Genetic continuity of Indo-Iranian speakers since the Iron Age in southern Central Asia | Scientific Reports – Nature.com


Modern Indo-Iranian genetic affinities with ancient samples

To explore the relation between present-day Central Asian individuals and the Eurasian genomic diversity, ancient and modern, we first performed a Principal Component Analysis (PCA) (Fig. 1b, Supplementary fig. S1 and S2) on 1915 modern genomes and projected 3102 ancient genome-wide data onto it. Regarding the present-day Eurasian diversity, the three top Principal Components (PCs) roughly mimic the geographical repartition of modern populations: the PC1 (3% of variance) discriminates between Eastern and Western Eurasian individuals, the PC2 between South Asian and modern European individuals, and the PC3 discriminates against the Baikal populations from the East Asian cluster (see Supplementary fig. S1). Present-day Indo-Iranian individuals from Central Asia cluster together on the first three PCs while Turko-Mongol individuals form a gradient from the Indo-Iranian cluster to ancient Baikal samples on PC3, in agreement with cultural categorization instead of geography. However, a substructure appears within the Indo-Iranian group with the Yaghnobis (TJY) falling closely to the Western cluster, while the Tajiks populations (TJA, TJE, TAB) stretch toward the Baikal cluster, indicating some additional East Asian or Baikal Hunter-Gatherer (BHG) proximity.

For the ancient individuals, Bronze Age, Iron Age, and historical steppe individuals fall on a cline stretching up from European to East Asian groups, with Western_Steppe individuals clustering on the bottom of the European cluster and Central_Steppe individuals spreading from the Western_Steppe cluster to the Okunevo_BA cluster close to Baikal and Siberian modern individuals. The ancient individuals of southern Central Asia (Neolithic, Bronze Age and Iron Age) follow a cline stretching from Neolithic Iranian individuals (Iran_N) to present-day Iranians and Yaghnobis.

Contrastingly, the Iron Age samples (Turkmenistan_IA and Ksirov_Kushan individuals) are located close to modern Indo-Iranian populations, although slightly negative values on the first axis and positive values on the third axis suggest an addition of Baikal ancestry in the present-day Indo-Iranians. Finally, it appears from this PCA (Fig. 1c) that ancient and present-day Indo-Iranian populations from Central Asia form together a cline between Iranian Neolithic farmers and Central_Steppe Bronze Age, with a clear shift in ancestry toward Steppe between Bronze Age and Iron Age as observed before18, and a smaller shift toward eastern Asian ancestry between Iron Age and present-day. This shift is more pronounced for Tajiks than Yaghnobis.

To confirm our initial observations and identify genetic structures, we performed an unsupervised clustering analysis using ADMIXTURE47 on the same dataset used for the PCA (see Supplementary fig. S4, S5 and S6). Consistently with the PCA, we evidenced in all modern Indo-Iranians the presence of a genetic component maximized in Iran Neolithic farmers (Iran_N, dark green; mean value for Yaghnobis: 37%; 25% for Tajiks), of another maximized in Eastern European Hunter-Gatherers (EEHG) and Western Scandinavian Hunter-Gatherers (WSHG) (pale green; mean value for Yaghnobis: 13%; for Tajiks: 10%) and of a third component (dark blue; mean value for Yaghnobis: 36%; for Tajiks: 29%) that is not completely maximized in any population of our dataset, but is found in present-day Europeans and in Anatolian Neolithic farmers (Anatolia_N). In addition, a fourth component maximized in Baikal Hunter-Gatherers (BHG: Shamanka_EN) and largely present in all modern Turko-Mongol populations (red; 50% on average) is also inferred to a lower extent in the modern Indo-Iranian populations, with a significantly smaller proportion in Yaghnobis than in Tajiks (mean value respectively 7% and 14%; t-test p-value = 2.10–16). Finally, the Tajiks present a small proportion (4%) of modern East Asian ancestry (pink component, maximized in the Han population), which is largely present in all Turko-Mongol populations from Central Asia (mean value 10%), and around 8% of the component maximized in present-day South Asian populations (orange), which are both absent in Yaghnobis.

