their Neanderthal from just f-stats because of their Denisovan. For East Asians it might also be that their extra archaic affinity comes from the Denisovan that the Salkhit paper detected. I don't know how much the European ancients are considered for the usual suggestion that there was only one Neanderthal admixture event. Maybe all admixing Neanderthal was very similar, I think around the Altai they also got replaced by Vindija ones?

[TanTin]

Let me add something about this Z- requirement.
It is valid only when we do statistics for some group of individuals or for some population.
However we can run F-statistics for individuals only.
In our case: ZK - is one individual only.
BK are 4 or 5, but we may run Fstats for for whichever we like. 
Laos_Hoabinhian.SG  - is pretty much the same..
So to avoid this inconvenience with Z-requirement - the simple step to bypass it:  just select one individual  and run the F-test with it.
In such case F-stats will just count the number of shared alleles between these 4 individuals and it will provide results in the form of simple count and percentage.. 
Z-number doesn't matter in such case. More especially in the case where est  is near 0 .  
Dividing near 0 value with some small number is not much recommended.
If our test populations are a group with significant numbers, then we must take into account Z-value. however when working with individuals we may just ignore this.

Z is only a requirement to show there was gene flow. Maybe you can't prove a negative but if Z is close to zero you have no evidence of gene flow, that's how f-stats are usually used.
Don't really know about bias but f4 should be less affected by it if I go by this.
https://uqrmaie1.github.io/admixtools/ar...stats.html

F4( TEST, NA19012.SG ,  Vindija_snpAD.DG, Loschbour.DG )

For Africans that will be positive because Japan and Loschbour are both OoA bottlenecked. Will a simple PCA not work? First component should separate modern and archaic.

F-stats will measure the difference. Imagine 2 individuals who have almost the same profile.
If they share the same markers - there will be no big difference. More equal they are to each other - lower is the est (F4 estimation) . se (standard error ) could be very low too. 
Az Z is calculated by est/se, then Z will be also near 0 for such example because est is low.

So we may have very closely related individuals, but F4 to show 0 est /  Z near 0 and no   evidence of gene flow . Because this function will measure the difference.  If no difference  - we may see such 0 results, however these could be some individuals from the same population, sharing the same markers.

[Desdonas]

The affinity for Zlaty Kun in Iranian HG's is likely from their Baradostian ancestry, which we don't have samples of, and which is not Zlaty Kun, just some early branch of Eurasians. Perhaps this is what those early researchers called "Basal Eurasian": a mixture of Baradostian (admixture calculators go for Zlaty Kun as a proxy in the absence of actual Baradostian samples) and "ANA" (quite possibly Emiran). Zlaty Kun had Neanderthal ancestry though, and probably so did Baradostian.

If I recall, even Iberomaurusians had higher than average Neanderthal ancestry, which doesn't make a lot of sense if operating under the assumption that ANA is really some sort of genuinely African type ancestry. I'm skeptical that the ANA-side of IBM is really some deeply rooted local Aterian relict and not just some very old Egyptian or Levantine (Emiran as you mentioned) reflux.

They clearly have European HG ancestry, maybe all of their non-ANA is simply WHG-like, so that's where their Neanderthal comes from. I also think Aterian is too old, haplogroup E probably came from East Africa less than 40 thousand years ago, and that should be most of the African part.

Aren't Nilote's y-hg A mostly/all A-M13? Y-full and FtDNA both have the TRMCA of M13 at ~10KBC. Probably a later incursion from the South.
Y-hg B seems to have most of it's diversity in Central Africa.
At some point in the past, it is possible East Africa may not have had y-hg A or B, just E (and D0).

Could be just a bottleneck? I used to think CT is Eurasian because it would be unlikely that C and F (and D?) all went extinct in Africa, but there really isn't all that much diversity in Africa, something like 10 branches at the OoA time depth with E and D0 both there from CT. So likely a lot went extinct, plus sampling is also way behind, maybe something could turn up eventually.

What kind of lame Basal Eurasian is that if it's not more related to populations with Basal Eurasian?

