Volga Tatars

new

Admin
Administrator

Posts: 73,018

Volga Tatars Oct 21, 2023 5:23:45 GMT

Quote

Post by Admin on Oct 21, 2023 5:23:45 GMT

Results
Population structure
In our study, collection points spanned the latitudes from the Northern Europe to Caucasus and longitudes from Eastern Europe to far shores of Siberia. The results of unsupervised analysis of the “extended” dataset with ADMIXTURE [44, 45] varying the number of ancestral clusters (K) from 2 to 10 (Additional file 2: Figure S2) were most informative for K = 6 (Fig.2), which resulted in lowest cross-validation error and meaningful distribution of components between studied populations.

Fig. 2

Admixture proportions in studied populations, K = 6. Populations from the Extended dataset. Abbreviated population codes: NSK - Russians from Novosibirsk; STV -Starover Russians; ARK: Bashkirs from Arkhangelskiy district; BRZ - Bashkirs from Burzyansky district

To enable direct comparison with previous world-wide studies by National Genographic Project [7, 8, 10, 31, 39, 48], ADMIXTURE analysis in supervised mode were performed for K = 9. For each chip type, we selected a subset of SNPs (~30 K and ~60 K for the two chip types) that matched National Genographic Project chip Geno 2.0 [7].

Among the profiled populations, the degrees of ancient admixture varied dramatically. While the populations residing in the Caucasus and East Siberia regions were mostly represented by a single component (dark green and dark blue, correspondingly), the samples from the Volga-Ural region exhibit substantial admixture of European and Asian components (red, dark green, and light blue). Importantly, for each of the populations, the ratio between components was characteristic of that population (Additional files 3: Tables S2 and S3). Moreover, we observed a subdivision of the European genetic component into two clusters, one most prevalent in the Caucasus (dark green) and another with highest frequency in Northern Europe (red). These two European components jointly account for 50% - 90% of admixture vectors in both Turkic and Uralic speakers of Volga-Ural region, while in Finno-Ugric speakers in Northeast Europe and in all Slavic populations these components account for almost 90% of the gene pool.

Slavic speakers – Russians, Ukrainians, and Belarussians - are similar in their genetic composition. Ukrainians and Belarusians have almost identical proportions of the two “European” components and have virtually no “Asian” admixture. Russians from Siberia - both Novosibirsk residents’ and Russian Eastern Orthodox Old Believers’ (Starover) samples - being genetically close to Slavs residing in Europe, also have between five to 10 % of Central Siberian ancestry (light blue).

Asian genetic ancestry of the profiled populations is represented by four components. Two of them, light green (Beringian) and dark blue (East Siberian), are geographically confined to Northeast Asia, while showing only minor impact on populations west of Ural Mountains. On the contrary, East Asian (pink) and Central Siberian (light blue) components are also present in populations to the west of Ural.

In Central Asian Turkic speakers, including Kazakh and Uzbek, East Asian genetic influence is dominant (>35%), while in Bashkir it is detected at somewhat lower levels (~ 20%). Importantly, in Western Turkic speakers, like Chuvash and Volga Tatar, the East Asian component was detected only in low amounts (~ 5%).

The light blue genetic component dominates genetic landscape of populations inhabiting West and Central Siberia: Ugric-speaking Khanty and Mansi, Samoyedic speaking Selkups and linguistically isolated Ket. However, this ancestry component is present not only in Siberia, but also on the western side of Ural Mountains, though at somewhat lower frequencies - 20-30% in Komi (16% on average) and Udmurt (27% on average) who belong to the Permic branch of Uralic languages. Interestingly, similar levels of this ancestry component (16–23%) are also exhibited by Turkic speaking Chuvash (20% on average) and Bashkir (17% on average), while Tatar, who also reside in the Volga region and have related linguistic and cultural profiles, only show at most 15% (10% on average) of this genetic component. Even lower levels of this ancestry component (<5%) were observed in Turkic speakers of Central Asia.

The Beringian component (light green) is confined exclusive to indigenous populations of Eskimo, Chukchi and Koryak. The East Siberian (dark blue) component is represented by Turkic and Samoyedic speakers of Central Siberian plateau: Yakut, Dolgan and Nganasan. This component is also found at moderate frequencies in Mongolic and Turkic speakers in Baikal region and Central Asia (5–15%), and, at low but discernible frequencies (1–5%), in Turkic speakers residing in Volga-Ural region.