The ADMIXTURE analysis is also congruent with the PCA concerning the ancient groups (see Supplementary fig. S4 and S5). Indeed, Iron Age southern Central Asian individuals present a remarkably similar profile to Yaghnobis’ profile: for instance, the individual labelled as Turkmenistan_IA has a profile with about 25% of WSHG/EEHG component, 30% of Iran_N component and 35% of the Anatolian farmer ancestry component but missing BHG ancestry (Fig. 2). Bronze Age Central Steppe pastoralists show a similar profile except for a significant increase in Iranian ancestry, and Western Steppe pastoralists have the beige component maximized in Western European Hunter-Gatherers (WEHG), which is absent in modern Indo-Iranian populations.

Figure 2
figure2

ADMIXTURE analysis of 5019 individuals (3102 ancient and 1915 modern). The results for a subset of the dataset (present-day Indo-Iranian individuals and ancient populations discussed in the main text) are displayed for K = 10, which has the lowest cross validation value (0.994). The full analysis is shown in SI. In the first column, the modern individuals from Central Asia; second column, the ancient individuals from southern Central Asia; third column, ancient individuals from the Steppe; last column, miscellaneous individuals discussed in the main text. (Figure done with ggplot2 v. 3.3.3 R package https://cran.r-project.org/web/packages/ggplot2/index.html).

Thus, modern Indo-Iranian speaking populations appear as midway between Central Steppe and southern Central Asia Bronze Age populations, quite similarly to the Turkmenistan Iron Age individuals, with a limited impulse from eastern and southern Asian groups.

Population continuity within the Indo-Iranians

To formally test for genetic continuity with Iron Age southern Central Asia and the limited admixture with Baikal-related populations at the source of the present-day Indo-Iranian speaking populations, we performed D-statistics, f3-statistics and qpAdm modelling on the same dataset used for the PCA et ADMIXTURE analyses as well as on a dataset formed by shotgun sequences from 3 Yaghnobis (TJY), 19 Tajiks (TJE) and 24 Turkmens (TUR)48 as well as the ancient genomes for a final set of ~ 700k SNPs.

We identified and characterized gene flows that occurred since the Iron Age by computing D-statistics of the form D(Mbuti, Ancient population ; Turkmenistan_IA, present day Indo-Iranian) for every ancient population in our dataset (Fig. 3, Table S5). These statistics are expected positive when gene flows occurred from the Ancient population to the present-day Indo-Iranians. For the Yaghnobis, only one individual, an Iron Age individual from Nepal genetically close to East Asian populations (Nepal_Chokhopani_2700BP.SG)45, has a significantly positive D-statistic (Z > 3). Tajik individuals (TJE) display a higher number of ancient populations (N = 41) for which D-statistic is positive; the common characteristic of these ancient populations is to exhibit a large amount of BHG ancestry, consistently with the ADMIXTURE analysis (Fig. 2). We also note that the Tajiks present a positive D-statistic with an historical individual from India (Great Andaman) (Fig. 3) showing a possible connection with South Asia. Thus, modern Indo-Iranian populations descend from groups related to those present in Turkmenistan as early as Iron Age, with a contribution from another East Asian population who brought the BHG ancestry and, except for Yaghnobis, a contribution from a South Asian population.

Figure 3
figure3

Gene flow in Indo-Iranian populations since Iron Age. Positive D-statistics (Z > 3) of the form D(Mbuti, Ancient population; Turkmenistan_IA, TJY/TJE/TUR). A positive D-statistic demonstrates that a gene flow occurred from the ancient population to the Indo-Iranian or Turkmen population compared to Turkmenistan_IA. The estimated statistic ± 3 standard errors is indicated. (Figure done with ggplot2 v. 3.3.3 R package https://cran.r-project.org/web/packages/ggplot2/index.html).