What I was trying to illustrate regarding point 3 was this.
Congo_Mbuti.DG Sunghir.SG      ZlatyKun.SG Ust_Ishim.DG  0.00278  4.44  1127764
Congo_Mbuti.DG MA1.SG          ZlatyKun.SG Ust_Ishim.DG  0.00285  3.89  791393
Congo_Mbuti.DG Andaman_100BP.SG ZlatyKun.SG Ust_Ishim.DG  0.00289  4.22  1111692
Congo_Mbuti.DG Iran_Wezmeh_N.SG ZlatyKun.SG Ust_Ishim.DG -0.00048 -0.75 1126449
If Iran_N is ~60% of it's ancestry from something in the neighborhood of Sunghir/MA1/Andaman, and there is no connection between ZlatyKun and Iran_N, shouldn't that last stat have 60% of the d-value of the first 3?

Well, what could be wrong there...

Code:

Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_Ust_Ishim.DG -8.73E-05 0.000547693 -0.159441747 0.87332085 1024958
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG China_Tianyuan 0.001100443 0.000547264 2.010810096 0.044345519 841640
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_Kostenki14 0.004344356 0.000534625 8.125992513 4.44E-16 1003557
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Russia_MA1_HG.SG 0.006491183 0.000560707 11.57679177 5.40E-31 729232
Mbuti.DG Iran_GanjDareh_N Czechia_Bohemia_UP_HG Indian_GreatAndaman_100BP.SG 0.001585569 0.000522755 3.033100144 0.002420553 1016556

It might be something in Ust-Ishim, the same thing that also makes him less related to East Eurasians than he should be.
The other option is this is proof there was Iran/Basal gene flow into West Eurasians.

Let me add something about this Z- requirement.
It is valid only when we do statistics for some group of individuals or for some population.
However we can run F-statistics for individuals only.
In our case: ZK - is one individual only.
BK are 4 or 5, but we may run Fstats for for whichever we like. 
Laos_Hoabinhian.SG  - is pretty much the same..
So to avoid this inconvenience with Z-requirement - the simple step to bypass it:  just select one individual  and run the F-test with it.
In such case F-stats will just count the number of shared alleles between these 4 individuals and it will provide results in the form of simple count and percentage.. 
Z-number doesn't matter in such case. More especially in the case where est  is near 0 .  
Dividing near 0 value with some small number is not much recommended.
If our test populations are a group with significant numbers, then we must take into account Z-value. however when working with individuals we may just ignore this.

Z is only a requirement to show there was gene flow. Maybe you can't prove a negative but if Z is close to zero you have no evidence of gene flow, that's how f-stats are usually used.
Don't really know about bias but f4 should be less affected by it if I go by this.
https://uqrmaie1.github.io/admixtools/ar...stats.html

F4( TEST, NA19012.SG ,  Vindija_snpAD.DG, Loschbour.DG )

For Africans that will be positive because Japan and Loschbour are both OoA bottlenecked. Will a simple PCA not work? First component should separate modern and archaic.

In fact, until now I still believe that BE split from other Eurasians before 55-60 kya. But if there is not much Neanderthal in Oceanians and Onge, then the Neanderthal pulse shared by all Eurasians at around 55 kya may be the similar size. The extra Neanderthal in Oase, Bacho Kiro and East Asians is secondary. Furthermore, BE may not be very basal, while Crown Eurasian may be more derived than expected.

It was Hoabinhian, Jomon, and maybe Longlin and AASI. For Oceanians I don't know because I can't really calculate their Neanderthal from just f-stats because of their Denisovan. For East Asians it might also be that their extra archaic affinity comes from the Denisovan that the Salkhit paper detected. I don't know how much the European ancients are considered for the usual suggestion that there was only one Neanderthal admixture event. Maybe all admixing Neanderthal was very similar, I think around the Altai they also got replaced by Vindija ones?

1. IBM's non-ANA ancestry, except the Near Eastern Pinarbasi like component, there may also be some Goyet-like geneflow. Mtdna-M1 has an epicenter in Africa/IBM-like genepool, which makes some people even consider that M1 is ANA, while MxM1 is OoA. In fact, M1 could be associated with an unsampled Goyet-related population from Europe. (Migrated through the Mediterranean?)

2. Y-hg E has a TMRCA of about 52kya rather than 40kya. But during 52-40kya, E may diversified in East Africa, and then expanded to North Africa and West Africa.

3. The Altai-like Neanderthal lineage was indeed replaced by a new Neanderthal lineage (Chagyrskaya). From the graph below we can infer that Chagyrskaya is significantly Vindija-related.

