Global25 automated nMonte for South/Central Asian members

[Coldmountains]

Using Iran Chl or Hasanlu/Hajji Firuz IA alone failed surprisingly, he needed a distinctively Levantine source , Ebla or Levant BA, and a lot of it. Even if you look at the stats , he still needs a more Levantine source, not Iran more Syrian/Lebanese type ancestry. I guess there was more gene flow from Levant than I thought. Yes, Most Kandahar/Helmand Pashtuns definitely would have gotten further Plateau Iranian ancestry, along with Balochis, when it arrives is a good question. What was the subclade of the R1a for Ghaznavid 2959??

Without ancient dna from Afghanistan or Arachosia we are using here only indirect proxies and not direct ancestors of Pashtuns or Pashtun-admixed people. Shar-I-Sokhta and Alakh_MLBA are in my opininion too eastern in terms of ancestry. Kazakhstan_MLBA_OyDzhaylau is also too steppe-shifted. I am not excluding Levantine during the early Islamic period but i have seen that many Pashtuns score Levantine-stuff when we use just IVC and BMAC genomes. So i rather think the "Levantine" shift effected most western Pashtuns and is probably caused by different layers of West Asian geneflow (including minor levantine/arabic) starting from the achaemenid period and intensyfing with the arrival of Islam. Most Afghan Pashtuns when they are not very Dardic-like score on Global25 more or at least as much Levantine-like ancestry as this Udegram sample. But maybe something is not correct with global25 or my model itself here. Neverthless some of the Y-DNA G clades could have a Mesopotomian origin for example like G-M283, which closest parallel clades are found among Jews and Italians.


Distance: 0.8763% / 0.00876309
Target: Pashtun_North_Afghanistan_AVG
56.6 Gandhara_IA
21.9 Gandhara_IA (Udegram IA)
11.2 Levant_Megiddo_MLBA
10.3 Early_Turk

Distance: 1.3997% / 0.01399703
Target: PAK_Udegram_MA_Ghaznavid:I7716
74.7 Gandhara_IA (Udegram IA)
19.5 LBA/EIA_Central_Asia (TKM_IA + Dashti_Kozy)
4.1 Levant_Hazor_MLBA
1.4 Levant_Megiddo_MLBA
0.3 Early_Turk

Distance: 1.3062% / 0.01306230
Target: Pashtun_Kandahar_AVG
59.7 Gandhara_IA (Udegram IA)
22.0 LBA/EIA_Central_Asia (TKM_IA + Dashti_Kozy)
11.8 Levant_Megiddo_MLBA
3.9 Saka_Hun-like
2.6 Early_Turk

If i remember correctly the Ghaznavid 2959 sample is of too low quality to say much about his R1a. He was found positive for Z93 but more can not be said about him. Based on modern Pasthun R1a in the same region he sould be YP413+ but this is of course speculative.

[pegasus]

Without ancient dna from Afghanistan or Arachosia we are using here only indirect proxies and not direct ancestors of Pashtuns or Pashtun-admixed people. Shar-I-Sokhta and Alakh_MLBA are too eastern in terms of ancestry and too old. Kazakhstan_MLBA_OyDzhaylau is also too steppe-shifted. I am not excluding Levantine during the early Islamic period but i have seen that many Pashtuns score Levantine-stuff when we use just IVC and BMAC genomes. So i rather think the "Levantine" shift effected most western Pashtuns and is probably caused by different layers of West Asian geneflow starting from the achaemenid period and intensyfing with the arrival of Islam . Most Afghan Pashtuns when they are not very Dardic-like score on Global25 more or at least as much Levantine-like ancestry as this Udegram sample. But maybe something is not correct with global25 itself here. Neverthless some of the Y-DNA G clades could have a Mesopotomian origin for example like G-M283.
Distance: 0.8763% / 0.00876309
Target: Pashtun_North_Afghanistan_AVG
56.6 Gandhara_IA
21.9 Gandhara_IA (Udegram IA)
11.2 Levant_Megiddo_MLBA
10.3 Early_Turk

Distance: 1.3997% / 0.01399703
Target: PAK_Udegram_MA_Ghaznavid:I7716
74.7 Gandhara_IA (Udegram IA)
19.5 LBA/EIA_Central_Asia (TKM_IA + Dashti_Kozy)
4.1 Levant_Hazor_MLBA
1.4 Levant_Megiddo_MLBA
0.3 Early_Turk

Distance: 1.3062% / 0.01306230
Target: Pashtun_Kandahar_AVG
59.7 Gandhara_IA (Udegram IA)
22.0 LBA/EIA_Central_Asia (TKM_IA + Dashti_Kozy)
11.8 Levant_Megiddo_MLBA
3.9 Saka_Hun-like
2.6 Early_Turk

If i remember correctly the Ghaznavid 2959 sample is of too low quality to say much about his R1a. He was found positive for Z93 but more can not be said about him. Based on modern Pasthun R1a in the same region he sould be YP413+ but this is of course speculative.