Then, we formally test if the contributions detected with D-statistics are due to admixture events that occurred since the Iron Age. We first computed f3-statistic49 of the form f3(TJY/TJA/TJE/TAB ; Source1, Source2), that is expected to be negative (Z < -3) if the Indo-Iranian populations can be modeled as admixed between the two sources (Table S7). Only combinations implying a population from East Asia ancestry (like the XiongNu) and westerner populations representing the components seen in the ADMIXTURE analysis (Iranian Neolithic, Anatolian farmer, and Steppe ancestry) were significant (Fig. 2). These statistics attest to the existence of an actual admixture between a population probably presenting a mix of Iran Neolithic, BMAC, Anatolian early neolithic and Bronze Age Steppe ancestry with a population with a strong affinity to the BHG ancestry. The Yaghnobi population has significantly fewer pairs with a negative f3-statistic than the Tajik populations, probably due to their long-term isolation. We also specifically calculated f3-statistic of the form f3(TJY/TJA/TJE/TAB; Ancient population, Turkmenistan_IA) and obtained several negative f3-statistics always with the same ancient populations implied in the positive D-statistic (see Supplementary fig. S3, Table S6) showing that Indo-Iranians can be successfully modelled as the admixture of Iron Age Turkmenistan and BHG-related population.

We then modelled Yaghnobi and Tajik populations using qpAdm23 to estimate mixture proportions. To test which proximal populations fit the best in our model, we used the rotating method23 and we excluded all combinations with a p-value ≤ 0.01. We first tried a two-ways admixture testing several possibilities among rotating sources. For the Yaghnobis, the only model retained was the one with ~ 93–88% from Turkmenistan_IA and ~ 7–12% ancestry from XiongNu (Table 1). With 3-ways modelling, we could not reject different models for TJY: 3 models imply 90% ancestry from Turkmenistan_IA and 7% ancestry from XiongNu, and around 3% of ancestry from Europe_EN, BMAC or Ukraine_Scythian; we also obtained a model with Ukraine_Scythian, BMAC and XiongNu inferring the older admixture at the origin of Turkmenistan_IA (Table 1). When testing for more admixture sources, we obtained only two 4-ways models and one 5-ways model (Supp. Data). One interesting model is a 4-ways model with 17% Ukrainian Scythians, 60% Turkmenistan_IA, 14% BMAC and 8% XiongNu, i.e. this model shows a close affinity of Yaghnobis with Western Steppe-like populations.

Table 1 Plausible models for Yaghnobis (TJY), Tajiks (TJA, TAB), Turkmens (TUR) and Turkmenistan_IA (TurkIA) as a mixture of two or three sources obtained with qpAdm.

To model Tajiks, all 2-ways admixture models were excluded and we obtained one 3-ways admixture model (p-value = 0.49) implying around 17% ancestry from XiongNu, almost 75% ancestry from Turkmenistan_IA, and around 8% ancestry from a South Asian individual (Indian_GreatAndaman_100BP)50 representing a deep ancestry in South Asia (Table 1).

Thus, the qpAdm modelling shows that at least 90% of the ancestry of current Indo-Iranian ancestry is modelized as inherited from Iron Age individuals from southern Central Asia with an affinity with BMAC. Consequently, Indo-Iranians present a strong genetic continuity in the region since the Iron Age with anecdotic admixture with BHG ancestry related individuals, and, for the Tajiks, with South Asian ancestry related populations possibly after Iron Age.

Finally, we used DATES18 to estimate the number of generations since the admixture events. We obtained 35 ± 15 generations for the admixture between Turkmenistan_IA and XiongNu-like populations at the origins of the Yaghnobis, i.e. an admixture event dating back to ~ 1019 ± 447 years ago considering 29 years per generation51. For Tajiks (TJE, TAB, TJA) we obtained dates from ~ 546 ± 138 years ago (18.8 ± 4.7 generations) to ~ 907 ± 617 years ago (31.2 ± 21.3 generations) for the West/East admixture. We also obtained a date of ~ 944 ± 300 years ago for the admixture with the South Asian population.

