Elevated Eurasian Steppe Ancestry in Kurds

Here I investigate how much more genetic drift Kurds share with various ancient Eurasian Steppe samples, in comparison with some of their geographic neighbors. I do this formally using D-Statistics (Dstats) from ADMIXTOOLS¹, a software bundle, readily available from the Reich Labs website at Harvard University. ADMIXTOOLS has been fully described in Patterson et al, 2012².

The output of qpDstat is informative about the direction of gene flow. So for 4 populations (W, X, Y, Z) as follows:

If the Z-score is +ve, then the gene flow occured either between W and Y or X and Z; If the Z-score is -ve, then the gene flow occured either between W and Z or X and Y.

In the following tests, I examine D ( W/SC/S Asian, Iraqi Kurd, Steppe sample, Outgroup). For Eurasians, using a Eurasian ascertained panel, and an outgroup, such as Chimp, I can rule out gene flow between W/SC/S Asian & Chimp (W & Z), as well as between Iraqi Kurd & Chimp (X & Z), for Eurasians. Therefore, a -ve D in this case can be interpreted as greater overall geneflow, Iraqi Kurd <—> Steppe sample than geneflow W/SC/S Asian <—> Steppe sample, and conversely, a +ve D can be interpreted as greater geneflow, W/SC/S Asian <—> Steppe sample than geneflow, Iraqi Kurd <—> Steppe sample.

Thus far, I had not explored the source of this excess steppe geneflow (compared to their neighbors) to Kurds, because I had assumed that this steppe geneflow had filtered down to Kurds from some of their northern neighbors such as Georgians and Armenians. This was a reasonable assumption, because groups such as Georgians form a buffer population between Kurdistan and the Pontic Caspian Steppe, where ancients associated with the Yamnaya Steppe pastorists used to roam. Apparently, I was mistaken, as my analysis using Dstats showed that the Iraqi Kurd samples had more steppe ancestry than their northern Georgian, Armenian, and Turk neighbors! So how can we explain this….To me the most reasonable explanation would be that Kurds received a dose of steppe geneflow relatively recently via some steppe admixed groups which had settled in Kurdistan, such as Parthians, Sarmatians, Saka, Alans, Scythians and Mitanni.

The other surprise is that Dstats show Kurds as sharing a greater amount of derived alleles with Eurasian Steppe groups in spite of my Kurd samples having greater admixture from outside the “Steppe’ clade, such as African, SW Asian and S Asian, than their Georgian and Armenian neighbors. Many will wonder how this is significant….Well, any African, SW Asian or S Asian admixture in excess of that carried by the Steppe populations, such as in my Kurd samples will tend to actually decrease shared drift between Kurds and Steppe populations. Thus since my Kurd samples carry more African, SW Asian, and S Asian admixture than Georgians and Armenians, the Dstat output, should have theoretically shown Georgians and Armenians with greater steppe gene flow than Kurds . However, we see the opposite.

Therefore, to account for the steppe admixture levels in the Kurd samples, input into Kurds is necessary from a population with significantly greater steppe ancestry than Georgians, Armenians, and Turks have to offer. Thus, I suspect groups such as Parthians, Sarmatians, Saka, Alans, Scythians and Mitanni, which had historically settled in Kurdistan had something to do with this.

The following are some definitions:

Iranian – Average of 35 Iranian samples from Mazendaran, Loristan, and Shiraz.

Turkish – Average of 75 Turk samples from Adana, Aydin, Balekisir, Istanbul, Kayseri, and Trabzon.

Kurd C3 – A Kurdish sample from N Iraq.

Yamnaya_S = Yamnaya Samara,
Yamnaya_K = Yamnaya Kalmykia

Steppe samples are defined as follows:

EHG_61	I0061
Yamnaya_S2	I0357
Yamnaya_S3	I0370
Poltavka1	I0371
Poltavka2	I0374
Yamnaya_S4	I0429
Yamnaya_S5	I0438
Poltavka4	I0440
Yamnaya_S6	I0443
Yamnaya_S7	I0444
Afansievo	RISE509
Afansievo	RISE511
Yamnaya_K2	RISE548
Steppe_Eneolithic	I0122
Scythian_IA	I0247
Steppe_MLBA	I0232
Steppe_MLBA	I0359
Steppe_MLBA	I0422
Steppe_MLBA	I0423
Steppe_MLBA	I0424
Steppe_MLBA	I0430
Steppe_MLBA	I0432
Sintashta	RISE395
Andronovo1	RISE500
Andronovo2	RISE503
Andronovo3	RISE505

Throughout, a -ve D value indicates that Iraqi Kurd (Kurd C3) shares more genetic drift with the particular steppe sample than Pop1, which most likely translates to more steppe ancestry. Conversely, a +ve D value indicates that the Pop1 sample shares more genetic drift with the steppe sample than Iraqi Kurd. The tables are sorted with the samples that Iraqi Kurd shares the most drift with on top.