[Horatio McCallister]

The affinity for Zlaty Kun in Iranian HG's is likely from their Baradostian ancestry, which we don't have samples of, and which is not Zlaty Kun, just some early branch of Eurasians. Perhaps this is what those early researchers called "Basal Eurasian": a mixture of Baradostian (admixture calculators go for Zlaty Kun as a proxy in the absence of actual Baradostian samples) and "ANA" (quite possibly Emiran). Zlaty Kun had Neanderthal ancestry though, and probably so did Baradostian.

If I recall, even Iberomaurusians had higher than average Neanderthal ancestry, which doesn't make a lot of sense if operating under the assumption that ANA is really some sort of genuinely African type ancestry. I'm skeptical that the ANA-side of IBM is really some deeply rooted local Aterian relict and not just some very old Egyptian or Levantine (Emiran as you mentioned) reflux.

They clearly have European HG ancestry, maybe all of their non-ANA is simply WHG-like, so that's where their Neanderthal comes from. I also think Aterian is too old, haplogroup E probably came from East Africa less than 40 thousand years ago, and that should be most of the African part.

Aren't Nilote's y-hg A mostly/all A-M13? Y-full and FtDNA both have the TRMCA of M13 at ~10KBC. Probably a later incursion from the South.
Y-hg B seems to have most of it's diversity in Central Africa.
At some point in the past, it is possible East Africa may not have had y-hg A or B, just E (and D0).

Could be just a bottleneck? I used to think CT is Eurasian because it would be unlikely that C and F (and D?) all went extinct in Africa, but there really isn't all that much diversity in Africa, something like 10 branches at the OoA time depth with E and D0 both there from CT. So likely a lot went extinct, plus sampling is also way behind, maybe something could turn up eventually.

1. The point isn't that IBM had any Neanderthal ancestry at all, it's that their Neanderthal levels were higher than you would expect being a half/half West Eurasian + ANA mix. I went back to the original IBM paper's supplements ( Pleistocene North African genomes link Near Eastern and sub-Saharan African human populations | Science) to see what they had to say about this - "Considering the dual ancestry of the Taforalt individuals, we can explain the Altai affinity in Taforalt as a dilution of its Natufian-related ancestry with its significant proportion (~36.5%) of sub-Saharan African ancestry." However, when looking at the f4s they used to determine this, some things stood out as unusual:

Imgur: The magic of the Internet

IBM has comparable Z scores to Natufians and Levant_N, even though the latter obviously have proportionately much less ANA than IBM does, although to be fair they do have lowish SNP counts. Also Iran_N has lowest Neanderthal out of all the Eurasians in that test and even lower than some of the Africans. There are some other odd results too like the high scores for Mozabites and Ju_Huan North, I would be interested in seeing if someone here could possibly re-run these or similar tests as a sanity check.

2. There have been lots of papers over the last few years showing that the divergence between Eurasians and Africans is very deep, going back 100k years. I don't think there's any real reason to assume that this proto-Eurasian coalescence (CT and L3) needed to have taken place in Africa. The below paper makes the case that the proto-Eurasian population spent an extended period of time coalescing in the neighborhood of Arabia for tens of thousands of years before disseminating across Eurasia as a whole, which sounds a lot more likely to me than the conventional OoA scenario whereby a branch of humans leaves East Africa 55,000-60,000 years ago and rapidly colonizes and replaces Neanderthals and Denisovans from all parts of Eurasia.

pnas.2213061120

[Desdonas]

The affinity for Zlaty Kun in Iranian HG's is likely from their Baradostian ancestry, which we don't have samples of, and which is not Zlaty Kun, just some early branch of Eurasians. Perhaps this is what those early researchers called "Basal Eurasian": a mixture of Baradostian (admixture calculators go for Zlaty Kun as a proxy in the absence of actual Baradostian samples) and "ANA" (quite possibly Emiran). Zlaty Kun had Neanderthal ancestry though, and probably so did Baradostian.

If I recall, even Iberomaurusians had higher than average Neanderthal ancestry, which doesn't make a lot of sense if operating under the assumption that ANA is really some sort of genuinely African type ancestry. I'm skeptical that the ANA-side of IBM is really some deeply rooted local Aterian relict and not just some very old Egyptian or Levantine (Emiran as you mentioned) reflux.

They clearly have European HG ancestry, maybe all of their non-ANA is simply WHG-like, so that's where their Neanderthal comes from. I also think Aterian is too old, haplogroup E probably came from East Africa less than 40 thousand years ago, and that should be most of the African part.