Admin
Administrator

Posts: 73,018

Volga Tatars Oct 22, 2023 5:11:29 GMT

Quote

Post by Admin on Oct 22, 2023 5:11:29 GMT

f3 population test
ADMIXTURE-guided ancestry clustering suggests that, in most of the populations studied, the genetic background is complex, as it includes at least three hypothetical ancestral components: European, East Asian and North Asian (Siberian). Because the ADMIXTURE-guided ancestry clustering cannot be used in lieu of the formal test of admixture, three-population test [40] was conducted to confirm proposed admixture events. Surrogate populations were selected for each of three ancestral components, followed by f3 test of admixture, to find out whether a target population is admixed between two source populations. The combinations of source and target populations and the Z score in the f3 test are reported in Additional file 3: Table S4.

Admixture scenarios were tested in sampled populations geographically grouped as Northern Europe, Volga-Ural region and Central Asia and, while assuming two-way mixture between European and East Asian, or between European and Central Siberian surrogate populations. Significant Z scores (Z < −3) for admixture between East European (Belarus, Ukrainian) and either East Asian (Tuva, Yakut) or Central Siberian (Khanty, Ket) populations were obtained for all populations of Northern Europe and Volga-Ural region.

Note, that the value of f3(X; A, B) (and the corresponding Z score) is negative if X is a mixture of A and B. Among all pairs compared, the most negative f3 values were obtained for populations of Volga-Ural region using East Slavs and Siberians as source groups (Additional file 3: Table S4). This suggests that both, Western (including ancestors of East Slavs) and Eastern (Siberian) sources of the formation of Volga-Ural populations, which could be also seen from the ADMIXTURE plot (Fig. 2). Although Ukrainians and Belarus received similar scores in the f3 test, Belarus turned out to be a slightly better proxy for East European component in populations of Volga-Ural region.

Populations from Central Asia also received significant negative Z scores in the f3 test using East Asia and Europe populations as a source, although their best supported surrogate populations were different. For Uzbek and Kazakh, the best surrogate for a European source was Abkhaz, while for Volga-Ural populations it was either Belarus or Ukrainian. For Uzbek, the best East Asian surrogate was Yakut, while for Kazakh and Kyrgyz it was Tuva (followed by Yakut).

Analysis of IBD blocks
To identify shared identical-by-descent (IBD) blocks, we used the fastIBD algorithm implemented in the Beagle package [46]. We first calculated total amount of IBD in centimorgans (cM) shared between the populations, which was then averaged per pair of individuals (Additional file 3: Table S6). Since the length of IBD blocks is anti-correlated with their respective age, analysing the distribution of length of blocks allows us to examine patterns in ancestry sharing with temporal resolution [49]. We analysed the amount of shared IBD in two length bins: 1–3 cM (ancient blocks) and 4–10 cM (recent blocks) (Fig. 3 and Additional file 3: Table S5). We focused on three regions, providing the densest sample coverage: Caucasus, Volga basin and Siberia. Populations from the Caucasus share most of IBD blocks between themselves and the amount of shared IBD ranges between 3.26 cM to 12.39 cM per pair, which is roughly comparable to the amount of IBD blocks shared between Eastern European populations in our dataset (Additional file 3: Table S5). A conspicuous exception are the Chechens, who share almost all detected IBD blocks among themselves and only a scant amount of IBD with neighbouring Caucasus populations. This may have happened because the Chechen sample is the only representative of the North-East Caucasus in our dataset. Low amount of IBD shared outside the cluster of Caucasian populations suggests a lack of recent ancestry links with Uralic, Slavic or Turkic people (except for Turkic from the Caucasus) present in our dataset.

