PDA

View Full Version : Pashtun and Afghan Global25 Results



Coldmountains
05-19-2020, 11:28 PM
I collected some Pashtun Global25 coordinates and created some averages for Pashtuns from Kandahar, Kabul/Paghman and North Afghanistan (only based on academic samples)




Pashtun_AFG_AVG,0.08723,0.03295834,-0.06955018,0.03952345,-0.05904121,0.02909764,0.00391666,0.00097882,-0.02268352,-0.01853845,-0.00507825,0.00061773,0.00071191,-0.00693966,0.01425057,0.01454466,-0.00323191,0.00045666,0.00213682,-0.01301003,-0.00270348,-0.00637,0.00419415,-0.00490012,0.00404218
Pashtun_Kabul_Paghman,0.08840267,0.02166433,-0.06901333,0.042636,-0.05601033,0.03151467,0.00219333,-0.00238467,-0.02358833,-0.01372833,-0.00481733,0.00069933,-0.00153633,-0.01215667,0.012441,0.02081633,0.00343367,-0.001816,-0.00020967,-0.01271433,-0.00311967,-0.004122,0.00579267,-0.006346,0.00247467
Pashtun_Kandahar,0.0878716,0.04285514,-0.0716276,0.0375326,-0.06286306,0.028484,0.00382266,0.0024766,-0.02020694,-0.0215404,-0.00470914,0.0009394,0.001358,-0.00428486,0.01628646,0.01279926,-0.0060672,0.00113146,0.0035948,-0.01514066,-0.00375146,-0.0073118,0.00804394,-0.00605686,0.0032408
Pashtun_North_Afghanistan,0.084988,0.02775767,-0.06662467,0.039729,-0.05570233,0.02770333,0.00579667,0.001846,-0.02590633,-0.01834533,-0.00595433,0.000000,0.00188333,-0.00614733,0.012667,0.011182,-0.005172,0.00160467,0.00205333,-0.00975467,-0.00054067,-0.00704833,-0.00382067,-0.00152633,0.00694533

Also i modelled these Pashtuns and various Pashtun groups, which we already had in the offical global25 spreadsheet. The amount of Steppe and BMAC-like ancestry is similar in all Pashtun subgroups here but there seems to be significant differences in terms of ASI-like and East Asian-like ancestry. Generally ASI- or IVC-like ancestry is peaking among most eastern Pashtuns and East Asian-like ancestry among Pashtuns in North Afghanistan. There is also some minor Iran_C-like shift among Kandahari Pashtuns. WSHG on the otherhand peaks here among eastern Pashtuns, which also tend to be the most IVC/ASI-shifted Pashtuns.

https://i.ibb.co/qDc836v/ADASD.jpg (https://ibb.co/S6Mg4Sz)

Unfornuately i have not found Afghan Pashtun global25 coordinates east of Kabul but Pashtuns from regions like Nangarhar are probably inbetween Kabuli Pashtuns and Tarkalani/Yusufzai Pashtuns here. PAK_Udegram_MA_Ghaznavid is on this pca also on this position and likely from East Afghanistan.

https://i.ibb.co/PwbnYXr/newplot-3.png (https://ibb.co/t8NthjH)


Generally it is assumed that Pashtuns from Kandahar and southeastern Afghanistan are most representative for Proto-Pashtuns and have the least Pre-Pashtun admixture. So i tried to model Pashtuns as mix of Kandahari Pashtuns + Non-Pashtun groups (Pamiri-like East Iranics, Dardics, Burusho, Uzbeks, Turkmens,..). Based on these model Proto-Pashtun-like ancestry is around 40-50% among eastern Pashtuns in Pakistan and around 50-60% among Pashtuns in northern Afghanistan,but this is probably quite different from individual to individual and not fully accurate.

https://i.ibb.co/JQJYcGy/Unbenannt.jpg (https://ibb.co/k1CLGw9)

pegasus
05-20-2020, 12:44 AM
I collected some Pashtun Global25 coordinates and created some averages for Pashtuns from Kandahar, Kabul/Paghman and North Afghanistan (only based on academic samples)





Also I modeled these Pashtuns and various Pashtun groups, which we already had in the offical global25 spreadsheet. The amount of Steppe and BMAC-like ancestry is similar in all Pashtun subgroups here but there seems to be significant differences in terms of ASI-like and East Asian-like ancestry. Generally ASI- or IVC-like ancestry is peaking among most eastern Pashtuns and East Asian-like ancestry among Pashtuns in North Afghanistan. There is also some minor Iran_C-like shift among Kandahari Pashtuns. WSHG on the otherhand peaks here among eastern Pashtuns, which also tend to be the most IVC/ASI-shifted Pashtuns.

https://i.ibb.co/qDc836v/ADASD.jpg (https://ibb.co/S6Mg4Sz)

Unfornuately i have not found Afghan Pashtun global25 coordinates east of Kabul but Pashtuns from regions like Nangarhar are probably inbetween Kabuli Pashtuns and Tarkalani/Yusufzai Pashtuns here. PAK_Udegram_MA_Ghaznavid is on this pca also on this position and likely from East Afghanistan.

https://i.ibb.co/PwbnYXr/newplot-3.png (https://ibb.co/t8NthjH)


Generally it is assumed that Pashtuns from Kandahar and southeastern Afghanistan are most representative for Proto-Pashtuns and have the least Pre-Pashtun admixture. So i tried to model Pashtuns as mix of Kandahari Pashtuns + Non-Pashtun groups (Pamiri-like East Iranics, Dardics, Burusho, Uzbeks, Turkmens,..). Based on these model Proto-Pashtun-like ancestry is around 40-50% among eastern Pashtuns in Pakistan and around 50-60% among Pashtuns in northern Afghanistan,but this is probably quite different from individual to individual and not fully accurate.

https://i.ibb.co/JQJYcGy/Unbenannt.jpg (https://ibb.co/k1CLGw9)

Those IVC scores are ridiculously depressed( looks like your using a Paniya proxy even) they would be plotting west of where Pamiris do but they are significantly shifted southwards for a reason. Their ethnogenesis is much later than Indo Aryan ones, so its likely their Proto Pashtun substrate mixed with some population on the SPGT cline or a related population. This is in line with the Iran paper which had some Sistanis cluster with the Western shifted SPGT samples like Barikot, so that really limits what their ASI substrate population would be. Given Steppe ancestry is present by 1600 BC, its unlikely Proto Pashtuns mix with actual IVC people but groups that formed post it, but regardless IVC substrates are needed. So they are unique in that aspect they harbor both Indo Aryan and East Iranic ancestry. Considering the greatest linguistic diversity among Pashto is the environs of Northern Balochistan, unfortunately, it is impossible their ASI substrate was from some tribal or HG population. DMXX and I both came to the conclusion there is a separate sub cline going from SPGT to some Steppe rich progenitor-like Kangju or intermediate between TKM IA and Kangju . ALL Pashtuns fall on it, even most Pamiris do. Yaghnobis did not because they lack any IVC ancestry or have very low levels of it. Within Pashtuns, there is significant intra diversity. Your model with modern populations is also very overfitted since Iskashimis fall on the same cline with other Pashtuns and they are related to each other. Some of the Tarkalani and Uthmankhels are contemporaneous with Kandahar Pashtuns, they need richer Steppe progenitors.


"sample": "Uthmankhel:Average",
"fit": 1.0052,
"PAK_Katelai_IA": 64.17,
"KAZ_Kangju": 35.83,

"sample": "Tarkalani:Average",
"fit": 1.0076,
"PAK_Katelai_IA": 67.5,
"KAZ_Kangju": 32.5,


"sample": "Tajik_Ishkashim:Average",
"fit": 0.9729,
"KAZ_Kangju": 53.33,
"PAK_Katelai_IA": 46.67,

sample": "Tajik_Shugnan:Average",
"fit": 0.947,
"KAZ_Kangju": 68.33,
"PAK_Katelai_IA": 31.67,


Those Afghan Pashtun samples have some extra Plateau Iranian ancestry they model better with it .