Aren't Nilote's y-hg A mostly/all A-M13? Y-full and FtDNA both have the TRMCA of M13 at ~10KBC. Probably a later incursion from the South.
Y-hg B seems to have most of it's diversity in Central Africa.
At some point in the past, it is possible East Africa may not have had y-hg A or B, just E (and D0).

Could be just a bottleneck? I used to think CT is Eurasian because it would be unlikely that C and F (and D?) all went extinct in Africa, but there really isn't all that much diversity in Africa, something like 10 branches at the OoA time depth with E and D0 both there from CT. So likely a lot went extinct, plus sampling is also way behind, maybe something could turn up eventually.

1. The point isn't that IBM had any Neanderthal ancestry at all, it's that their Neanderthal levels were higher than you would expect being a half/half West Eurasian + ANA mix. I went back to the original IBM paper's supplements ( Pleistocene North African genomes link Near Eastern and sub-Saharan African human populations | Science) to see what they had to say about this - "Considering the dual ancestry of the Taforalt individuals, we can explain the Altai affinity in Taforalt as a dilution of its Natufian-related ancestry with its significant proportion (~36.5%) of sub-Saharan African ancestry." However, when looking at the f4s they used to determine this, some things stood out as unusual:

Imgur: The magic of the Internet

IBM has comparable Z scores to Natufians and Levant_N, even though the latter obviously have proportionately much less ANA than IBM does, although to be fair they do have lowish SNP counts. Also Iran_N has lowest Neanderthal out of all the Eurasians in that test and even lower than some of the Africans. There are some other odd results too like the high scores for Mozabites and Ju_Huan North, I would be interested in seeing if someone here could possibly re-run these or similar tests as a sanity check.

2. There have been lots of papers over the last few years showing that the divergence between Eurasians and Africans is very deep, going back 100k years. I don't think there's any real reason to assume that this proto-Eurasian coalescence (CT and L3) needed to have taken place in Africa. The below paper makes the case that the proto-Eurasian population spent an extended period of time coalescing in the neighborhood of Arabia for tens of thousands of years before disseminating across Eurasia as a whole, which sounds a lot more likely to me than the conventional OoA scenario whereby a branch of humans leaves East Africa 55,000-60,000 years ago and rapidly colonizes and replaces Neanderthals and Denisovans from all parts of Eurasia.

pnas.2213061120

Anyway, the y-hg E related population seems to be outgroup of all Eurasian-proper. Although the deep divergence between Africans and Eurasians can be explained by deep ancestry like A00, A and B, but if there was an early back migration (E, D2) from Arabia to Africa, we may see the "Arabian Standstill" alleles in Africans significantly.

An alternative scenario is that, follow the conventional OoA theory but consider that the chronology of BT, CT, DE and D may be older than expected. The following graph shows that D2(D0) split from D1-M174 at about 71kya.

[Horatio McCallister]

Anyway, the y-hg E related population seems to be outgroup of all Eurasian-proper. Although the deep divergence between Africans and Eurasians can be explained by deep ancestry like A00, A and B, but if there was an early back migration (E, D2) from Arabia to Africa, we may see the "Arabian Standstill" alleles in Africans significantly.

You are ignoring the other half of the deep ancestry in Africans, the mtDNA side. Non-L3 lineages are the majority of mtDNA in Africans, that's where the bulk of their actual deep African ancestry is going to derive, not peripheral lineages like A00.

[TanTin]

1. The point isn't that IBM had any Neanderthal ancestry at all, it's that their Neanderthal levels were higher than you would expect being a half/half West Eurasian + ANA mix. I went back to the original IBM paper's supplements ( Pleistocene North African genomes link Near Eastern and sub-Saharan African human populations | Science) to see what they had to say about this - "Considering the dual ancestry of the Taforalt individuals, we can explain the Altai affinity in Taforalt as a dilution of its Natufian-related ancestry with its significant proportion (~36.5%) of sub-Saharan African ancestry." However, when looking at the f4s they used to determine this, some things stood out as unusual:

Imgur: The magic of the Internet



IBM has comparable Z scores to Natufians and Levant_N, even though the latter obviously have proportionately much less ANA than IBM does, although to be fair they do have lowish SNP counts. Also Iran_N has lowest Neanderthal out of all the Eurasians in that test and even lower than some of the Africans. There are some other odd results too like the high scores for Mozabites and Ju_Huan North, I would be interested in seeing if someone here could possibly re-run these or similar tests as a sanity check.

