Created G25 Calculator For South/Central Asians.

[geneticsikiru]

table (1).png
barchart (1).png

[geneticsikiru]

table (2).png
barchart (2).png

[Kapisa]

Any idea what could add instead, I needed to another proxy otherwise everyone is scoring very high IVC.

Bronze Age Central Asian & IAMC individuals are by far the best proxy for any excess west eurasian on top of what comes with IVC, I would rather not include TUR_Kaman-Kalehoyuk_MLBA either, rather BMAC or KGZ-BA

Sample	Fit	Shahr I Sokhta BA2	Bustan BA	CG CentralSteppeMLBA	Chokhopani 2700BP	S AASI Sim	MNG Mongol
Average (Barikot H)	1.53	53	21.5	21	4	0.5	0
Average (Udegram IA)	1.55	52	26	18	2.5	0	1.5

[geneticsikiru]

Bronze Age Central Asian & IAMC individuals are by far the best proxy for any excess west eurasian on top of what comes with IVC, I would rather not include TUR_Kaman-Kalehoyuk_MLBA either, rather BMAC or KGZ-BA

Sample	Fit	Shahr I Sokhta BA2	Bustan BA	CG CentralSteppeMLBA	Chokhopani 2700BP	S AASI Sim	MNG Mongol
Average (Barikot H)	1.53	53	21.5	21	4	0.5	0
Average (Udegram IA)	1.55	52	26	18	2.5	0	1.5

Thanks for the tip, for some pops CHN is ok, but for others they score too much of it.

[CopperAxe]

Bronze Age Central Asian & IAMC individuals are by far the best proxy for any excess west eurasian on top of what comes with IVC, I would rather not include TUR_Kaman-Kalehoyuk_MLBA either, rather BMAC or KGZ-BA

Sample	Fit	Shahr I Sokhta BA2	Bustan BA	CG CentralSteppeMLBA	Chokhopani 2700BP	S AASI Sim	MNG Mongol
Average (Barikot H)	1.53	53	21.5	21	4	0.5	0
Average (Udegram IA)	1.55	52	26	18	2.5	0	1.5

Something with Central Asian forager ancestry definitely is required for South Asians as it is pretty clear that Indo-Aryans would have in it excess relative to what was already present in the genepool of urban agriculturalists of sc-asia, however the amount in most cases would directly correlate with steppe ancestry so if that isn't high then it won't be much of a requirement in models. The steppe ancestry in models would drop down a bit but it would be more accurate anyways. What is to be noted though is that the "WSHG" ancestry we see around 4000 BC and onwards is more like a Kumsay type profile rather than just Tyumen/Sosonivoy types, let alone Tarim_EMBA which people really overuse a lot. The Kyzlbulak_mlba2 sample is pretty interesting in that regard as it is roughly 50% Airgyrzhal-like. You also have an iron age sample from Liushui that models as 50/50 Khovsgol_BA/Aigyrzhal_BA but with next to no steppe_mlba, so these types were definitely still around in some areas of Central Asia during the LBA.

[Kapisa]

Something with Central Asian forager ancestry definitely is required for South Asians as it is pretty clear that Indo-Aryans would have in it excess relative to what was already present in the genepool of urban agriculturalists of sc-asia, however the amount in most cases would directly correlate with steppe ancestry so if that isn't high then it won't be much of a requirement in models. The steppe ancestry in models would drop down a bit but it would be more accurate anyways. What is to be noted though is that the "WSHG" ancestry we see around 4000 BC and onwards is more like a Kumsay type profile rather than just Tyumen/Sosonivoy types, let alone Tarim_EMBA which people really overuse a lot. The Kyzlbulak_mlba2 sample is pretty interesting in that regard as it is roughly 50% Airgyrzhal-like. You also have an iron age sample from Liushui that models as 50/50 Khovsgol_BA/Aigyrzhal_BA but with next to no steppe_mlba, so these types were definitely still around in some areas of Central Asia during the LBA.

Yup. And these kumsay like individuals are also present in Sintashta sites as outliers.