"sample": "Custom:Pashtun_AFG_AVG",
"fit": 0.97,
"PAK_Katelai_IA": 52.5,
"KAZ_Kangju": 32.5,
"IRN_Seh_Gabi_C": 15


Afghan Pashtuns who have acquired significant and recent Plateau Iranian ancestry or/and Turkic ancestry via Uzbeks or Hazaras can shift accordingly but most still fall in that classic BMAC-Central Steppe- IVC ancestral blend.

Target: Zhob_scaled
Distance: 1.9537%
38.8 TKM_Gonur1_BA
29.6 IRN_Shahr_I_Sokhta_11466
26.8 RUS_Krasnoyarsk_MLBA
4.8 NPL_Chokhopani_2700BP


Target: Zhob_scaled
Distance: 2.6322%
49.2 TKM_Gonur1_BA
22.6 IRN_Shahr_I_Sokhta_11459
22.2 RUS_Krasnoyarsk_MLBA
6.0 NPL_Chokhopani_2700BP

https://i.imgur.com/Xr3MtuN.png
There is this notion the less AASI there is , the less IVC but that's not always the case because of how cosmopolitan those IVC samples are.

"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.3243,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I11466": 30.83,
"UZB_Sappali_Tepe_BA": 30.83,
"CustomGroup_CentralSteppeMLBA": 30,
"MNG_Hovsgol_BA": 8.33,



"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.0579,
"CustomGroup_CentralSteppeMLBA": 31.67,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I11466": 29.17,
"UZB_Sappali_Tepe_BA": 24.17,
"IRN_Seh_Gabi_C": 7.5,
"Mongola": 7.5,


"sample": "Custom:Kandahar_scaled",
"fit": 1.5874,
"UZB_Sappali_Tepe_BA": 35,
"CustomGroup_CentralSteppeMLBA": 30.83,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I8726": 30.83,
"Mongola": 3.33,


"sample": "Custom:Kandahar_scaled",
"fit": 1.3646,
"UZB_Sappali_Tepe_BA": 39.17,
"CustomGroup_CentralSteppeMLBA": 30.83,
"CustomGroup_IVCp--TKM_Gonur2_BA_I2123": 26.67,
"Mongola": 3.33,



He is the only sample I have seen model well with 8726 and its inline with the archaeological notes. Interestingly he models well with Gonur2 as well.

• 201, Grave 201, LS1 (I8726): Context date of 3100-3000 BCE. Genetically male. This individual was one of two burials, deposited one above the other, in a catacomb grave dated to the early Period I, phases 10/9. The grave had a few pots and was otherwise poor except for the inclusion of an alabaster vessel. Salvatori et al. (175) note that a distinctive pottery type among the grave goods belongs to a distinctive cluster of graves at Shahr-i-Sokhta that is “possibly local or northeast oriented (Kandahar area)”

Phase I and II of Mundigak is essentially a Western extension of Zhob culture (IVC related), III and IV it transforms into part of the Jiroft culture and you see that because the pottery changes with Quetta Ware further west ie Chlorite vases.

Coldmountains
05-20-2020, 06:22 AM
Those IVC scores are ridiculously depressed( looks like your using a Paniya proxy even) they would be plotting west of where Pamiris do but they are significantly shifted southwards for a reason. Their ethnogenesis is much later than Indo Aryan ones, so its likely their Proto Pashtun substrate mixed with some population on the SPGT cline or a related population. This is in line with the Iran paper which had some Sistanis cluster with the Western shifted SPGT samples like Barikot, so that really limits what their ASI substrate population would be. Given Steppe ancestry is present by 1600 BC, its unlikely Proto Pashtuns mix with actual IVC people but groups that formed post it, but regardless IVC substrates are needed. So they are unique in that aspect they harbor both Indo Aryan and East Iranic ancestry. Considering the greatest linguistic diversity among Pashto is the environs of Northern Balochistan, unfortunately, it is impossible their ASI substrate was from some tribal or HG population. DMXX and I both came to the conclusion there is a separate sub cline going from SPGT to some Steppe rich progenitor-like Kangju or intermediate between TKM IA and Kangju . ALL Pashtuns fall on it, even most Pamiris do. Yaghnobis did not because they lack any IVC ancestry or have very low levels of it. Within Pashtuns, there is significant intra diversity. Your model with modern populations is also very overfitted since Iskashimis fall on the same cline with other Pashtuns and they are related to each other. Some of the Tarkalani and Uthmankhels are contemporaneous with Kandahar Pashtuns, they need richer Steppe progenitors.


"sample": "Uthmankhel:Average",
"fit": 1.0052,
"PAK_Katelai_IA": 64.17,
"KAZ_Kangju": 35.83,

"sample": "Tarkalani:Average",
"fit": 1.0076,
"PAK_Katelai_IA": 67.5,
"KAZ_Kangju": 32.5,


"sample": "Tajik_Ishkashim:Average",
"fit": 0.9729,
"KAZ_Kangju": 53.33,
"PAK_Katelai_IA": 46.67,

sample": "Tajik_Shugnan:Average",
"fit": 0.947,
"KAZ_Kangju": 68.33,
"PAK_Katelai_IA": 31.67,


Those Afghan Pashtun samples have some extra Plateau Iranian ancestry they model better with it .

"sample": "Custom:Pashtun_AFG_AVG",
"fit": 0.97,
"PAK_Katelai_IA": 52.5,
"KAZ_Kangju": 32.5,
"IRN_Seh_Gabi_C": 15


Afghan Pashtuns who have acquired significant and recent Plateau Iranian ancestry or/and Turkic ancestry via Uzbeks or Hazaras can shift accordingly but most still fall in that classic BMAC-Central Steppe- IVC ancestral blend.

Target: Zhob_scaled
Distance: 1.9537%
38.8 TKM_Gonur1_BA
29.6 IRN_Shahr_I_Sokhta_11466
26.8 RUS_Krasnoyarsk_MLBA
4.8 NPL_Chokhopani_2700BP


Target: Zhob_scaled
Distance: 2.6322%
49.2 TKM_Gonur1_BA
22.6 IRN_Shahr_I_Sokhta_11459
22.2 RUS_Krasnoyarsk_MLBA
6.0 NPL_Chokhopani_2700BP

https://i.imgur.com/Xr3MtuN.png
There is this notion the less AASI there is , the less IVC but that's not always the case because of how cosmopolitan those IVC samples are.

"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.3243,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I11466": 30.83,
"UZB_Sappali_Tepe_BA": 30.83,
"CustomGroup_CentralSteppeMLBA": 30,
"MNG_Hovsgol_BA": 8.33,

"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.2814,
"UZB_Sappali_Tepe_BA": 36.67,
"CustomGroup_CentralSteppeMLBA": 28.33,
"IRN_Shahr_I_Sokhta_BA2": 26.67,

"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.0579,
"CustomGroup_CentralSteppeMLBA": 31.67,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I11466": 29.17,
"UZB_Sappali_Tepe_BA": 24.17,
"IRN_Seh_Gabi_C": 7.5,
"Mongola": 7.5,


"sample": "Custom:Kandahar_scaled",
"fit": 1.5874,
"UZB_Sappali_Tepe_BA": 35,
"CustomGroup_CentralSteppeMLBA": 30.83,
"CustomGroup_IVCp--IRN_Shahr_I_Sokhta_BA2_I8726": 30.83,
"Mongola": 3.33,


"sample": "Custom:Kandahar_scaled",
"fit": 1.3646,
"UZB_Sappali_Tepe_BA": 39.17,
"CustomGroup_CentralSteppeMLBA": 30.83,
"CustomGroup_IVCp--TKM_Gonur2_BA_I2123": 26.67,
"Mongola": 3.33,



He is the only sample I have seen model well with 8726 and its inline with the archaeological notes. Interestingly he models well with Gonur2 as well.