It seems these results are from older dataset. As the group names in the latest datasets are different.
So I did some selection of groups similar to those in the test.
Will show the results below.

[TanTin]

Show Content

Spoiler
  pop1          pop2                pop3  pop4                        est      se      z        p      n
  <chr>        <chr>                <chr>  <chr>                      <dbl>    <dbl>  <dbl>    <dbl>  <dbl>
1 Primate_Chimp Altai_Neanderthal.DG Yoruba Primate_Chimp          -0.0959  0.000530 -181.  0        592996
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Altai_Neanderthal.DG    0.0798  0.000570  140.  0        592996
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Russia_Ust_Ishim.DG      0.00168 0.000352    4.76 1.96e- 6 592969
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Russia_MA1_HG.SG        0.00181 0.000358    5.05 4.41e- 7 425966
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Luxembourg_Loschbour.DG  0.00156 0.000290    5.39 6.87e- 8 592749
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Bulgaria_N              0.00146 0.000268    5.44 5.20e- 8 508318
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Iran_GanjDareh_N        0.00117 0.000231    5.05
 4.39e- 7 562236
8 Primate_Chimp Altai_Neanderthal.DG Yoruba Iran_HajjiFiruz_N        0.00139 0.000247    5.62 1.96e- 8 539981
9 Primate_Chimp Altai_Neanderthal.DG Yoruba Iranian                  0.00133 0.000177    7.52 5.30e-14 592996
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Mozabite                0.00125 0.000159    7.88 3.18e-15 592996

1 Primate_Chimp Altai_Neanderthal.DG Yoruba Georgia_Kotias.SG      0.00162  0.000298  5.43 5.70e- 8 591680
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Georgia_Satsurblia.SG  0.00113  0.000365  3.09 1.98e- 3 352936
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Ju_hoan_North          0.00116  0.000170  6.82 9.27e-12 592996
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Egyptian              0.00111  0.000160  6.95 3.71e-12 592996
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_Natufian        0.000790 0.000522  1.51 1.30e- 1 115788
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_Natufian_contam 0.00138  0.000403  3.43 5.94e- 4 249108
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Algerian              0.00106  0.000167  6.36 1.97e-10 592996
8 Primate_Chimp Altai_Neanderthal.DG Yoruba South_Africa_400BP.SG  0.000189 0.000159  1.19 2.36e- 1 592957
9 Primate_Chimp Altai_Neanderthal.DG Yoruba South_Africa_1900BP.SG 0.000863 0.000249  3.47 5.15e- 4 592913
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Saharawi              0.000907 0.000164  5.52 3.46e- 8 592996

1 Primate_Chimp Altai_Neanderthal.DG Yoruba Saharawi.DG            0.00106  0.000219    4.85    1.22e- 6 585088
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Palestinian            0.00137  0.000171    7.99    1.31e-15 592996
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Lebanon_MBA.SG        0.00127  0.000234    5.42    6.00e- 8 576979
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_MLBA            0.00116  0.000206    5.62    1.92e- 8 551487
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_PPNB            0.00125  0.000351    3.56    3.72e- 4 397782
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Morocco_Iberomaurusian 0.000884  0.000238    3.72    1.97e- 4 576053
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Hadza1                0.0000783 0.000170    0.460  6.45e- 1 592952
8 Primate_Chimp Altai_Neanderthal.DG Yoruba Mbuti                  0.000251  0.000139    1.80    7.17e- 2 592996
9 Primate_Chimp Altai_Neanderthal.DG Yoruba Mende                  0.0000743 0.0000849  0.875  3.81e- 1 592996
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Ethiopia_4500BP.SG    0.0000932 0.000244    0.382  7.03e- 1 592974
11 Primate_Chimp Altai_Neanderthal.DG Yoruba Yoruba                0        0        NaN    NaN        592996
12 Primate_Chimp Altai_Neanderthal.DG Yoruba JPT.SG                0.00159  0.000223    7.13    1.04e-12 592995

Dataset used for these f4 stats: v52.2_HO_public
As these results are not very friendly in table, I will show them as diagram.