Shared Genetic Drift with Steppe Populations; Iraqi Kurd vs Armenians

The following table shows that Iraqi Kurd shares more drift than Armenians with the majority of the steppe samples (-ve Ds), especially with the Khvalynsk Eneolithic and Scythian IA samples. This translates to most likely more steppe ancestry for Iraqi Kurd than Armenians. There are only 2 samples with which Armenians share more drift with than Iraqi Kurd; the Mesolithic Kotias Hunter Gatherer sample and Yamnaya S6. It should be noted here that Z values under absolute 1 are considered insignificant at this SNP level.

Table 1 – D-stats of the form D [Armenians, Kurd-C3; Steppe cultures, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with all steppe cultures than Armenians, and significantly more drift with Steppe-EMBA, Scythians, Steppe-MLBA, and EHG than Armenians share with the aforementioned.

Shared Genetic Drift with Steppe Populations; Iraqi Kurd vs Georgians

As was the case with Armenians, but to a lesser degree, the following table shows that Iraqi Kurd shares more drift than Georgians with the majority of the steppe samples (-ve Ds), especially with the Andronovo 2, Yamnaya S7 and Scythian IA, and EHG samples. This translates to most likely more steppe ancestry for Iraqi Kurd than Georgians. Georgians share more drift than Iraqi Kurd with only are only 2 CHG samples, and 3 steppe samples.

**Table 2 –** D-stats of the form D [Georgians, Kurd-C3; Steppe cultures, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with all steppe cultures than Georgians, and significantly more drift with most Steppe-EMBA, Scythians, Steppe-MLBA, and EHG than Georgians share with the aforementioned.

Shared Genetic Drift with Steppe Populations; Iraqi Kurd vs Iranian Zoroastrians

Here again we see that Iraqi Kurd shares more drift with almost all the steppe samples (-ve D) than Zoroastrians, especially with the Scythian IA, Yamnaya S7, Andronovo, and 13K year old Satsurbila CHG samples.

**Table 3 –** D-stats of the form D [Zoroastrians, Kurd-C3; Steppe cultures, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with all steppe cultures than Iranian Zoroastrians, especially with the Iron Age Volga river Scythian sample and Middle to Late Bronze Age proto Indo-Iranian Andronovo and Sintashta cultures than Iranian Zoroastrians share with the aforementioned.

Shared Genetic Drift with Steppe Populations; Iraqi Kurd vs Turks

Again we see that Iraqi Kurd shares more drift with almost all the steppe samples (-ve D) than Turks, especially with the Scythian IA, 13K year old Satsurbila CHG, and Khvalynsk Eneolithic samples.

**Table 4 –** D-stats of the form D [X , Kurd-C3; Steppe Eneolithic, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with the Khvalynsk Eneolithic steppe samples than most modern and ancient West Asians. The significantly greater drift Kurd-C3 shares with Khvalynsk Eneolithic cultures than Neolithic and Bronze Age West Asian samples share with those indicate a significant amount of steppe admixture into Kurds since the Neolithic and Bronze Ages.

Kurds may have received some steppe gene flow via Scythians

The Western Scythian IA sample seems to consistently show up at the top of the table for Iraqi Kurd, which is probably not a coincidence considering Scythians ruled parts of Kurdistan. To see whether Dstats could corroborate this, I ran D (Asians/Ancients, Kurd, Scythian, Outgroup). The results showed that Iraqi Kurd shared more genetic drift with Scythian IA than virtually all W/SC/S Asians and Ancients. This indicates a significant amount of admixture into Kurds from Western Scythians or the related Sarmatians and Alans. This is consistent with the Kurdish language as well as Farsi being one of the closest related extant languages to some Saka languages.

I plan to follow-up with haplotype comparison tests for shared IBD DNA segments using Beagle later, to see whether the patterns seen here can be corroborated.

**Table 5 –** D-stats of the form D [X, Kurd-C3; Scythian-IA, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with the Volga river Iron Age Scythian sample than all the shown modern West and SC Asian populations, as well as the ancient West Asian samples. This indicates a significant amount of admixture into Kurds from Iron Age western Scythians, Sarmatians, and Alans.