• 201, Grave 201, LS1 (I8726): Context date of 3100-3000 BCE. Genetically male. This individual was one of two burials, deposited one above the other, in a catacomb grave dated to the early Period I, phases 10/9. The grave had a few pots and was otherwise poor except for the inclusion of an alabaster vessel. Salvatori et al. (175) note that a distinctive pottery type among the grave goods belongs to a distinctive cluster of graves at Shahr-i-Sokhta that is “possibly local or northeast oriented (Kandahar area)”

Phase I and II of Mundigak is essentially a Western extension of Zhob culture (IVC related), III and IV it transforms into part of the Jiroft culture and you see that because the pottery changes with Quetta Ware further west ie Chlorite vases.

I not claimed that they are 10% IVC i used this to model ASI-like ancestry. But the higher they scored ASI-like ancestry here the higher they also scored IVC-like ancestry when using the most ASI-shifted Iran-Shar-i-Sokhta sample it is not from individual to individual exactly like that but IVC is much higher among eastern Pashtuns than among Pashtuns in Kabul or Kandahar. Also Proto-Pashtuns are from southeastern Afghanistan and formed there so Kangyu or Iron Age groups from Tajikistan or Uzbekistan are not directly relevant to model Proto-Pashtun-like ancestry because it would be very unlikely that such kind of population existed in southeastern Afghanistan or migrated to Swat to spread Pashtun language/culture in the first place. Many of the Pashtun migrations just happened recently (last 500 years).

What happened before Proto-Pashtuns in southeastern Afghanistan t is not fully clear because we lack the right reference pops but we actually found Pashtun-related R1a-YP413 in the Tazabagyab culture which or another closely related culture could actually brought directly Iranic ancestry into southeastern Afghanistan. YP413 was not found in any Saka-related group in Central Asia later and also has a rather South Central Asian/South Asian distribution. Pre-Proto-Pashtuns had a very old presence in southern Afghanistan and probably are related to the Pargyetae of Arachosia,who were mentioned by Ptolemy. Pashtuns sit on this cline, you showed, in my opinion because they have relatively high steppe ancestry + some WSHG/East Asian which maybe was much lower among Proto-Pashtuns in southeastern Afghanistan. But this kind of mix could make their Iranic-like ancestry looking superfically similar to Kangju. So i am not sure their Iranic ancestry can be modelled with Kangju.


https://i.ibb.co/CBqppq6/Unbenannt.jpg (https://ibb.co/tBSwwSz)


Also many of the Afghan Pashtuns here are indeed clustering more west and closer to Ishashmi than to Kamboj or Kho for example.

Distance to: Pashtun_Kandahar
0.03415025 Uthmankhel
0.03415025 Uthmankhel
0.03465725 Tarkalani
0.03465725 Tarkalani
0.04206959 Tajik_Ishkashim
0.04206959 Tajik_Ishkashim
0.04211571 Yusufzai
0.04211571 Yusufzai
0.04310913 Kalash
0.04310913 Kalash
0.04572289 Pashtun
0.04572289 Pashtun
0.04772078 Balochi
0.04772078 Balochi
0.04877939 Kho_Singanali
0.04877939 Kho_Singanali
0.04915042 Brahui
0.04915042 Brahui
0.05010114 Tajik_Shugnan
0.05010114 Tajik_Shugnan
0.05622411 Tajik_Rushan
0.05622411 Tajik_Rushan
0.05640107 Parsi_Pakistan
0.05640107 Parsi_Pakistan
0.05679218 Parsi_India
0.05679218 Parsi_India
0.05691338 Kamboj
0.05691338 Kamboj
0.06124077 Tajik
0.06124077 Tajik

Distance to: Pashtun_North_Afghanistan
0.03014273 Uthmankhel
0.03014273 Uthmankhel
0.03119565 Tajik_Ishkashim
0.03119565 Tajik_Ishkashim
0.03139264 Tarkalani
0.03139264 Tarkalani
0.03821791 Yusufzai
0.03821791 Yusufzai
0.03833795 Pashtun
0.03833795 Pashtun
0.03915481 Kho_Singanali
0.03915481 Kho_Singanali
0.04178534 Kalash
0.04178534 Kalash
0.04441438 Tajik_Shugnan
0.04441438 Tajik_Shugnan
0.04899644 Tajik
0.04899644 Tajik
0.05286248 Tajik_Rushan
0.05286248 Tajik_Rushan
0.05321879 Kamboj
0.05321879 Kamboj
0.05698529 Jatt_Pathak
0.05698529 Jatt_Pathak
0.05730701 Ror
0.05730701 Ror
0.05880353 Balochi
0.05880353 Balochi
0.05911766 Brahui
0.05911766 Brahui

Distance to: Pashtun_Kabul_Paghman
0.02709563 Tarkalani
0.02709563 Tarkalani
0.02872563 Uthmankhel
0.02872563 Uthmankhel
0.03332909 Kho_Singanali
0.03332909 Kho_Singanali
0.03375375 Yusufzai
0.03375375 Yusufzai
0.03419835 Tajik_Ishkashim
0.03419835 Tajik_Ishkashim
0.03677858 Pashtun
0.03677858 Pashtun
0.04308332 Kalash
0.04308332 Kalash
0.04877653 Tajik_Shugnan
0.04877653 Tajik_Shugnan
0.04926301 Kamboj
0.04926301 Kamboj
0.05279816 Jatt_Pathak
0.05279816 Jatt_Pathak
0.05321686 Ror
0.05321686 Ror
0.05381213 Tajik
0.05381213 Tajik
0.05778941 Tajik_Rushan
0.05778941 Tajik_Rushan
0.06120619 Kohistani
0.06120619 Kohistani
0.06129390 Brahui
0.06129390 Brahui

pegasus
05-20-2020, 07:48 AM
I not claimed that they are 10% IVC i used this to model ASI-like ancestry. But the higher they scored ASI-like ancestry here the higher they also scored IVC-like ancestry when using the most ASI-shifted Iran-Shar-i-Sokhta sample it is not from individual to individual exactly like that but IVC is much higher among eastern Pashtuns than among Pashtuns in Kabul or Kandahar. Also Proto-Pashtuns are from southeastern Afghanistan and formed there so Kangyu or Iron Age groups from Tajikistan or Uzbekistan are not directly relevant to model Proto-Pashtun-like ancestry because it would be very unlikely that such kind of population existed in southeastern Afghanistan or migrated to Swat to spread Pashtun language/culture in the first place. Many of the Pashtun migrations just happened recently (last 500 years).

What happened before Proto-Pashtuns in southeastern Afghanistan t is not fully clear because we lack the right reference pops but we actually found Pashtun-related R1a-YP413 in the Tazabagyab culture which or another closely related culture could actually brought directly Iranic ancestry into southeastern Afghanistan. YP413 was not found in any Saka-related group in Central Asia later and also has a rather South Central Asian/South Asian distribution. Pre-Proto-Pashtuns had a very old presence in southern Afghanistan and probably are related to the Pargyetae of Arachosia,who were mentioned by Ptolemy. Pashtuns sit on this cline, you showed, in my opinion because they have relatively high steppe ancestry + some WSHG/East Asian which maybe was much lower among Proto-Pashtuns in southeastern Afghanistan. But this kind of mix could make their Iranic-like ancestry looking superfically similar to Kangju. So i am not sure their Iranic ancestry can be modelled with Kangju.


https://i.ibb.co/CBqppq6/Unbenannt.jpg (https://ibb.co/tBSwwSz)


Also many of the Afghan Pashtuns here are indeed clustering more west and closer to Ishashmi than to Kamboj or Kho for example.