[Horatio McCallister]

Show Content

Spoiler
  pop1          pop2                pop3  pop4                        est      se      z        p      n
  <chr>        <chr>                <chr>  <chr>                      <dbl>    <dbl>  <dbl>    <dbl>  <dbl>
1 Primate_Chimp Altai_Neanderthal.DG Yoruba Primate_Chimp          -0.0959  0.000530 -181.  0        592996
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Altai_Neanderthal.DG    0.0798  0.000570  140.  0        592996
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Russia_Ust_Ishim.DG      0.00168 0.000352    4.76 1.96e- 6 592969
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Russia_MA1_HG.SG        0.00181 0.000358    5.05 4.41e- 7 425966
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Luxembourg_Loschbour.DG  0.00156 0.000290    5.39 6.87e- 8 592749
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Bulgaria_N              0.00146 0.000268    5.44 5.20e- 8 508318
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Iran_GanjDareh_N        0.00117 0.000231    5.05
 4.39e- 7 562236
8 Primate_Chimp Altai_Neanderthal.DG Yoruba Iran_HajjiFiruz_N        0.00139 0.000247    5.62 1.96e- 8 539981
9 Primate_Chimp Altai_Neanderthal.DG Yoruba Iranian                  0.00133 0.000177    7.52 5.30e-14 592996
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Mozabite                0.00125 0.000159    7.88 3.18e-15 592996

1 Primate_Chimp Altai_Neanderthal.DG Yoruba Georgia_Kotias.SG      0.00162  0.000298  5.43 5.70e- 8 591680
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Georgia_Satsurblia.SG  0.00113  0.000365  3.09 1.98e- 3 352936
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Ju_hoan_North          0.00116  0.000170  6.82 9.27e-12 592996
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Egyptian              0.00111  0.000160  6.95 3.71e-12 592996
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_Natufian        0.000790 0.000522  1.51 1.30e- 1 115788
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_Natufian_contam 0.00138  0.000403  3.43 5.94e- 4 249108
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Algerian              0.00106  0.000167  6.36 1.97e-10 592996
8 Primate_Chimp Altai_Neanderthal.DG Yoruba South_Africa_400BP.SG  0.000189 0.000159  1.19 2.36e- 1 592957
9 Primate_Chimp Altai_Neanderthal.DG Yoruba South_Africa_1900BP.SG 0.000863 0.000249  3.47 5.15e- 4 592913
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Saharawi              0.000907 0.000164  5.52 3.46e- 8 592996

1 Primate_Chimp Altai_Neanderthal.DG Yoruba Saharawi.DG            0.00106  0.000219    4.85    1.22e- 6 585088
2 Primate_Chimp Altai_Neanderthal.DG Yoruba Palestinian            0.00137  0.000171    7.99    1.31e-15 592996
3 Primate_Chimp Altai_Neanderthal.DG Yoruba Lebanon_MBA.SG        0.00127  0.000234    5.42    6.00e- 8 576979
4 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_MLBA            0.00116  0.000206    5.62    1.92e- 8 551487
5 Primate_Chimp Altai_Neanderthal.DG Yoruba Israel_PPNB            0.00125  0.000351    3.56    3.72e- 4 397782
6 Primate_Chimp Altai_Neanderthal.DG Yoruba Morocco_Iberomaurusian 0.000884  0.000238    3.72    1.97e- 4 576053
7 Primate_Chimp Altai_Neanderthal.DG Yoruba Hadza1                0.0000783 0.000170    0.460  6.45e- 1 592952
8 Primate_Chimp Altai_Neanderthal.DG Yoruba Mbuti                  0.000251  0.000139    1.80    7.17e- 2 592996
9 Primate_Chimp Altai_Neanderthal.DG Yoruba Mende                  0.0000743 0.0000849  0.875  3.81e- 1 592996
10 Primate_Chimp Altai_Neanderthal.DG Yoruba Ethiopia_4500BP.SG    0.0000932 0.000244    0.382  7.03e- 1 592974
11 Primate_Chimp Altai_Neanderthal.DG Yoruba Yoruba                0        0        NaN    NaN        592996
12 Primate_Chimp Altai_Neanderthal.DG Yoruba JPT.SG                0.00159  0.000223    7.13    1.04e-12 592995

Dataset used for these f4 stats: v52.2_HO_public
As these results are not very friendly in table, I will show them as diagram.

Thank you very much for running these. Not to be too much of a burden, but are you able to re-run these using the 1240k panel samples? I am curious if the greater SNP count in 1240k might change some of these results.