**Table 6 –** D-stats of the form D [X, Kurd-C3; Andronovo-MLBA, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with the MLBA Andronovo culture than most of the shown modern West and SC Asian populations, as well as the ancient West Asian samples. This indicates a significant amount of admixture into Kurds from the Middle to Late Bronze Age proto Indo-Iranian Andronovo culture. Table 5 indicates that this admixture was likely mediated into Kurds by the Iron Age western Scythians, Sarmatians, and Alans.

**Table 7 –** D-stats of the form D [Anatolian Turks, Kurd-C3; Steppe cultures, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with steppe cultures than Turks, and significantly greater drift with the Volga River Western Scythian sample than Turks share with the aforementioned.

**Table 7 –** D-stats of the form D [X, Kurd-C3; EHG, Chimp] show that the Kurmanji Kurd sample C3 shares greater genetic drift with EHG than many modern West Asian populations, and significantly greater drift with the EHG sample than its ancestral West Asian populations. Shared drift with EHG by modern West Asians is a good predictor of steppe admixture into those populations since most Bronze Age and later steppe populations were significantly EHG admixed to the exclusion of ancient West Asians.

Professor Izady on the Sindhi Kurds of Iraq and the Mittani

Professor Izady who has taught at Columbia and Harvard states the following³:

“The name ‘Mittani’ survives today in the Kurdish clans of Mattini and Millani/Milli who inhabit the exact same geographical areas of Kurdistan as the ancient Mittani. The name “Mittan,” however, is a Hurrian name rather than Aryan. At the onset of Aryan immigration into Kurdistan, only the aristocracy of the high-ranking warrior groups were Aryans, while the bulk of the people were still Hurrian in all manners. The Mittani aristocratic house almost certainly was from the immigrant Sindis, who survive today in the populous Kurdish clan of Sindi—again—in the same area where the Mittani kingdom once existed. These ancient Sindi seem to have been an Indic, and not Iranic group of people, and in fact a branch of the better known Sindis of India-Pakistan, that has imparted its name to the River Indus and in fact, India itself. (footnote 8) While the bulk of the Sindis moved on to India, some wondered into Kurdistan to give rise to the Mittani royal house and the modern Sindi Kurds.Expectedly, the Mittani pantheon includes names like Indra, Varuna, Suriya and Nasatya is typically Indic. The Mittanis could have introduced during this early period some of the Indic/Vedic tradition that appears to be manifest in the Kurdish religion of Yazdanism.The avalanche of the Indo-European tribes, however, was to come about 1200 BC, raining havoc on the economy and settled culture in the mountains and lowlands alike. The north was settled by the Haigs who are known to us now as the Armenians, while the rest of the mountains became targets of settlement of various Iranic peoples, such as the Medes, Persian, Scythians, Sarmatians, and Sagarthians (whose name survives in the name of the Zagros mountains).By 850 BC, the last Hurrian states had been extinguished by the invading Aryans, whose sheer number of immigrants must have been considerable. These succeeded over time to change the Hurrian language(s) of the people in Kurdistan, as well as their genetic make-up. By about the 3rd century BC, the Aryanization of the mountain communities was virtually complete.Since the star of the Mittani shown brightest in 1500 BC, Aryan dynasties of various size and influence continued their appearance in various corners of Kurdistan. None, however, was to match, and in fact surpass the Mittanis as the Medians. The rise of the Medes from their capital at Ecbatana (modern Hamadan) in 727 coincided with the fall of the last major Hurrian kingdom: the Mannaeans. Ignoring the proud legacy of the Hurrian states and even the Aryan empire of the Mittanis which can squarely be claimed by on every ground by modern Kurds, it is the Medes that the Kurds have grown most fund of. Medes are claimed regularly by the Kurds and pronounced by others to be the ancestors of them. This is strange, when realizing how many millennia of cultural and ethnic evolution preceded the rise of Medes into Kurdistan. In reality, Medes are no more the ancestors of the modern Kurds as all other Halafian, Hurrian and Mittani who came before them or the legion of other peoples and states that came after them. Nonetheless, today, even the first Kurdish satellite television transmitter is given the name “Med TV” (Kurdish for “Median TV”). Fascination of the Kurds with the Median Federation (a.k.a., Empire) that ended in 549 BC remains supreme, indeed.”

References:

1- Admixtools

2- Patterson et al, 2012

3- Kurdistanica.com