[Split] New Yazidi Kurd Data (+ Iranian Azeri, Other Kurd & Lor Analysis)
With barely any new relevant samples added, these two runs are still my best fits using BA-IA averages in the spreadsheet.
distance%=1.7727"
Helves_scaled
Levant_Canaanite_MBA,41.6
IRN_Hajji_Firuz_IA,35.4
Kura-Araxes_ARM_Kaps,23
"distance%=1.6887"
Helves_scaled
Levant_Canaanite_MBA,39
IRN_Hajji_Firuz_IA,24
Kura-Araxes_ARM_Kaps,20.6
IRN_Hasanlu_IA,16.4
But I'm starting to think that we might not get aDNA from Syria or Iraq in years.
Agamemnon (07-15-2019), Dewsloth (07-15-2019), DMXX (07-15-2019), khanabadoshi (07-16-2020), Seabass (07-02-2020)
Interesting though that the Yezidi average scores more Turkish/East Eurasian related than the Kurdish one. I would've guessed it to be the other way around.Originally Posted by StarDS9
"distance%=1.5456"
Kurdish
IRN_Hajji_Firuz_C,71
TKM_IA,25.4
Anatolia_Ottoman,3.6
"distance%=1.8217"
Ezid
IRN_Hajji_Firuz_C,67.2
TKM_IA,27.8
Anatolia_Ottoman,5
I've run an iteration of Helves' model (with the addition of the Dinka) for all individual W/NW Iranian populations in G25 (excluding Iranian Azeris, who've been removed, quite unfortunately). I'm the closest representative to them right now (over half).
None of the Yezidis surpass StarDS9, Turkmenistan_IA-wise (he, I and the Iranian user 'NK19191' are in the top 2-5% for steppe-related admixture in NW/N Iranian-related peoples). They also don't look much different to other Kurdish groups with respect to the other components.Code:Sample Details Fit Map Anatolia Ottoman--MA2195 Dinka IRN Hajji Firuz C TKM IA 1 Custom:AGUser_DMXX 3.0487 Open Map 7.8 0.8 62.2 29.2 2 Custom:AGUser_Mori_yek 2.1435 Open Map 4.8 0.8 76.8 17.6 3 Custom:AGUser_StarDS9 2.6264 Open Map 5 0 67.2 27.8 4 IRN_Hasanlu_IA:F38 2.6882 Open Map 1 0 84.4 14.6 5 Iranian_Lor:LORII19 Modern; 2.58 Open Map 6.2 1.2 77 15.6 6 Iranian_Lor:LORII34 Modern; 3.2072 Open Map 5 1.4 79.2 14.4 7 Iranian_Lor:LORII36 Modern; 3.3557 Open Map 0 1.6 80.6 17.8 8 Iranian_Lor:LORII44 Modern; 3.3644 Open Map 4.6 1.2 76.4 17.8 9 Iranian_Lor:LORII48 Modern; 2.7461 Open Map 6 0 76 18 10 Iranian_Lor:LORII50 Modern; 2.8128 Open Map 6 1 77.4 15.6 11 Iranian_Lor:LORII51 Modern; 2.5972 Open Map 4.4 1.8 82 11.8 12 Iranian_Lor:LORII52 Modern; 2.885 Open Map 2.4 0.2 75.8 21.6 13 Iranian_Lor:LORII58 Modern; 3.082 Open Map 5 0 77.6 17.4 14 Iranian_Lor:LORII9 Modern; 2.1486 Open Map 8.6 1 79.4 11 25 Kurdish:KRD-14 2.7405 Open Map 2.8 0 83.8 13.4 26 Kurdish:KRD-31 3.149 Open Map 3 0 78.4 18.6 27 Kurdish:KRD-40 2.232 Open Map 6.4 0 76.8 16.8 28 Kurdish:KRD-43 2.9044 Open Map 3 0.4 79.4 17.2 29 Kurdish:KRD-45 2.9222 Open Map 1.4 0.8 72 25.8 30 Kurdish:KRD-49 3.4237 Open