[TanTin]

Thank you very much for running these. Not to be too much of a burden, but are you able to re-run these using the 1240k panel samples? I am curious if the greater SNP count in 1240k might change some of these results.

The results with V52 1240k  are similar, but not better.

[TanTin]

Few notes:
1. The names of the groups don't match exactly to each other in the both datasets, so I had to replace some groups with other.
2. Despite the higher number of snips, for some populations in these F-stats we see low number of snips.
3. F-stats are calculated as percentage
number of matching snips / total number of snips

So by increasing the total number we may see decrease of the result for some groups, but it may also increase for other , depending on the range of the snips. If we remove the noise, the results should be the same for both datasets.

[TanTin]

And there is something you need to see:

R2V_52 <- f4( prefix,  "Chimp.REF" ,  "JPT.SG" ,  "Yoruba.DG"  ,  TEST)  #

Show Content

Spoiler

> R2V_52[ 1:20,]
# A tibble: 20 x 9
  pop1      pop2  pop3      pop4                              est      se      z        p      n
  <chr>    <chr>  <chr>    <chr>                            <dbl>    <dbl>  <dbl>    <dbl>  <dbl>
1 Chimp.REF JPT.SG Yoruba.DG Chimp.REF                    -0.111  0.00312  -35.5  7.31e-276 934295
2 Chimp.REF JPT.SG Yoruba.DG Altai_Neanderthal.DG          -0.0453  0.00135  -33.4  7.75e-245 934019
3 Chimp.REF JPT.SG Yoruba.DG Russia_Ust_Ishim.DG            0.0239  0.000871  27.4  7.46e-166 934244
4 Chimp.REF JPT.SG Yoruba.DG Russia_MA1_HG.SG              0.0313  0.000991  31.6  6.26e-219 545541
5 Chimp.REF JPT.SG Yoruba.DG Luxembourg_Loschbour.DG        0.0328  0.000942  34.8  5.63e-266 933966
6 Chimp.REF JPT.SG Yoruba.DG Bulgaria_N                    0.0318  0.00104  30.6  6.43e-206 807228
7 Chimp.REF JPT.SG Yoruba.DG Iran_GanjDareh_N              0.0269  0.000704  38.2  0        880886
8 Chimp.REF JPT.SG Yoruba.DG Iran_HajjiFiruz_N              0.0332  0.00111  30.0  4.72e-198 850756
9 Chimp.REF JPT.SG Yoruba.DG Iranian.DG                    0.0250  0.000811  30.8  5.89e-209 934269
10 Chimp.REF JPT.SG Yoruba.DG Mozabite.DG                    0.0219  0.000626  35.0  5.74e-268 934232
11 Chimp.REF JPT.SG Yoruba.DG Georgia_Kotias.SG              0.0244  0.000825  29.6  4.56e-192 932276
12 Chimp.REF JPT.SG Yoruba.DG Georgia_Satsurblia.SG          0.0233  0.000920  25.3  1.56e-141 579577
13 Chimp.REF JPT.SG Yoruba.DG Ju_hoan_North.DG              -0.0107  0.000412 -25.9  1.98e-148 934268
14 Chimp.REF JPT.SG Yoruba.DG Egypt_ThirdIntermediatePeriod  0.0460  0.00204  22.5  1.44e-112 514527
15 Chimp.REF JPT.SG Yoruba.DG Israel_Natufian                0.0162  0.000902  18.0  3.76e- 72 235018
16 Chimp.REF JPT.SG Yoruba.DG Israel_Natufian_contam        0.0209  0.000870  24.0  1.72e-127 367172
17 Chimp.REF JPT.SG Yoruba.DG Morocco_EN.SG                  0.0153  0.000702  21.9  6.59e-106 512200
18 Chimp.REF JPT.SG Yoruba.DG Morocco_LN.SG                  0.0325  0.000768  42.4  0        166429
19 Chimp.REF JPT.SG Yoruba.DG Morocco_unknown.SG            0.00930 0.00102    9.13 6.85e- 20 111811
20 Chimp.REF JPT.SG Yoruba.DG South_Africa_400BP.SG        -0.00149 0.000218  -6.83 8.46e- 12 934292
> R2V_52[ 20:33,]
# A tibble: 14 x 9
  pop1      pop2  pop3      pop4                        est      se      z          p      n
  <chr>    <chr>  <chr>    <chr>                      <dbl>    <dbl>  <dbl>      <dbl>  <dbl>
1 Chimp.REF JPT.SG Yoruba.DG South_Africa_400BP.SG  -0.00149  0.000218  -6.83  8.46e- 12 934292
2 Chimp.REF JPT.SG Yoruba.DG South_Africa_1900BP.SG -0.0133  0.000488 -27.2    6.39e-163 934183
3 Chimp.REF JPT.SG Yoruba.DG Saharawi.DG            0.0224  0.000598  37.4    1.18e-305 934262
4 Chimp.REF JPT.SG Yoruba.DG Palestinian.DG          0.0229  0.000755  30.3    2.57e-202 934274
5 Chimp.REF JPT.SG Yoruba.DG Lebanon_MBA.SG          0.0232  0.000768  30.2    2.74e-200 913609
6 Chimp.REF JPT.SG Yoruba.DG Israel_MLBA            0.0266  0.000695  38.3    0        865224
7 Chimp.REF JPT.SG Yoruba.DG Israel_PPNB            0.0211  0.000835  25.3    3.12e-141 614209
8 Chimp.REF JPT.SG Yoruba.DG Morocco_Iberomaurusian  0.0276  0.00121  22.8    1.71e-115 893474
9 Chimp.REF JPT.SG Yoruba.DG Turkey_N.SG            0.0241  0.000794  30.3    3.61e-202 934263
10 Chimp.REF JPT.SG Yoruba.DG Mbuti.DG              -0.00756  0.000340 -22.2    1.69e-109 934283
11 Chimp.REF JPT.SG Yoruba.DG Mende.DG              -0.000798 0.000234  -3.41  6.51e-  4 934238
12 Chimp.REF JPT.SG Yoruba.DG Ethiopia_4500BP.SG      0.00178  0.000306  5.83  5.47e-  9 934244
13 Chimp.REF JPT.SG Yoruba.DG Yoruba.DG              0        0        NaN    NaN        934295
14 Chimp.REF JPT.SG Yoruba.DG JPT.SG                  0.0428  0.00113  37.7    0        934295