Admin
Administrator

Posts: 73,018

Volga Tatars Oct 23, 2023 4:31:51 GMT

Quote

Post by Admin on Oct 23, 2023 4:31:51 GMT

Fig. 3

Sharing of ancient and recent IBD blocks between populations in focus regions. IBD sharing is calculated as a sum of IBD segments averaged per pair of individuals. Sharing of IBD blocks is calculated separately for short, ancient blocks (1–3 cM) and more recent 4–10 cM blocks. Recent IBD blocks are typically shared inside the populations, while sharing of ancient IBD blocks is more complex. Darker colors correspond to higher amounts of shared IBD

The Volga-Ural region is populated by three major language and cultural groups: Uralic, Turkic and Slavic speakers. Bashkir and Tatar are major Turkic groups in the region. Although both ethnic groups live in the same region and their languages are mutually intelligible, we surprisingly detected only a limited amount of ancient IBD blocks shared between them, and their overall IBD sharing pattern is different: Tatar share moderate amount of IBD (3.55–7.35 cM per pair) with all neighbouring populations, while Bashkir share most of their ancient blocks (on average 8.62 cM per pair) with Khanty, a group of Uralic speakers from Western Siberia. We speculate that this disparity between cultural and genetic affinities of Tatar and Bashkir can be attributed to a phenomenon of cultural dominance: the population ancestral to Bashkir adopted the Turkic language during Turkic expansion from the east (language replacement event).

European and Siberian Uralic speakers are separated by the Ural Mountain range. This separation affects sharing of recent IBD blocks: Komi and Udmurt share in the interval of 4–10 cM almost double the amount of IBD they share with Khanty. Interestingly, the situation is different when we look at sharing of ancient IBD blocks: the amount of ancient IBD blocks that Udmurt and Komi share with Khanty (10.63 and 8.62 cM per pair, respectively) is comparable to the amount of ancient IBD shared between them (11.38 cM per pair). These data agree well with the ethnic history of European and Asian Uralic speakers [50]. Udmurt and Komi belong to the Permic branch of the Uralic language family and share a recent origin, as demonstrated by common short IBD blocks, their split from their ancestral Uralic population is dated back to the first half of the 1st millennium BC. Split between ancestors of Asian Uralic people, represented in this paper by Khanty, and ancestors of modern European Uralic ethnic groups (Udmurt and Komi) is much older and dated back approximately to 3rd millennium BC. All modern Uralic populations share common genetic substrate inherited from some ancient Uralic people, reflected in long and similar size ancient IBD blocks shared between Udmurt, Komi and Khanty. All analysed native Siberian populations exhibit high levels of intrapopulation sharing of IBD (Fig 3), which is in line with observed long runs of homozygosity in these populations. High rates of shared IBD blocks were detected also between pairs of Siberian populations, particularly between Ket and Khanty.

Following B Yunusbayev, M Metspalu, E Metspalu, A Valeev, S Litvinov, R Valiev, V Akhmetova, E Balanovska, O Balanovsky, S Turdikulova, et al. [19], who observed that the number of shared IBD blocks decline exponentially with the distance between populations, we have calculated linear regression between geographic distance and logarithm of IBD for all pairs of populations from the “Core” dataset (see Additional file 3: Tables S7 and S8). This resulted in the following equation: log 10 IBD = 1.772 − 0.0005975 × (Distance in kilometers), adjusted R 2 = 0.4923, p-value <10−16.

Although Russian Starovers appeared in Siberia only about 300 years ago, details of their geographic origin are not clear. Starovers share roughly twice as much of ancient blocks with Komi and Karelian than between themselves, or with other Slavic speaking groups (Additional file 3: Table S5). This finding, corroborated by results of GPS and reAdmix (described below), strongly points to Northern European ancestral ties of Russian Starovers. The most pronounced difference between predicted and observed IBD rate was found for Siberian Russians, the residents of Novosibirsk and the Starovers, and for the Chinese outgroup (Additional file 3: Tables S7 and S8, Fig. 4). This was expected due to relatively recent relocation of Eastern European Slavic individuals to Siberia. Interestingly, one of the Caucasian ethnicities, the Chechens, also share fewer IBD blocks with Chinese than would be expected from the geographical distance separating these populations. This outlier could be possibly explained by underestimation of the degree of apparent geographical isolation of Chechens, occupying hard to reach highlands of Caucasus range, since elevation is not considered by the regression. From patterns of IBD-sharing and from ADMIXTURE-based analysis we see that Caucasus populations differ from their neighbors in the European part of the Russian Federation. One of the factors being that some Caucasus tribes reside at high altitudes of 2000 m above sea level or more, where they have been genetically isolated for centuries [51,52,53,54].