Map 2 0 75.6 22.4 31 Kurdish:KRD-5 2.5963 Open Map 3.6 0 78.8 17.6 32 Kurdish:KRD-8 2.9149 Open Map 2.8 0.6 80.6 16 33 Kurdish:kurd1101 Modern; 2.0978 Open Map 3.4 0.2 78 18.4 34 Kurdish:kurd1156 Modern; 2.3603 Open Map 4.4 0 86 9.6 35 Kurdish:kurd1159 Modern; 2.3393 Open Map 2.8 0 73.6 23.6 36 Ezid:EZI-001 3.1904 Open Map 1.8 0.2 72.4 25.6 37 Ezid:EZI-010 2.2194 Open Map 5.2 0.2 68.8 25.8 38 Ezid:EZI-012 2.9591 Open Map 4.6 1 71.8 22.6 39 Ezid:EZI-013 3.0367 Open Map 2.6 0 77.8 19.6 40 Ezid:EZI-020 3.1023 Open Map 4.8 0 68 27.2 41 Ezid:EZI-029 3.248 Open Map 5.6 2 78.8 13.6 42 Ezid:EZI-060 2.9524 Open Map 3.6 0.4 81.4 14.6 43 Ezid:EZI-074 3.4684 Open Map 5.4 0.4 77 17.2
As we already knew, Kurmanji Kurds, Lors and Iranian Azeris are basically MLBA steppe and Medieval Turkish-shifted variations of a Hasanlu_IA type population (I personally think this was an early Median). We can safely add Yezidis to that list, now. The Turkmenistan_IA admixture differences between Lurs vs. Kurmanjis + Azeris + Yezidis might represent additional Parthian admixture in us relative to them. IMO, that would support the idea that "we" are predominantly Median-derived, with variable amounts of additional TRK_IA (?Parthian) and Medieval Turkish.
I'll calculate the means for these shortly (will list the "KRD" and "kurd" averages separately, as Star indicated to us months ago that the latter group are Anatolian-shifted Turkish Kurds).
I re-ran the model again, with the addition of my AASI_NW simulation. I realised that some of the Medieval Turkish scores were inflated (particularly for Lors). The majority of the fits improved.
The averages:Code:Sample Details Fit Map Anatolia Ottoman--MA2195 Dinka IRN Hajji Firuz C Simulated AASI NW By DMXX TKM IA 1 Custom:AGUser_DMXX 2.7087 Open Map 5.8 0.8 64 2.8 26.6 2 Custom:AGUser_Mori_yek 2.0318 Open Map 3.6 0.6 77.4 1.8 16.6 3 Custom:AGUser_StarDS9 2.5547 Open Map 3.6 0 69.2 1.6 25.6 4 Ezid:EZI-001 3.1753 Open Map 1.8 0.2 73.4 1 23.6 5 Ezid:EZI-010 2.3619 Open Map 2.6 0.2 69.8 0.6 26.8 6 Ezid:EZI-012 2.9742 Open Map 2.6 1 72.8 1 22.6 7 Ezid:EZI-013 2.8617 Open Map 1.2 0 79.2 1.4 18.2 8 Ezid:EZI-020 2.6099 Open Map 2 0 70.2 3 24.8 9 Ezid:EZI-029 3.0945 Open Map 3 1.8 81.8 2.6 10.8 10 Ezid:EZI-060 2.6866 Open Map 2.4 0.2 83 2 12.4 11 Ezid:EZI-074 3.118 Open Map 2.2 0.4 77.8 2.2 17.4 12 IRN_Hasanlu_IA:Average 2.8876 Open Map 1.6 0 87 0.8 10.6 13 Iranian_Lor:LORII19 Modern; 2.326 Open Map 3.8 1.2 79.6 3 12.4 14 Iranian_Lor:LORII34 Modern; 2.9853 Open Map 3.2 1.2 79.8 2.2 13.6 15 Iranian_Lor:LORII36 Modern; 3.4889 Open Map 0 1 84.6 1.2 13.2 16 Iranian_Lor:LORII44 Modern; 2.9632 Open Map 3.4 1 78.2 2.8 14.6 17 Iranian_Lor:LORII48 Modern; 