Your probably right, hopefully 2959 should fall more in the regional pattern, but he has way too much ANF on G25. Though this Levantine shift in this guy is likely isolated and not indicative of Pashtuns , extra West Asian ancestry would be of Plateau Iranian origin not beyond that esp on a massive scale . Though even then the extra West Asian input would not exceed that of the BMAC or IVC related ancestry. It just goes to show a lot of genomes are still needed.

[ThaYamamoto]

I just used the sims on Genoplot, so they should be up to date.

I figured it out, using DMXX's NW SPGT derived sim is causing this massive inflation of Barcin. Sticking to the SiSBA3 derived sim and the moderns derived south sim fix this issue. Those EEF amounts were way too high for SA, something about the SGPT sim that messes things up.

edit: the gonur derived and sisBA2 sims work really well as well, particularly for me.

[ViktorL1]

Viktor_scaled,0.0774,-0.103584,0.006411,0.007752,-0.044316,0.013108,0.00611,0.002538,-0.02168,-0.010387,-0.027606,-0.002997,0.006244,-0.007294,0.0019,0.001193,-0.008084,0.002534,-0.001634,-0.002126,-0.003743,-0.004081,-0.008011,-0.00253,-0.001317

Viktor,0.0068,-0.0102,0.0017,0.0024,-0.0144,0.0047,0.0026,0.0011,-0.0106,-0.0057,-0.017,-0.002,0.0042,-0.0053,0.0014,0.0009,-0.0062,0.002,-0.0013,-0.0017,-0.003,-0.0033,-0.0065,-0.0021,-0.0011

Target: Viktor_scaled
Distance: 4.4979% / 0.04497900
27.0 Yamnaya_RUS_Samara
23.8 Han
12.2 Nganassan
11.8 Anatolia_Tepecik_Ciftlik_N
9.4 IRN_Ganj_Dareh_N
8.8 Kura-Araxes_ARM_Kaps
5.6 IRN_Shahr_I_Sokhta_BA2
1.4 BRA_LapaDoSanto_9600BP

Finally got my G25 coordinates/analysis. I am Tajik, all known ancestry from Tajikistan (Khujand region).

[pegasus]

Viktor_scaled,0.0774,-0.103584,0.006411,0.007752,-0.044316,0.013108,0.00611,0.002538,-0.02168,-0.010387,-0.027606,-0.002997,0.006244,-0.007294,0.0019,0.001193,-0.008084,0.002534,-0.001634,-0.002126,-0.003743,-0.004081,-0.008011,-0.00253,-0.001317

Viktor,0.0068,-0.0102,0.0017,0.0024,-0.0144,0.0047,0.0026,0.0011,-0.0106,-0.0057,-0.017,-0.002,0.0042,-0.0053,0.0014,0.0009,-0.0062,0.002,-0.0013,-0.0017,-0.003,-0.0033,-0.0065,-0.0021,-0.0011

Target: Viktor_scaled
Distance: 4.4979% / 0.04497900
27.0 Yamnaya_RUS_Samara
23.8 Han
12.2 Nganassan
11.8 Anatolia_Tepecik_Ciftlik_N
9.4 IRN_Ganj_Dareh_N
8.8 Kura-Araxes_ARM_Kaps
5.6 IRN_Shahr_I_Sokhta_BA2
1.4 BRA_LapaDoSanto_9600BP

Finally got my G25 coordinates/analysis. I am Tajik, all known ancestry from Tajikistan (Khujand region).

Your posting after a very long time. Nice to see you back
Even though your ethnically Tajik, genetically your pretty much similar to Uzbeks.


sample: khujand
distance: 1.8275
Sappali_Tepe_BA: 30.5
Sintashta_MLBA: 30.5
Late_Xiongnu_Han: 20
Slab_Grave_EIA_1: 19

[agent_lime]

Viktor_scaled,0.0774,-0.103584,0.006411,0.007752,-0.044316,0.013108,0.00611,0.002538,-0.02168,-0.010387,-0.027606,-0.002997,0.006244,-0.007294,0.0019,0.001193,-0.008084,0.002534,-0.001634,-0.002126,-0.003743,-0.004081,-0.008011,-0.00253,-0.001317