Distance to: Pashtun_Kandahar
0.03415025 Uthmankhel
0.03415025 Uthmankhel
0.03465725 Tarkalani
0.03465725 Tarkalani
0.04206959 Tajik_Ishkashim
0.04206959 Tajik_Ishkashim
0.04211571 Yusufzai
0.04211571 Yusufzai
0.04310913 Kalash
0.04310913 Kalash
0.04572289 Pashtun
0.04572289 Pashtun
0.04772078 Balochi
0.04772078 Balochi
0.04877939 Kho_Singanali
0.04877939 Kho_Singanali
0.04915042 Brahui
0.04915042 Brahui
0.05010114 Tajik_Shugnan
0.05010114 Tajik_Shugnan
0.05622411 Tajik_Rushan
0.05622411 Tajik_Rushan
0.05640107 Parsi_Pakistan
0.05640107 Parsi_Pakistan
0.05679218 Parsi_India
0.05679218 Parsi_India
0.05691338 Kamboj
0.05691338 Kamboj
0.06124077 Tajik
0.06124077 Tajik

Distance to: Pashtun_North_Afghanistan
0.03014273 Uthmankhel
0.03014273 Uthmankhel
0.03119565 Tajik_Ishkashim
0.03119565 Tajik_Ishkashim
0.03139264 Tarkalani
0.03139264 Tarkalani
0.03821791 Yusufzai
0.03821791 Yusufzai
0.03833795 Pashtun
0.03833795 Pashtun
0.03915481 Kho_Singanali
0.03915481 Kho_Singanali
0.04178534 Kalash
0.04178534 Kalash
0.04441438 Tajik_Shugnan
0.04441438 Tajik_Shugnan
0.04899644 Tajik
0.04899644 Tajik
0.05286248 Tajik_Rushan
0.05286248 Tajik_Rushan
0.05321879 Kamboj
0.05321879 Kamboj
0.05698529 Jatt_Pathak
0.05698529 Jatt_Pathak
0.05730701 Ror
0.05730701 Ror
0.05880353 Balochi
0.05880353 Balochi
0.05911766 Brahui
0.05911766 Brahui

Distance to: Pashtun_Kabul_Paghman
0.02709563 Tarkalani
0.02709563 Tarkalani
0.02872563 Uthmankhel
0.02872563 Uthmankhel
0.03332909 Kho_Singanali
0.03332909 Kho_Singanali
0.03375375 Yusufzai
0.03375375 Yusufzai
0.03419835 Tajik_Ishkashim
0.03419835 Tajik_Ishkashim
0.03677858 Pashtun
0.03677858 Pashtun
0.04308332 Kalash
0.04308332 Kalash
0.04877653 Tajik_Shugnan
0.04877653 Tajik_Shugnan
0.04926301 Kamboj
0.04926301 Kamboj
0.05279816 Jatt_Pathak
0.05279816 Jatt_Pathak
0.05321686 Ror
0.05321686 Ror
0.05381213 Tajik
0.05381213 Tajik
0.05778941 Tajik_Rushan
0.05778941 Tajik_Rushan
0.06120619 Kohistani
0.06120619 Kohistani
0.06129390 Brahui
0.06129390 Brahui

The model suggests BMAC populations mixed with Steppe people and proceeded to mix with tribals or HGs . Those Kabul River Pashtuns ( Abasin /Spin Ghar) types are heavily SPGT like and u can see it from their AASI values, that Paghman one even with Turkic related ancestry is still harboring close to 10% AASI and it shows in his models. Southern Pashtuns are a different for sure, because they typically need a more Zagrosian shifted IVC source on average which makes sense since the Southern IVC was more Iran_N rich (Balochistan) but a more BMAC heavy source as well, so they are unique and some have later plateau Iranian ancestry, which is not surprising because of intermarriages over the past millennia. That being said their substrate population would still be closely related with the more Western shifted SPGT samples, there is very little room for what their substrate population can be. I just look at Southern Pashtuns as Pamiri like but much more IVC shifted with a dash of plateau Iranian ancestry, though so have much more but some on the other hand like that Kandahar sample above doesn't need it at all. Pashtuns from other regions show a shift according to the region, though not always. It really is a case by case basis.

Coldmountains
05-20-2020, 06:39 PM
The model suggests BMAC populations mixed with Steppe people and proceeded to mix with tribals or HGs . Those Kabul River Pashtuns ( Abasin /Spin Ghar) types are heavily SPGT like and u can see it from their AASI values, that Paghman one even with Turkic related ancestry is still harboring close to 10% AASI and it shows in his models. Southern Pashtuns are a different for sure, because they typically need a more Zagrosian shifted IVC source on average which makes sense since the Southern IVC was more Iran_N rich (Balochistan) but a more BMAC heavy source as well, so they are unique and some have later plateau Iranian ancestry, which is not surprising because of intermarriages over the past millennia. That being said their substrate population would still be closely related with the more Western shifted SPGT samples, there is very little room for what their substrate population can be. I just look at Southern Pashtuns as Pamiri like but much more IVC shifted with a dash of plateau Iranian ancestry, though so have much more but some on the other hand like that Kandahar sample above doesn't need it at all. Pashtuns from other regions show a shift according to the region, though not always. It really is a case by case basis.

I created a Bronze Age model because i tried to use components which don't overlap too much and are not too similar with each other. Of course this is not perfect and i see the problems with it. The idea was to see how Pashtun subgroups/tribes differ from each other in terms of basic ancient components. Using later Iron Age genomes from Central Asia and Pakistan, we often see that components behave unstable and fluctuate extremely from individual to individual because they overlap quite a lot with each other and share some kind of specific ancestry.

It is generally seen by most linguists and historians, that Proto-Pashtuns orginated in Arachosia and migrated in the early mediveal/medieval period from the this region into Gandhara and the Hindukush region more in the north. So ancient genomes from Arachosia would be most relevant for modelling Proto-Pashtuns.

But Iron Age groups from Arachosia will be not that different to Pakistan_IA genomes, we already have on Global25, that is why modelling Proto-Pashtun ancestry among modern day Pashtuns is not easy. If we had some Iron Age Arachosians we could compare them to modern day Pashtuns, but i think they will cluster pretty close to Kandahari and southern Pashtuns just with less East Asian. There don't seem to be a big genetic change in the region and if southern Pashtuns had significant recent ancestry from the West (Iran_C shifted groups) they would also have less Steppe compared to other South Central Asians in the region, but actually they score around 30% and are on of the most steppe-shifted groups in the region. They have significantly higher Steppe than eastern Pashtuns who seem to be around 25%.

So i conclude, that they are quite good as reference for Proto-Pashtun ancestry in other Pashtun subgroups more in the north and east.

pegasus
05-20-2020, 09:24 PM
I created a Bronze Age model because i tried to use components which don't overlap too much and are not too similar with each other. Of course this is not perfect and i see the problems with it. The idea was to see how Pashtun subgroups/tribes differ from each other in terms of basic ancient components. Using later Iron Age genomes from Central Asia and Pakistan, we often see that components behave unstable and fluctuate extremely from individual to individual because they overlap quite a lot with each other and share some kind of specific ancestry.

It is generally seen by most linguists and historians, that Proto-Pashtuns orginated in Arachosia and migrated in the early mediveal/medieval period from the this region into Gandhara and the Hindukush region more in the north. So ancient genomes from Arachosia would be most relevant for modelling Proto-Pashtuns.

But Iron Age groups from Arachosia will be not that different to Pakistan_IA genomes, we already have on Global25, that is why modelling Proto-Pashtun ancestry among modern day Pashtuns is not easy. If we had some Iron Age Arachosians we could compare them to modern day Pashtuns, but i think they will cluster pretty close to Kandahari and southern Pashtuns just with less East Asian. There don't seem to be a big genetic change in the region and if southern Pashtuns had significant recent ancestry from the West (Iran_C shifted groups) they would also have less Steppe compared to other South Central Asians in the region, but actually they score around 30% and are on of the most steppe-shifted groups in the region. They have significantly higher Steppe than eastern Pashtuns who seem to be around 25%.