2.6995 Open Map 2.8 0 76.4 1.4 19.4 18 Iranian_Lor:LORII50 Modern; 2.3388 Open Map 3.6 0.6 78.8 3 14 19 Iranian_Lor:LORII51 Modern; 2.3501 Open Map 2.2 1.6 82.2 2 12 20 Iranian_Lor:LORII52 Modern; 2.821 Open Map 1.2 0.4 76.8 1.2 20.4 21 Iranian_Lor:LORII58 Modern; 3.1582 Open Map 3.8 0 78 0.4 17.8 22 Iranian_Lor:LORII9 Modern; 2.3132 Open Map 5.4 1 82 2 9.6 23 Kurdish:KRD-14 2.5446 Open Map 1.8 0 85 1.8 11.4 24 Kurdish:KRD-31 3.3255 Open Map 2.2 0 82.4 0.6 14.8 25 Kurdish:KRD-40 2.5066 Open Map 4.4 0 79.8 1.2 14.6 26 Kurdish:KRD-43 2.963 Open Map 1.8 0.4 82.2 1.2 14.4 27 Kurdish:KRD-45 2.9515 Open Map 1.4 0.6 72.6 0.2 25.2 28 Kurdish:KRD-49 3.3333 Open Map 1.6 0 75.8 1.2 21.4 29 Kurdish:KRD-5 2.799 Open Map 2.6 0 83.2 1.6 12.6 30 Kurdish:KRD-8 3.1965 Open Map 2.8 0.6 85 1 10.6 31 Kurdish:kurd1101 Modern; 2.0664 Open Map 1.6 0.2 79.2 1.4 17.6 32 Kurdish:kurd1156 Modern; 2.4298 Open Map 2.4 0 87.6 0.6 9.4 33 Kurdish:kurd1159 Modern; 2.4273 Open Map 2.6 0 74.8 0.4 22.2
It looks like the omission of AASI_NW caused some of the non-Yazidi Kurds to experience a slight elevation in their Turkmenistan_IA and Medieval Turkish scores.Code:Pop Anatolian_Ottoman Dinka HF_C AASI TKM_IA EZID 2.2 0.5 76.0 1.7 19.6 LOR 2.9 0.8 79.6 1.9 14.7 KRD 2.3 0.2 80.8 1.1 15.6 kurd 2.2 0.1 80.5 0.8 16.4
This run corroborates Helves' observation - Yazidis are about 15-20% more Turkmenistan_IA shifted in comparison to Kurds from Iraq and Turkey. One of them (EZ-10) is in the same "maxed out" strata as myself and Star (25-26%).
There also does appear to be some Medieval Turkish ancestry, which appears to fairly regularly hover between 2-3%. Yezidis resemble Lors when it comes to AASI-derived admixture (it's slightly more than what's observed in other Kurdish groups).
I've just calculated the non-Yezidi Kurdish stats for Turkmenistan_IA admixture - Averages around 16.7%, SD ~5.42. That puts StarDS9 in the top 4-5% of non-Yezidi Kurds. The total NW/W Iranian-related average (all pops above included exluding Hasanlu_IA and myself) is basically the same (16.8%, SD ~5.1).
To be honest, Iranian Azeris probably wouldn't change the Turkmenistan_IA pan-regional average. From memory, their distribution for steppe-related admixture (excluding East Eurasian admix) was in the same range as Kurmanji Kurds. The main difference lay in East Eurasian-related admixture, and a subsequent depression in prehistoric Iranian plateau admix (you can see that reflected in my own scores - I'm less Hajji Firuz_ChL than everyone else).
Last edited by DMXX; 07-15-2019 at 04:21 PM. Reason: correction
Wonder why our runs give so different results using the same pops? I use scaled coordinates and pen=0.