Sample	Fit	Kumsay EBA	Aigyrzhal BA	Yamnaya RUS Samara	Globular Amphora
Average (Sintashta MLBA o1)	4.47	100	0	0	0
Average (Sintashta MLBA o3)	6.56	96	0	0	4
Average (Sintashta MLBA o2)	2.19	35.5	0	50	14.5
Average (Sintashta MLBA)	2.10	0	0	68.5	31.5

[CopperAxe]

Yup. And these kumsay like individuals are also present in Sintashta sites as outliers.

Sample	Fit	Kumsay EBA	Aigyrzhal BA	Yamnaya RUS Samara	Globular Amphora
Average (Sintashta MLBA o1)	4.47	100	0	0	0
Average (Sintashta MLBA o3)	6.56	96	0	0	4
Average (Sintashta MLBA o2)	2.19	35.5	0	50	14.5
Average (Sintashta MLBA)	2.10	0	0	68.5	31.5

I think some of these others need actual WSHG proper and in one case EHG even. I think those outliers were also buried in unmarked graves outside the citadels rather than the regular graves so who knows what their context their was exactly. What matters the most is that the Kazakh steppes and especially the western parts were inhabited by these peoples, because you have the Syr Darya and Amu Darya which emptied in the (former) aral sea from the central asian mountain passes so you can imagine how they spread in those regions.

[Kapisa]

Thanks for the tip, for some pops CHN is ok, but for others they score too much of it.

Depends on the question one is asking. If your model is trying to find ancestral populations for modern south Asians then best approach is to be consistent in terms of time period. For example, if we want to test the hypothesis that Iron Age populations of South Asia descend from an interaction of Bronze Age populations of South and Central Asia then we can use populations from that time period for that model. Using historic era individuals which are already admixed with (IVC+BMAC+steppe) would result in skewing results as it would not be obvious how much actual "steppe" or "IVC" one carries. Some historic profiles are very close to modern day population, others are not that prevalent anymore today, like I6893 (Saidu Sharif H) and (Junmachanyilian HE2) which are possibly admixed with something additional; Greek and later Persian era admixture!!??

Sample	Fit	Shahr I Sokhta BA2	Bustan BA	CG CentralSteppeMLBA	Chokhopani 2700BP	Mycenaean Palace of Nestor BA
Average (Butkara H)	1.58	73.5	11.5	11	4	0
Average (Aligrama H)	1.59	68.5	8.5	18.5	4.5	0
Average (Saidu Sharif H)	1.08	64.5	10	20	4.5	1
Average (Gujar Pakistan)	1.21	64.5	13.5	19.5	2	0.5
Median (Sindhi)	1.50	63.5	14	16.5	3	3
I6893 (Saidu Sharif H)	3.19	60	0	26	4	10
KhatriNonOutlier Average (auro)	1.07	57	16.5	22	3	1.5
Average (Kohistani)	1.17	55	17.5	22	5.5	0
Average (Kamboj)	1.06	53.5	15.5	22	3	6
Average (Barikot H)	1.53	53.5	21	21	4.5	0
Average (Junmachanyilian HE2)	2.67	45	32.5	13.5	0	9
KNT005 (Otyrar Antiquity)	2.00	19	49	26	0.5	5.5

[T90]

Hi Guys,

I made some changes to the model, let me know what you think.

Target: 1


Distance: 2.6469% / 0.02646926
45.2 IRN_Shahr_I_Sokhta_BA2
18.8 CHN_Junmachanyilian_HE2
18.0 KAZ_Mys_MLBA
11.2 S_AASI_Sim
2.8 NPL_Samdzong_1500BP
2.6 RUS_Sintashta_MLBA
1.2 MNG_Mongol
0.2 TUR_Kaman-Kalehoyuk_MLBA



Target: 2


Distance: 2.7095% / 0.02709453
48.0 IRN_Shahr_I_Sokhta_BA2
22.4 CHN_Junmachanyilian_HE2
17.0 KAZ_Mys_MLBA
6.6 S_AASI_Sim
6.0 NPL_Samdzong_1500BP

[geneticsikiru]

Made pca of results:

new pca plot.png