So i conclude, that they are quite good as reference for Proto-Pashtun ancestry in other Pashtun subgroups more in the north and east.

Thats not what linguists think at all, the greatest linguistic diversity of Pashto languages is not in Arachosia its in the Northern Balochistan/Sulaiman zone and for good and logical reasons, you have Tarino which has archaic features which you actually see in Pamiri languages but basal with other Pashto dialects. Also, there is the specific absorption of the Indo Aryan substratum but really its that Para Munda and/or Dravidian sound structure absorbed into Indo Aryan again that limits where the location would be, so at best eastern borderlands of Arachosia with Balochistan, but its due to the proximity to the later. Also, I am no linguist by profession but as a speaker of Indo Aryan and Iranic languages you can pick up on the contrasts. I don't think Proto Pashtuns formed that late either , they would have to be significantly Steppe rich, much more than what modern Pashtuns from the Jnub have , also Pashtuns from Khost who are even more Steppe shifted, that supports that. It looks most of the main populations were formed in the IA with internal movements in the Classical Age and Antiquity. Though I don't think there was a massive onslaught of them in the Gandhara region, many of them moved there from adjacent areas but they still fall on that continuum established millennia earlier. Ergo, even with populations on the Western limits of that cultural zone , there is still significant continuity.


"sample": VelvetNono_AGUser",
"fit": 1.7089,
"PAK_Katelai_IA": 75,
"TJK_Ksirov_H_Kushan": 25,

"sample": "VelvetNono_AGUser",
"fit": 1.8284,
"PAK_Katelai_IA": 59.17,
"PAK_Barikot_IA": 27.5,
"KAZ_Kangju": 13.33

"sample": "VelvetNono_AGUser",
"fit": 1.6423,
"PAK_Katelai_IA": 85.83,
"Iberia_Northeast_Empuries2": 7.5,
"KAZ_Kangju": 6.67,


Other models for those samples they interestingly take in ancestry from Hellenic populations, whether this is legit will have to be tested on qpAdm when their sample sets and software is updated. Though they are interesting and testament for Afghanistan being a cultural crossroad.


"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.0491,
"PAK_Katelai_IA": 45.83,
"PAK_Loebanr_IA_o": 38.33,
"Iberia_Northeast_Empuries2": 11.67,
"MNG_Xiong_Nu_East_Asian": 4.17,

"sample": "Custom:Pashtun_North_Afghanistan",
"fit": 0.7238,
"PAK_Loebanr_IA": 55.83,
"KAZ_Kangju": 32.5,
"IRN_Seh_Gabi_C": 6.67,
"Iberia_Northeast_Empuries2": 5,

Mingle
05-21-2020, 08:02 AM
Thats not what linguists think at all, the greatest linguistic diversity of Pashto languages is not in Arachosia its in the Northern Balochistan/Sulaiman zone and for good and logical reasons, you have Tarino which has archaic features which you actually see in Pamiri languages but basal with other Pashto dialects.

Distinguishing Arachosia from the Sulaiman Zone seems like nitpicking. This is a map of Arachosia:

https://upload.wikimedia.org/wikipedia/commons/2/26/Afghanistan_region_during_500_BC.jpg

Seems to overlap into North Balochistan. Besides, North Balochistan is often considered part of the "Greater Kandahar" cultural zone.

Anyways, North Balochistan borders Kandahar. The closest people to North Balochistani Pashtuns are going to be Kandahari Pashtuns, not Pamiris or whatever. And we don't have any G25 coordinates from NB but we do from Kandahar, so it makes sense to use them as a proxy for Proto-Pashtuns.

I don't see how arguing against them originating in Kandahar means that Kangju are the next best fit. Nothing shows that Proto-Pashtuns from North Balochistan should resemble Kangju.


I don't think Proto Pashtuns formed that late either , they would have to be significantly Steppe rich, much more than what modern Pashtuns from the Jnub have , also Pashtuns from Khost who are even more Steppe shifted, that supports that. It looks most of the main populations were formed in the IA with internal movements in the Classical Age and Antiquity. Though I don't think there was a massive onslaught of them in the Gandhara region, many of them moved there from adjacent areas but they still fall on that continuum established millennia earlier. Ergo, even with populations on the Western limits of that cultural zone , there is still significant continuity.

Khost Pashtuns' high Steppe is probably from the previous East Iranic inhabitants of the region before Pashtuns reached it. Southerners probably have less Steppe than them cause they're further south bordering Steppe-deprived Balochistan/Gedrosia.

Coldmountains
05-21-2020, 08:48 AM
Thats not what linguists think at all, the greatest linguistic diversity of Pashto languages is not in Arachosia its in the Northern Balochistan/Sulaiman zone and for good and logical reasons, you have Tarino which has archaic features which you actually see in Pamiri languages but basal with other Pashto dialects. Also, there is the specific absorption of the Indo Aryan substratum but really its that Para Munda and/or Dravidian sound structure absorbed into Indo Aryan again that limits where the location would be, so at best eastern borderlands of Arachosia with Balochistan, but its due to the proximity to the later. Also, I am no linguist by profession but as a speaker of Indo Aryan and Iranic languages you can pick up on the contrasts. I don't think Proto Pashtuns formed that late either , they would have to be significantly Steppe rich, much more than what modern Pashtuns from the Jnub have , also Pashtuns from Khost who are even more Steppe shifted, that supports that. It looks most of the main populations were formed in the IA with internal movements in the Classical Age and Antiquity. Though I don't think there was a massive onslaught of them in the Gandhara region, many of them moved there from adjacent areas but they still fall on that continuum established millennia earlier. Ergo, even with populations on the Western limits of that cultural zone , there is still significant continuity.


"sample": VelvetNono_AGUser",
"fit": 1.7089,
"PAK_Katelai_IA": 75,
"TJK_Ksirov_H_Kushan": 25,

"sample": "VelvetNono_AGUser",
"fit": 1.8284,
"PAK_Katelai_IA": 59.17,
"PAK_Barikot_IA": 27.5,
"KAZ_Kangju": 13.33

"sample": "VelvetNono_AGUser",
"fit": 1.6423,
"PAK_Katelai_IA": 85.83,
"Iberia_Northeast_Empuries2": 7.5,
"KAZ_Kangju": 6.67,


Other models for those samples they interestingly take in ancestry from Hellenic populations, whether this is legit will have to be tested on qpAdm when their sample sets and software is updated. Though they are interesting and testament for Afghanistan being a cultural crossroad.


"sample": "Custom:Pashtun_Kabul_Paghman",
"fit": 1.0491,
"PAK_Katelai_IA": 45.83,
"PAK_Loebanr_IA_o": 38.33,
"Iberia_Northeast_Empuries2": 11.67,
"MNG_Xiong_Nu_East_Asian": 4.17,

"sample": "Custom:Pashtun_North_Afghanistan",
"fit": 0.7238,
"PAK_Loebanr_IA": 55.83,
"KAZ_Kangju": 32.5,
"IRN_Seh_Gabi_C": 6.67,
"Iberia_Northeast_Empuries2": 5,

Arachosia and the Northern Baluchistan/Suleiman zone are not really cleary distinct regions and it does not really make such a big difference from a genetic point of view, where exactly in Arachosia Pashtuns orginated. More important is that southern Pashtuns speak more conservative Pashto and populate a region with no documented substrate/local non-pashtun languages in the last 1500-2000 years. Central Pashtuns with higher steppe on the otherside populate regions where some Pre-Pashtun languages like Ormuri are still spoken. Also Ptolemy around 2000 years ago mentioned a tribe called Pargyetae in Arachosia, which ethnonym is related to Parsa/Pashtun.