That's why. I don't use pen=0. The fits are "better" but the results are noisier, even if they appear more conveniently in line with expectations.
The member who created the nMonte script (Huij), a mathematician, doesn't recommend using it. Even if I hadn't done the background reading last year, I'd have deferred to the expert/tool creator's guidance nonetheless.
Last edited by DMXX; 07-15-2019 at 06:45 PM. Reason: clarification
IIRC (need to dig through to find that post) Huij also recommended what we call the "unscaled" sheet since the eigenvals used to re-create the "scaled" sheet could be problematic. I think Pen=0.001 with unscaled was his recommendation. Not sure what he thinks of pen with the scaled data.
DMXX (07-16-2019)
Ezid's I think have had less outside influence then other Kurmanjis, which is why they are more east shifted compared to other Kurmanjis. Ezidis would probably be similar to Soran and Southern Kurds, as on Gedmatch calculators their scores resembled them.
Out of all the Kurdish groups, I would guess the Goran and Feyli to be most TM_IA shifted as their languages are believed to be offshoot of Parthian. So far as I can tell all the Kurds are Kurmanjis on Globe 25.
I wonder if Poi can re-add the Azeri_Iran back on the Globe25?
DMXX (07-15-2019)
I will do that today. I have a commit history of all changes, including data changes, so I can just re-import the data as custom groups. So, in effect, Azeri_Iran becomes a non-Davidski group. We have numerous examples of those, for example of my own ethnic group, various Kerala groups, etc. That should do the trick. ETA will be likely in a few hours.
DMXX (07-16-2019)
I've just calculated the results for the Iranian Azeris. The updated list is as follows:
The group averages:Code:Sample Details Fit Map Anatolia Ottoman--MA2195 Dinka IRN Hajji Firuz C Simulated AASI NW By DMXX TKM IA 1 Custom:AGUser_DMXX 2.7087 Open Map 5.8 0.8 64 2.8 26.6 2 Custom:AGUser_Mori_yek 2.0318 Open Map 3.6 0.6 77.4 1.8 16.6 3 Custom:AGUser_StarDS9 2.5547 Open Map 3.6 0 69.2 1.6 25.6 4 Ezid:EZI-001 3.1753 Open Map 1.8 0.2 73.4 1 23.6 5 Ezid:EZI-010 2.3619 Open Map 2.6 0.2 69.8 0.6 26.8 6 Ezid:EZI-012 2.9742 Open Map 2.6 1 72.8 1 22.6 7 Ezid:EZI-013 2.8617 Open Map 1.2 0 79.2 1.4 18.2 8 Ezid:EZI-020 2.6099 Open Map 2 0 70.2 3 24.8 9 Ezid:EZI-029 3.0945 Open Map 3 1.8 81.8 2.6 10.8 10 Ezid:EZI-060 2.6866 Open Map 2.4 0.2 83 2 12.4 11 Ezid:EZI-074 3.118 Open Map 2.2 0.4 77.8 2.2 17.4 12 IRN_Hasanlu_IA:Average 2.8876 Open Map 1.6 0 87 0.8 10.6 13 Iranian_Lor:LORII19 Modern; 2.326 Open Map 3.8 1.2 79.6 3 12.4 14 Iranian_Lor:LORII34 