Chimp.REF JPT.SG Yoruba.DG Egypt_ThirdIntermediatePeriod  0.0460  0.00204  22.5  1.44e-112 514527
14 Chimp.REF JPT.SG Yoruba.DG JPT.SG                  0.0428  0.00113  37.7    0        934295

There is no mistake. This is a result by calculating  514527  snips. This  Egypt_ThirdIntermediatePeriod  Egyptians are near the same as the people from Japan.
When we compare them to the Chimp.  However South Africans show more affinity to the Chimp side. 

[Megalophias]

In the original Iberomaurusian paper they had stats of f4(Chimp, African; Natufian, Taforalt) or similar, where even Ju|'hoan had significant positive results, showing the affinity between ANA and Sub-Saharan Africans. I was looking for other stats along those lines, but I couldn't find any. Does anyone have any? Preferably not with the old dodgy Natufian samples, and ideally with ancient South Africans or Malawians, since Ju|'hoan do have recent East African gene flow.

PS or f4(West Eurasian, North African; Malawi, South Africa)?

[TanTin]

In the original Iberomaurusian paper they had stats of f4(Chimp, African; Natufian, Taforalt) or similar, where even Ju|'hoan had significant positive results, showing the affinity between ANA and Sub-Saharan Africans. I was looking for other stats along those lines, but I couldn't find any. Does anyone have any? Preferably not with the old dodgy Natufian samples, and ideally with ancient South Africans or Malawians, since Ju|'hoan do have recent East African gene flow.

PS or f4(West Eurasian, North African; Malawi, South Africa)?

I am not sure what F4 stats exactly do you need. Here is one such example.
Tell me more precise what do you need so I will generate it. 

[Megalophias]

Thanks TanTin, that is helpful. Does seem like ANA has a bit of West African affinity, but doesn't tell us what causes this.

f4(Chimp, South_Africa_1900BP; Israel_Natufian, Morocco_Iberomaurusian) would be good, to replicate the findings in the previous paper.

Do you have the ancient Malawi foragers (Hora and Fingira)? If you do then f4(Malawi, South_Africa_1900BP; Israel_Natufian, Morocco_Iberomaurusian) and f4(Chimp, Malawi; Israel_Natufian, Morocco_Iberomaurusian) would be nice.