Based on most linguistic evidences Pashto just formed 1500-2000 years ago. Their is not such a great difference between Pashtun dialects to locate it much more back in time.


Except for a few details, Paṣ̌tō dialects can be derived from a prototype not essentially different from the classical 10th/16th century literary language; they do not to any significant extent help us to reconstruct a more archaic form of Paṣ̌tō. There is only one dialect which stands decidedly apart, i.e., Waṇ(ecī) (or Tarīno) spoken in northeast Baluchistan between Harnai and Loralai, and now being more and more influenced by and pushed back by ordinary Paṣ̌tō.

Central Pashtuns from Khost populate regions, which were once populated by other steppe rich East Iranics (Ormuri, Parachi,..) and even some eastern Pashtuns can score around 30% steppe but on average Kandahari Pashtuns have righer steppe than eastern Pashtuns. Proto-Pashtuns were not always necessarily more steppe-shifted than the locals they assimilated/mixed with. So the Pashtuns with highest steppe admix are not necessarily most similar to Proto-Pashtuns.

I don't get significant Greek admixture for modern day Pashtuns at least for the regional/tribal averages.

Target: Pashtun_North_Afghanistan_AVG
Distance: 0.4877% / 0.00487700
46.6 BMAC
28.0 Steppe
17.6 IRN_Shahr_I_Sokhta_BA2
7.2 East_Asian
0.6 GRC_Mycenaean

Target: Pashtun_Kandahar_AVG
Distance: 0.9486% / 0.00948592
46.2 BMAC
28.0 Steppe
18.4 IRN_Shahr_I_Sokhta_BA2
4.2 East_Asian
2.2 Iran_C
1.0 GRC_Mycenaean

Target: Pashtun_Kabul_AVG
Distance: 0.7279% / 0.00727939
41.6 BMAC
25.2 Steppe
20.8 IRN_Shahr_I_Sokhta_BA2
6.6 East_Asian
3.0 Botai
2.8 GRC_Mycenaean

Greek admix is showing here up because you use too eastern and too steppe shifted Iron Age reference groups for Pashtuns. Kangju is not directly relevant for Pashtuns and Pakistan_IA genomes are too ASI and generally too eastern shifted for most Afghan Pashtuns. We lack good IA reference groups for Pashtuns. TKM_IA is too steppe-shifted and Pakistan_IA too eastern shifted. Unfornuately we will probably not so soon get ancient dna from Afghanistan or Arachosia.

Coldmountains
05-21-2020, 11:07 AM
Interestingly the regional/tribal averages form two clines, which meet where Kandahri Pashtuns are positioned on this pca. Pashtuns from North Afghanistan and Kabul sit on a cline, which probably shows East Iranic/Turkic-admixed Pashtuns in Northern Afghanistan and Central Pashtunistan. The other cline which i think most modern day Pashtuns belong to has Kandahari on one end and Pakistan_IA groups (especially Saidu Sharif) on the other end.

https://i.ibb.co/VWKLjNg/Pashtun-PCA.png (https://ibb.co/0jwnDXC)

gudar
05-21-2020, 12:09 PM
Southern Pashtuns have elevated steppe ancestry and low South Asian. South Afghanistan was once also called Sistan/Sakastan. Maybe they are indeed mixed with Saka. Northeastern Tajiks from mountainous regions (Panjshir, Badakhshan...) are very similar to Pamiri just more South Asian-shifted than other Pamiri or Tajiks.

one irani guy on another forum posted that many sistani from sistan balochistan province along afghan/pak border get high steppe

Gee
05-21-2020, 12:15 PM
Southern Pashtuns have elevated steppe ancestry and low South Asian. South Afghanistan was once also called Sistan/Sakastan. Maybe they are indeed mixed with Saka. Northeastern Tajiks from mountainous regions (Panjshir, Badakhshan...) are very similar to Pamiri just more South Asian-shifted than other Pamiri or Tajiks.

one irani guy on another forum posted that many sistani from sistan balochistan province along afghan/pak border get high steppe

Not tajiks from pansher. Theyre just like pashtuns actually, after seeing half pansheris and pansheri tajik results. Though badakshanis would be more like pamiris a lot

Coldmountains
05-21-2020, 01:40 PM
Southern Pashtuns have elevated steppe ancestry and low South Asian. South Afghanistan was once also called Sistan/Sakastan. Maybe they are indeed mixed with Saka. Northeastern Tajiks from mountainous regions (Panjshir, Badakhshan...) are very similar to Pamiri just more South Asian-shifted than other Pamiri or Tajiks.

one irani guy on another forum posted that many sistani from sistan balochistan province along afghan/pak border get high steppe

we can not fully answer this question before we not have ancient dna from Arachosia from various time periods to compare with. But i tend now to think that Saka had a low genetic impact in Afghanistan. Much of the extra East Asian some Pashtuns have can be explained with very recent admixture with Hazara, Tajiks and Uzbeks and early pre-Saka Indo-Iranian steppe groups who mixed with Central Asian locals (Botai,..)

Coldmountains
05-21-2020, 01:49 PM
Not tajiks from pansher. Theyre just like pashtuns actually, after seeing half pansheris and pansheri tajik results. Though badakshanis would be more like pamiris a lot

Pansjhiri are more shifted towards Kalash/Nuristani than many Pashtuns.

Gee
05-21-2020, 02:03 PM
Pansjhiri are more shifted to Kalash/Nuristani than many Pashtuns.

Not the ones i seen. They seem just pashtun tbh

gudar
05-21-2020, 02:04 PM
we can not fully answer this question before we not have ancient dna from Arachosia from various time periods to compare with. But i tend now to think that Saka had a low genetic impact in Afghanistan. Much of the extra East Asian some Pashtuns have can be explained with very recent admixture with Hazara, Tajiks and Uzbeks and early pre-Saka Indo-Iranian steppe groups who mixed with Central Asian locals (Botai,..)
yes , under sampling breeds lot of false theories

Coldmountains
05-21-2020, 10:17 PM
Interestingly the regional/tribal averages form two clines, which meet where Kandahri Pashtuns are positioned on this pca. Pashtuns from North Afghanistan and Kabul sit on a cline, which probably shows East Iranic/Turkic-admixed Pashtuns in Northern Afghanistan and Central Pashtunistan. The other cline which i think most modern day Pashtuns belong to has Kandahari on one end and Pakistan_IA groups (especially Saidu Sharif) on the other end.

https://i.ibb.co/VWKLjNg/Pashtun-PCA.png (https://ibb.co/0jwnDXC)

Based on this i created an improved calc to model Proto-Pashtun and Pre-Pashtun admix among modern day Pashtuns. Kandahari are used here as proxy for Proto-Pashtun-like admixture among modern day Pashtuns.

https://i.ibb.co/bmPq8XB/Unbenannt.jpg (https://ibb.co/mTG1M4q)

Pre-Pashtun Gandhara: Pakistan_IA+Kalash

East Iranic: TJK Kushan, Wusun, Kangju, Saka_Tian_Shan_O (Pamiri-like)

Turk: Karluk, Karakhanid

pegasus
05-22-2020, 06:03 PM
Arachosia and the Northern Baluchistan/Suleiman zone are not really cleary distinct regions and it does not really make such a big difference from a genetic point of view, where exactly in Arachosia Pashtuns orginated. More important is that southern Pashtuns speak more conservative Pashto and populate a region with no documented substrate/local non-pashtun languages in the last 1500-2000 years. Central Pashtuns with higher steppe on the otherside populate regions where some Pre-Pashtun languages like Ormuri are still spoken. Also Ptolemy around 2000 years ago mentioned a tribe called Pargyetae in Arachosia, which ethnonym is related to Parsa/Pashtun.