Admin
Administrator

Posts: 73,018

Volga Tatars Oct 24, 2023 3:45:30 GMT

Quote

Post by Admin on Oct 24, 2023 3:45:30 GMT

Fig. 4

Regression between logarithm of IBD (logIBD) and geographic distance between all pairs of studied Eurasian populations. Red line denotes the regression line and blue lines correspond to 95% prediction interval

At the next stage of analysis, Novosibirsk Russian and Chinese populations were excluded, leaving us with 325 population pairs. This analysis produced following equation: log 10 IBD = 1.394 − 0.0004121 × (Distance in kilometers), adjusted R 2 = 0.2774, p–value <10−16. Using this equation, we have identified pairs of populations (using the “core” dataset) that were more than two standard deviations away from the predicted log 10 IBD (Fig. 4). In this analysis, the departure from the regression line suggests unusual gene flow events, unaccounted for by the calculation of geographic proximity. Most significant departures were observed for Yakut with Kyrgyz, Kazakh, Buryat and Tuva, as well as for Ket with Khanty and Tuva combinations. Starover, who recently migrated to Siberia, again show genetic proximity to Karelian (Fig. 5).

Fig. 5

Departures from the expected IBD. Shown populations exceed the expected IBD sharing by more than two standard deviations. Departure from expected values is most pronounced among Siberian populations, and between Karel and Russian Starovers

Admin
Administrator

Posts: 73,018

Volga Tatars Oct 25, 2023 2:58:15 GMT

Quote

Post by Admin on Oct 25, 2023 2:58:15 GMT

GPS and reAdmix analyses of self-identified Russians
In our dataset, ethnic Russians constituted the largest population (N = 80). These samples came from two groups: residents of Novosibirsk district (39 samples, designated “NSK”) and the Siberian Starover (41 samples, designated “STV”). The Novosibirsk residents were thoroughly surveyed about their ancestors and selected only if reporting at least three preceding Russian generations, while members of the Starover cohort were all assumed to be “authentic” Russians. Since their resettlement from the European part of Russia in the seventeenth century, Russian Starovers deliberately adhered to a strict religious routine and avoided contact with neighbouring Native Siberian populations. Both the Starovers and most Novosibirsk residents are informally considered as “canonical Russians”. Nevertheless, only the Novosibirsk group represents a uniform sample from the modern Russian gene pool.

For all populations sampled in this study, SNP array data were compared to the worldwide collection of populations using Geno 2.0130 K ancestry-informative markers (AIMs) [55]. Both SNP platforms used in our analysis contain a subset of these markers. Chip 370 includes 60,730 AIMs, while chip 610 includes 90,231 AIMs. As demonstrated earlier [8], even in case if admixture vectors are determined with as little as 40,000 AIMs, the difference between the admixture vectors obtained from the complete set of AIMs and the reduced set does not exceed 3%. Therefore, reduction of AIMs to 60,000 or more resides within the range of natural variation and does not affect the accuracy of population assignment.

In the following analysis, we used admixture vectors obtained by ADMIXTURE software run with reference dataset from E Elhaik, T Tatarinova, D Chebotarev, IS Piras, C Maria Calò, A De Montis, M Atzori, M Marini, S Tofanelli, P Francalacci, et al. [8] in supervised mode (К = 9). When admixture vectors for Novosibirsk and Starover Russians were compared, relative weights of the “Northern European” component were found to differ by 2% (t-test p-value <0.009). The provenances of the samples were inferred by two algorithms, GPS [8] and reAdmix [39]. For each tested individual, GPS algorithm determines a location on a world map, where people with similar genotypes are most likely to reside. Notably, this algorithm is not suitable for analysis of recently mixed individuals, such as children of parents from two different ethnic groups. When subjected to GPS, a recently admixed sample would result in a report of high uncertainty of prediction.

To analyze modern Russians from Novosibirsk and Starover Russians, we used Russian diversity panel data genotyped on Geno 2.0 chip (Balanovsky et al., unpublished data). Nearly 37% of self-identified Russians from Novosibirsk were mapped to various Russian populations from European part of Russia: 13% to Tver region, 13% to Arkhangelsk region, 5% to Ryazan region and 3% to Don Cossacks and Vologda region each. Not surprisingly, as many as 27% of Novosibirsk residents were identified as Mordva: 24% as Erzya and 3% as Moksha (Fig. 6). These two subethnic groups were followed by Chuvash (16%), Karelian and Evenki (5% each). Many singular representations of other ethnic groups of the Russian Federation were also reported.