Modern; 2.9853 Open Map 3.2 1.2 79.8 2.2 13.6 15 Iranian_Lor:LORII36 Modern; 3.4889 Open Map 0 1 84.6 1.2 13.2 16 Iranian_Lor:LORII44 Modern; 2.9632 Open Map 3.4 1 78.2 2.8 14.6 17 Iranian_Lor:LORII48 Modern; 2.6995 Open Map 2.8 0 76.4 1.4 19.4 18 Iranian_Lor:LORII50 Modern; 2.3388 Open Map 3.6 0.6 78.8 3 14 19 Iranian_Lor:LORII51 Modern; 2.3501 Open Map 2.2 1.6 82.2 2 12 20 Iranian_Lor:LORII52 Modern; 2.821 Open Map 1.2 0.4 76.8 1.2 20.4 21 Iranian_Lor:LORII58 Modern; 3.1582 Open Map 3.8 0 78 0.4 17.8 22 Iranian_Lor:LORII9 Modern; 2.3132 Open Map 5.4 1 82 2 9.6 23 Kurdish:KRD-14 2.5446 Open Map 1.8 0 85 1.8 11.4 24 Kurdish:KRD-31 3.3255 Open Map 2.2 0 82.4 0.6 14.8 25 Kurdish:KRD-40 2.5066 Open Map 4.4 0 79.8 1.2 14.6 26 Kurdish:KRD-43 2.963 Open Map 1.8 0.4 82.2 1.2 14.4 27 Kurdish:KRD-45 2.9515 Open Map 1.4 0.6 72.6 0.2 25.2 28 Kurdish:KRD-49 3.3333 Open Map 1.6 0 75.8 1.2 21.4 29 Kurdish:KRD-5 2.799 Open Map 2.6 0 83.2 1.6 12.6 30 Kurdish:KRD-8 3.1965 Open Map 2.8 0.6 85 1 10.6 31 Kurdish:kurd1101 Modern; 2.0664 Open Map 1.6 0.2 79.2 1.4 17.6 32 Kurdish:kurd1156 Modern; 2.4298 Open Map 2.4 0 87.6 0.6 9.4 33 Kurdish:kurd1159 Modern; 2.4273 Open Map 2.6 0 74.8 0.4 22.2 1 Azeri_Iran:azerB38 Modern; 2.7471 Open Map 5.6 0.2 74.6 1.8 17.8 2 Azeri_Iran:azerB59 Modern; 3.3505 Open Map 8.8 0 74.8 2.2 14.2 3 Azeri_Iran:azerB61 Modern; 2.324 Open Map 6.2 0 75.8 1.4 16.6 4 Azeri_Iran:azerB64 Modern; 2.3044 Open Map 8.6 0 74 1.2 16.2 5 Azeri_Iran:azerB8 Modern; 3.1697 Open Map 6.4 0 78 2.2 13.4 6 Azeri_Iran:azerE1 Modern; 2.5158 Open Map 5.8 0.4 71.2 1 21.6 7 Azeri_Iran:azerE3 Modern; 2.5408 Open Map 4.2 0.2 78.8 1.2 15.6 8 Azeri_Iran:azerE6 Modern; 2.6718 Open Map 9 0 72.8 1.6 16.6 9 Azeri_Iran:azerE70 Modern; 2.929 Open Map 7 0.2 79.6 1.8 11.4 10 Azeri_Iran:azerE82 2.4317 Open Map 7.8 1 71 2 18.2 11 Azeri_Iran:azerE85 2.7717 Open Map 8 0.2 76.6 1.2 14 12 Azeri_Iran:azerE89 2.5121 Open Map 8 0 68.2 1.8 22 13 Azeri_Iran:azerE92 2.7473 Open Map 7.4 0.6 74.8 1.2 16 14 Azeri_Iran:azerF111 3.372 Open Map 6.6 0.4 63.6 1.2 28.2 15 Azeri_Iran:azerF4 2.4252 Open Map 6.8 0.6 73.2 1.8 17.6 16 Azeri_Iran:azerb72 Modern; 2.0873 Open Map 8 0.4 73.2 3.2 15.2
Iranian Azeris are slightly more Turkmenistan_IA shifted than non-Yazidi Kurds and Lors here, though they're behind Yazidis.Code:Group Sample Size Anatolian_Ottoman(MA2195) Dinka Hajji_Firuz_ChL AASI Turkmenistan_IA Azeri (Iran) 16 7.1 0.3 73.78 1.7 17.2 Kurdish (non-Yazidi, ?Turkey-Iraq) 11 2.3 0.2 80.7 1 15.8 Kurdish (Yazidi, ?) 9 2.2 0.5 76 1.7 19.6 Lor (Iran) 10 2.9 0.8 79.6 1.9 14.7
Posting Permissions
Reply With Quote