Based on most linguistic evidences Pashto just formed 1500-2000 years ago. Their is not such a great difference between Pashtun dialects to locate it much more back in time.



Central Pashtuns from Khost populate regions, which were once populated by other steppe rich East Iranics (Ormuri, Parachi,..) and even some eastern Pashtuns can score around 30% steppe but on average Kandahari Pashtuns have righer steppe than eastern Pashtuns. Proto-Pashtuns were not always necessarily more steppe-shifted than the locals they assimilated/mixed with. So the Pashtuns with highest steppe admix are not necessarily most similar to Proto-Pashtuns.

I don't get significant Greek admixture for modern day Pashtuns at least for the regional/tribal averages.

Target: Pashtun_North_Afghanistan_AVG
Distance: 0.4877% / 0.00487700
46.6 BMAC
28.0 Steppe
17.6 IRN_Shahr_I_Sokhta_BA2
7.2 East_Asian
0.6 GRC_Mycenaean

Target: Pashtun_Kandahar_AVG
Distance: 0.9486% / 0.00948592
46.2 BMAC
28.0 Steppe
18.4 IRN_Shahr_I_Sokhta_BA2
4.2 East_Asian
2.2 Iran_C
1.0 GRC_Mycenaean

Target: Pashtun_Kabul_AVG
Distance: 0.7279% / 0.00727939
41.6 BMAC
25.2 Steppe
20.8 IRN_Shahr_I_Sokhta_BA2
6.6 East_Asian
3.0 Botai
2.8 GRC_Mycenaean

Greek admix is showing here up because you use too eastern and too steppe shifted Iron Age reference groups for Pashtuns. Kangju is not directly relevant for Pashtuns and Pakistan_IA genomes are too ASI and generally too eastern shifted for most Afghan Pashtuns. We lack good IA reference groups for Pashtuns. TKM_IA is too steppe-shifted and Pakistan_IA too eastern shifted. Unfornuately we will probably not so soon get ancient dna from Afghanistan or Arachosia.

Whether they are relevant or not they serve as excellent proxies for Steppe, using modern-day populations to show interspecific variation is heavily flawed because of a plethora of issues, so it is pointless to model them as iterations of each other. They typically have 7-10% AASI , so you would need populations with more than that surely, the Katelai and Barikot cluster has groups with slightly more than that, so thats incorrect that they are too eastern. Loebanr outlier looks from further West, she has too low AASI, GD had that Paghman sample with her and Katelai. Velvet is from a region 200 km west of the SPGT cluster but still can be modeled as 80-85% SPGT like and he is not a lone ranger since former user Barnacle was exactly like him. As I stated earlier Southern Pashtuns need a more ANF enriched BMAC source coupled with a more Zagrosian shifted IVC source. There is very little room because their positions on the PCA are quite limited which I mentioned a few times prior.

I think you are projecting a misguided Slavic like ethnogenesis ( that too occurring in the MA) on these people is but its clearly far nuanced because the mixing is far more complicated because a) their Zagrosian ancestry is a mix of a varying Turan/BMAC and a very heterogeneous IVC population, further the BMAC population in itself is 10% IVC like . On the Steppe front , for Pashtuns their Steppe sources are clearly mixed with some taking more Okunevo like sources over Central Steppe EBA, others having an almost Western shift like that Kokcha sample. Its clear they harbor ancestries from different IIr streams. The cline for them runs East to West, the reason for the cline pulling upwards is those 'northern" populations are showing increased ENA/Turkic ancestry thats why but thats not what defines them . Post IA ASI substratum will definitely fall on that SPGT cline or a closely related population , they are not descended from a Balochi substrate. Ethnogenesis in the early MA during the rise of Turkic dynasties makes no sense as these populations were subjects of Turks not rulers. So a much earlier genesis is definite and one that clearly involves a much more Steppe rich progenitor and thats why Kangju works because it sits on that cline. Both Chinese and Indian chroniclers mention the Suleiman Range as the abode of Pashtuns in the 4th to 6th centuries CA, Arachosia ironically was considered Hellenic influenced even at that time, Indian sources can be a bit debatable but its in line with the Chinese description who tended to more meticulous with details.

Coldmountains
05-22-2020, 06:55 PM
Whether they are relevant or not they serve as excellent proxies for Steppe, using modern-day populations to show interspecific variation is heavily flawed because of a plethora of issues. They typically have 7-10% AASI , so you would need populations with more than that surely, the Katelai and Barikot cluster has groups with slightly more than that, so thats incorrect that they are too eastern. Loebanr outlier looks from further West, she has too low AASI, GD had that Paghman sample with her and Katelai. Velvet is from a region 200 km west of the SPGT cluster but still can be modeled as 80-85% SPGT like and he is not a lone ranger since former user Barnacle was exactly like him. As I stated earlier for Southern Pashtuns are more ANF enriched BMAC source coupled with a more Zagrosian shifted IVC source is needed. There is very little room because their positions on the PCA are quite limited which I mentioned few times prior.

I think your projecting a highly misguided Slavic like ethnogenesis ( that too occurring in the MA) on these people but its clearly far nuanced because the mixing is far more complicated because a) their Zagrosian ancestry is a mix of a varying Turan/BMAC and a very heterogeneous IVC population, further the BMAC population in itself is 10% IVC like . On the Steppe front , for Pashtuns their Steppe sources are clearly mixed with some taking more Okunevo like sources over Central Steppe EBA, others having an almost Western shift like that Kokcha sample. Its clear they harbor ancestries from different IIr streams. The cline for them runs East to West, the reason for the cline pulling upwards is those 'northern" populations are showing increased ENA/Turkic ancestry thats why but thats not what defies them . Post IA ASI substratum will definitely fall on that SPGT cline or a closely related population , they are not descended from a Balochi substrate. Ethnogenesis in the early MA during the rise of Turkic dynasties makes no sense as these populations were subjects of Turks not rulers. So a much earlier genesis is definite and one that clearly involves a much more Steppe rich progenitor and thats why Kangju works because it sits on that cline. Both Chinese and Indian chroniclers mention the Suleiman Range as the abode of Pashtuns in the 4th to 6th centuries CA, Arachosia ironically was considered considerably Hellenic influenced even at that time, Indian sources can be a bit debatable but its in line with the Chinese description who tended to more meticulous with details.

well there are many reason mentioned already in this thread why modern day South Pashtuns are indeed a good proxy for Proto-Pashtuns. You can continue to distract from the discussion by pointing to irrelevant geographic distinctions between Arachosia and the Suleiman mouintans but the point is that Proto-Pashtuns formed 1500-2000 years ago in Arachosia/Suleiman mountain region by a group closest to modern day Kandahari and South Pashtuns in North Baluchistan, which are on the Global25 South Asia pca exactly on the place where both Pashtun clines meet. Kangju or some Pamiri-like East Iranics are irrelevant here and using them you get unhistorical models. Pashtuns from NB would be probably slightly closer to Proto-Pashtuns, but we don't have their Global25 coordinates so Kandahari are for now the best proxy for Proto-Pashtun ancestry.

I don't really see much possible recent admixture in Arachosia especially when we consider that most Pashtuns for a long time were a pastoralist and semi-nomadic groups, which only recently mixed with urban groups. Arachosia had not such a big urban, sedentary or non-Pashtun population to mix with like Gandhara or Tajik-dominated regions in the north.

