Y-chromosomal analysis of clan structure of Kalmyks, the only European Mongol people, and their relationship to Oirat-Mongols of Inner Asia

TL;DR

This study uses Y-chromosome genotyping and sequencing to compare Kalmyks, the only Mongolic-speaking population in Europe, with Oirat-Mongols and neighboring groups. It finds that Kalmyks and Mongolian Oirats share a dominant paternal lineage, C3c1-M77, with a Kalmyk/Oirat-enriched sub-clade C3c1b-F6379 dating to ~1.5 KYA, predating the Kalmyk–Oirat split ~400 years ago.

Key findings

• Among 454 unrelated males, >50% of all studied samples belong to haplogroup C3; excluding Turkic-speaking Tsaatans/Tozhu Tuvans and Oirat-speaking Sart Kalmaks, C3 reaches 62.5% among Kalmyk and Mongolian Oirat subpopulations. The novel C3c1b-F6379 sub-clade comprises 40.3% of male lineages among Kalmyks and Oirats from Mongolia and occurs at minor frequencies among Mongol Tsaatans and Sart Kalmaks.
• A phylogeny of 37 high-coverage Y-chromosome sequences identifies C3c1b-F6379 with a TMRCA of ~1.5 KYA (95% HPD 1.1–2.1 KYA), before the Genghis Khan era and earlier than the Kalmyk–other Oirat split ~400 years ago. C3c1-M77 is also present in several geographically distant native Siberian populations.
• Kalmyk sub-groups and Western Mongolian Oirats cluster tightly in PCA; AMOVA shows differentiation among studied groups (FCT = 0.15670, P = 0.00293), while differentiation among geographically distant groups is low and insignificant (FCT = 0.01771, P = 0.13196). Sart Kalmaks show >50% hg R and ~13% hg C3c; Mongolian Tsaatans have ~96% N3a5a-F4205 and Q1a1b-M25; Tozhu Tuvans have N3a5a-F4205 at 15.2% and Q1a1b-M25 at 50.0%.
• Putative Khoshut rulers with possible genealogical links to Genghis Khan’s brother belong to haplogroup C3-M77. The STR network central node expands at 667 ± 155 years, while sequence-based TMRCA is 1076–2011 years; the putative Genghis Khan descendant haplotype is two mutational steps from the central node.

Methods

• Studied 454 unrelated male individuals: Kalmykia Torgut (58), Derbet (69), Buzava (52), Khoshut (28); Western Mongolia Torgut (47), Derbet (40), Khoshut (18), Tsaatan (23); Sart Kalmaks from Kyrgyzstan (61); Kalmyk students from Xinjiang, China (12); and Tozhu Tuvans from Russia (46). Genotyped Y-chromosomal biallelic markers, including newly designed markers F6379, B80, and B469.
• Microsatellite analysis used 23 STRs for 78 samples; a median-joining network for hg C3c1 used 17 Y-STRs from 50 hg C3c1-M77 samples (this study plus published data), with expansion time calculated by rho-statistic and a mutation rate of ~8.6 × 10−5 per locus per year.
• Whole Y-chromosome sequencing combined 10 novel and 28 published sequences; novel samples were sequenced on Illumina HiSeq 2500 with BigY capture. Variants were mapped/called with BWA-MEM, GATK, and bcftools, yielding 9.2 Mb usable sequence; phylogeny used BEAST v1.7.5 with hg C7a outgroup and hg C calibration age 50,865 years (95% CI 49,191–52,699). PCA, correspondence analysis, Fst, and AMOVA were also performed.

Limitations

• Stated: TMRCA estimates differ between STR network and sequence/BEAST analyses because of different methods and datasets; microsatellite mutation processes are less well understood, and repeat length changes can occur bidirectionally, potentially underestimating times. One newly sequenced sample was excluded from BEAST due to quality issues.
• Stated: reliable patrilineal identification of the Genghis Khan clan requires ancient DNA from archaeologically authentic remains belonging to the imperial family, which have not yet been found.
• Evident: some subgroup sample sizes are small (e.g., Khoshut 18/28, Tsaatan 23, Xinjiang 12), and the study relies on modern Y-chromosome data only; the Figure 1a caption also notes Xinjiang Kalmyk students were left out due to small sample size (<15), despite 12 being listed in the methods.