Viktor,0.0068,-0.0102,0.0017,0.0024,-0.0144,0.0047,0.0026,0.0011,-0.0106,-0.0057,-0.017,-0.002,0.0042,-0.0053,0.0014,0.0009,-0.0062,0.002,-0.0013,-0.0017,-0.003,-0.0033,-0.0065,-0.0021,-0.0011

Target: Viktor_scaled
Distance: 4.4979% / 0.04497900
27.0 Yamnaya_RUS_Samara
23.8 Han
12.2 Nganassan
11.8 Anatolia_Tepecik_Ciftlik_N
9.4 IRN_Ganj_Dareh_N
8.8 Kura-Araxes_ARM_Kaps
5.6 IRN_Shahr_I_Sokhta_BA2
1.4 BRA_LapaDoSanto_9600BP

Finally got my G25 coordinates/analysis. I am Tajik, all known ancestry from Tajikistan (Khujand region).

Do you have more Tajik results?

[ViktorL1]

I know! Glad to be back. Pegasus, is Sintashta the most plausible source for Steppe admixture in Southern Central Asians such as myself? And is Sappali Tepe a BMAC sample (mostly Iran_N, I’m guessing)? Thanks for providing that analysis.

I do know some other Tajiks (from other regions tho) who’ve been tested with 23andMe, could maybe persuade them to send their results to Eurogenes. Don’t have any other samples at the moment.

[agent_lime]

I know! Glad to be back. Pegasus, is Sintashta the most plausible source for Steppe admixture in Southern Central Asians such as myself? And is Sappali Tepe a BMAC sample (mostly Iran_N, I’m guessing)? Thanks for running that analysis.

I do know some other Tajiks (from other regions tho) who’ve been tested with 23andMe, could maybe persuade them to send their results to Eurogenes. Don’t have any other samples at the moment.

I know a few Tajiks as well. But they haven't shown interest in genetics. One of my good friends is a Pamiri Tajik. I'll ask her again.

[pegasus]

I know! Glad to be back. Pegasus, is Sintashta the most plausible source for Steppe admixture in Southern Central Asians such as myself? And is Sappali Tepe a BMAC sample (mostly Iran_N, I’m guessing)? Thanks for providing that analysis.

I do know some other Tajiks (from other regions tho) who’ve been tested with 23andMe, could maybe persuade them to send their results to Eurogenes. Don’t have any other samples at the moment.

Thats what the model preferred. Yes, Sappali is largely 65-70% Iran_N

[kamil154]

Don't you think that Simulated_AASI_by_DMXX inflates the AASI a little bit (3-4%)? This is what Kashmiri Pandit average scores on Genoplot (penalty off) with this particular model:

sample: Kashmiri Pandit:Average
distance: 1.8314
Ganj_Dareh_N: 31
Simulated_AASI_by_DMXX: 25.5
Sidelkino_HG: 9
Barcin_N: 20
Tyumen_HG: 14.5

penalty on:

sample: Kashmiri Pandit:Average
distance: 2.9794
Ganj_Dareh_N: 45
Simulated_AASI_by_DMXX: 23.5
Barcin_N: 15
Tyumen_HG: 8.5
Sidelkino_HG: 8

On other models the AASI for KPs limits itself to 20-21%. For me it has typically been at 24-26%. Or am I reading too much into it?

Some other groups (penalty on):

sample: Brahmin Gujarat:Average
distance: 2.1267
Ganj_Dareh_N: 40.5
Simulated_AASI_by_DMXX: 30
Barcin_N: 14
Tyumen_HG: 9.5
Sidelkino_HG: 6

sample: Bihar Brahmin Bhumihar:mean
distance: 3.1936
Ganj_Dareh_N: 36.5
Simulated_AASI_by_DMXX: 30
Barcin_N: 15
Tyumen_HG: 13
Sidelkino_HG: 5.5

sample: Brahmin Uttar Pradesh:Average
distance: 3.2536
Ganj_Dareh_N: 35.5
Simulated_AASI_by_DMXX: 34.5
Barcin_N: 12
Tyumen_HG: 10.5
Sidelkino_HG: 7.5

Well in my experience, AASI NW interacts a lot less with some of the other components and gives better fits for NW SA pops than AASI DMXX and so I prefer it. That being said using AASI NW only in tandem with Hoabinhian is a good idea IMO. I think excess ASE affinity is showing up as Hoabinhian on G25, its not meant to be taken as a literal ancestral vector. The two added up behave like just AASI DMXX on its own, but offer improved fit over just AASI DMXX.