You don't show me a better model for Pashtuns with relevant reference groups. I used both Dardic-like and Khatri-like sources for Pashtuns because they not got all of their Pre-Pashtun ancestry from Kalash-like populations. Gandhara was very diverse (Dardic-like, Khatri-like, Gujarat Bramin-like, Jatt-like,..) and modern day Pashtuns in Gandhara/eastern Afghanistan can be quite diverse depending on which group they mixed with and how much they mixed in the first place. But like i said it is not easy to differentiate between local and Proto-Pashtun ancestry, because Proto-Pashtuns and Pre-Pashtuns had both similar autosomal and uniparental dna. Also it were not just Indo-Aryan groups they mixed with in Central Pashtunistan there was also mixing with other East Iranics like Ormuri,what makes it even harder to differentiate between Proto-Pashtuns and Pre-Pashtuns. The situation is similar to Belarus, where Slavs rather recently mixed with closely related Balts, what makes it difficult to differentiate between Slavic and Baltic-like admixture.

You go to much back in time. Pashtuns expanded recently and formed rather recently compared to other groups. If you have another opinion show me a linguistic or historical source/study for Proto-Pashtuns forming or expanding from their homeland long before 1500-2000 years ago. There is no point in using BMAC or Turan_IA genomes to model Proto-Pashtun admixture here, because neither they are relevant for the time period nor for the location. This is like using Corded Ware as proxy for Slavic ancestry in Eastern Europe.

The last point that Pashtuns could not expand in Medieval times because they had Turkic rulers i don't really understand. Slavs also expanded into the Balkans during Avar rule for example. Turkic rule in the region was very decentralized and loosely. Early Pashtuns were often allied with them and we see this in the adoption of turkic terms and dynasties having both turkic and afghan characteristics like Khalji dynasty. Turkic rule rather helped Pashtuns to expand further eastwards.

Dr_McNinja
05-23-2020, 06:22 AM
Based on this i created an improved calc to model Proto-Pashtun and Pre-Pashtun admix among modern day Pashtuns. Kandahari are used here as proxy for Proto-Pashtun-like admixture among modern day Pashtuns.

https://i.ibb.co/bmPq8XB/Unbenannt.jpg (https://ibb.co/mTG1M4q)

Pre-Pashtun Gandhara: Pakistan_IA+Kalash

East Iranic: TJK Kushan, Wusun, Kangju, Saka_Tian_Shan_O (Pamiri-like)

Turk: Karluk, Karakhanid

What's the benefit of running with ADC set to 0.25x?

Coldmountains
05-23-2020, 06:25 AM
What's the benefit of running with ADC set to 0.25x?

The reference groups are too closely related unlike in the Bronze Age calc, where most of the source groups like Steppe, BMAC, East Asian, ... don't overlap very much in terms of autosomal dna. Late Iron Age Arachosians best represented by modern day South Pashtuns in my opinion were not that different from Pakistan_IA samples just more steppe, more BMAC/Iran_C and less IVC shifted. But on the otherside Proto-Pashtuns would also to some extent overlap with other East Iranic groups further north. So i used ADC 0.25 (similar to penalty on nmonte), which is recommended for calcs with closely related source groups.

Dr_McNinja
05-23-2020, 07:08 AM
The reference groups are too closely related unlike in the Bronze Age calc, where most of the source groups like Steppe, BMAC, East Asian, ... don't overlap very much in terms of autosomal dna. Late Iron Age Arachosians best represented by modern day South Pashtuns in my opinion were not that different from Pakistan_IA samples just more steppe, more BMAC/Iran_C and less IVC shifted. But on the otherside Proto-Pashtuns would also to some extent overlap with other East Iranic groups further north. So i used ADC 0.25 (similar to penalty on nmonte), which is recommended for calcs with closely related source groups.

Hm, should I use it when using multiple Steppe populations (like Sintashta, Kangju)

Coldmountains
05-23-2020, 07:19 AM
Hm, should I use it when using multiple Steppe populations (like Sintashta, Kangju)

Good question. You need to try both and check if the components/admixture results are without ADC 0.25 stable and logical based on what makes sense in the region. But i think because Kangju being more similar to South Asians and especially South Central Asians, it will be extremely favoured when you use penalty/ADC 0.25. So for Pashtuns and Tajiks i would not use it combined with penalty/ADC 0.25. Not sure about South Asians.

But i personally would also not use Sintashta and Kangju at the same time because Kangju has Sintashta admixture. Sintashta is ancestral to Kangju what can have weird effects if you use both. For Pashtuns we pretty much lack good reference groups from the Iron Age or earlier, which would represent the Pre-Proto-Pashtun Iranic element in the region. Kangju are too late to be ancestral to Pre-Proto-Pashtuns, which arrived probably around 2500-3000 years ago from the north in the Arachosia/Suleiman mountain region in my opinion

Coldmountains
05-23-2020, 02:59 PM
.....

Cynic
05-23-2021, 03:38 AM
The model suggests BMAC populations mixed with Steppe people and proceeded to mix with tribals or HGs . Those Kabul River Pashtuns ( Abasin /Spin Ghar) types are heavily SPGT like and u can see it from their AASI values, that Paghman one even with Turkic related ancestry is still harboring close to 10% AASI and it shows in his models. Southern Pashtuns are a different for sure, because they typically need a more Zagrosian shifted IVC source on average which makes sense since the Southern IVC was more Iran_N rich (Balochistan) but a more BMAC heavy source as well, so they are unique and some have later plateau Iranian ancestry, which is not surprising because of intermarriages over the past millennia. That being said their substrate population would still be closely related with the more Western shifted SPGT samples, there is very little room for what their substrate population can be. I just look at Southern Pashtuns as Pamiri like but much more IVC shifted with a dash of plateau Iranian ancestry, though so have much more but some on the other hand like that Kandahar sample above doesn't need it at all. Pashtuns from other regions show a shift according to the region, though not always. It really is a case by case basis.

I like to think of them as being a roughly equal mix between Baloch and Pamiri:

Target: Pashtun_Kandahar
Distance: 1.2250% / 0.01224975
55.2 Balochi
44.8 Tajik_Rushan

ViktorL1
07-22-2021, 09:00 PM
I collected some Pashtun Global25 coordinates and created some averages for Pashtuns from Kandahar, Kabul/Paghman and North Afghanistan (only based on academic samples)





Also i modelled these Pashtuns and various Pashtun groups, which we already had in the offical global25 spreadsheet. The amount of Steppe and BMAC-like ancestry is similar in all Pashtun subgroups here but there seems to be significant differences in terms of ASI-like and East Asian-like ancestry. Generally ASI- or IVC-like ancestry is peaking among most eastern Pashtuns and East Asian-like ancestry among Pashtuns in North Afghanistan. There is also some minor Iran_C-like shift among Kandahari Pashtuns. WSHG on the otherhand peaks here among eastern Pashtuns, which also tend to be the most IVC/ASI-shifted Pashtuns.

https://i.ibb.co/qDc836v/ADASD.jpg (https://ibb.co/S6Mg4Sz)

Unfornuately i have not found Afghan Pashtun global25 coordinates east of Kabul but Pashtuns from regions like Nangarhar are probably inbetween Kabuli Pashtuns and Tarkalani/Yusufzai Pashtuns here. PAK_Udegram_MA_Ghaznavid is on this pca also on this position and likely from East Afghanistan.

https://i.ibb.co/PwbnYXr/newplot-3.png (https://ibb.co/t8NthjH)


Generally it is assumed that Pashtuns from Kandahar and southeastern Afghanistan are most representative for Proto-Pashtuns and have the least Pre-Pashtun admixture. So i tried to model Pashtuns as mix of Kandahari Pashtuns + Non-Pashtun groups (Pamiri-like East Iranics, Dardics, Burusho, Uzbeks, Turkmens,..). Based on these model Proto-Pashtun-like ancestry is around 40-50% among eastern Pashtuns in Pakistan and around 50-60% among Pashtuns in northern Afghanistan,but this is probably quite different from individual to individual and not fully accurate.

https://i.ibb.co/JQJYcGy/Unbenannt.jpg (https://ibb.co/k1CLGw9)

Which model/calc are you using here? Could you post the coordinates?