Iron age Turkmenistan ancestry

Previous studies13,18 have already shown Turkmenistan_IA can be modelled as an admixture between BMAC and some steppe populations, and on the PCA (Fig. 1c), Turkmenistan_IA indeed belongs to the steppe cline. However, the steppes are split between several groups (Western steppe, Central steppe, Eastern steppe) depending on their amount of Eastern Asian ancestry. The ADMIXTURE analysis discriminates the Western and Central steppe ancestries by the presence of a red and mauve component (maximized respectively in East-Siberia and East Asia populations) in the latter, which is absent from Turkmenistan_IA, indicating an affinity with the Western steppe. Nevertheless, we noted that Andronovo or Sintashta individuals also lacked this component while being classified as Central_Steppe. Thus Central_Steppe group is highly heterogenous and gathers populations with some East-Asian ancestry like Karasuk or Central Saka and others more Western steppe-like as Andronovo and Sintashta. Furthermore, we obtained the higher f3-outgroup statistic of the form f3(Mbuti; Ancient pop, Turkmenistan_IA) for ancient populations from BMAC complex or West Eurasia, highlighting the double origin and affinity with the West. This affinity is further confirmed with D-statistics of form D(Mbuti, Turkmenistan_IA; Western_Steppe, Central_Steppe) (Fig. 4B) that are significantly negative (Z < -3) when a Western_Steppe population is opposed to a Central_Steppe population with an East Asian ancestry, like Central Saka or Karasuk (Fig. 4B). With D-statistic of the form D(Mbuti, Turkmenistan_IA; HG1 , HG2) – HG1 and HG2 belonging to WEHG, EEHG, WSHG, and BHG populations – we evidenced that the steppe populations admixed with BMAC lacked East Asian or Baikal component (Fig. 4A). Indeed, we only see significant D-statistics when BHG was confronted with the other HGs (Fig. 4A). Using HG populations avoids inferences from recent admixture; nevertheless, it failed to discriminate between most of the different steppe groups of this period at this level. This suggests that Turkmenistan_IA is devoid of the East Asian ancestry observed in several Central steppe groups as early as Bronze Age.

Figure 4
figure4

Absence of affinity of Turkmenistan_IA with East Asia ancestry shown by D-stat. In grey are non-significative (Z < 3) D-statistics, in blue significative positive D-statistics and in red significative negative one. Only populations with strong East Asian or BHG ancestry show a significative D-statistic. (Figure done with ggplot2 v. 3.3.3 R package https://cran.r-project.org/web/packages/ggplot2/index.html).

Finally, we tested different steppe populations which admixed with BMAC to model Turkmenistan_IA with qpAdm. We first constituted a set with Poltavka, Srubnaya (Western_Steppe) and 4 individuals from Russia labelled as Andronovo (Central_Steppe)52, to estimate the affinity with Europe and Western steppe previously highlighted with D-statistics and f3-statistics. We only obtained one model with 2 sources that we could not exclude (Table 1), and it implies an admixture of 43% BMAC and 57% Andronovo (p-value = 0.31) suggesting that Andronovo individuals are the best proxy for the steppe population which admixed with BMAC to form the Iron Age southern Central Asia group. When testing for the best model between Andronovo and Karasuk (Central steppe with East Asian component) to estimate the affinity with Asia, we produced a single fitting/ relevant model implying Andronovo (p-value = 0.51) with roughly the same proportions. Further tests explored the best model between Andronovo and Sintashta, two genetically close populations, and the single significantly outcome was the one with Andronovo and BMAC (p-value = 0.498) in the same proportions. Eventually, we tested the best model between the individuals labelled Andronovo and two populations belonging to the Andronovo-complex: Fedorovo Shoindykol18 and Alakul Lisakovskiy18. Once again, the only valid model was the one with Andronovo and BMAC. Overall, we can say that the Iron Age population from southern Central Asia emerges from the admixture of BMAC with a Bronze Age population close to the Andronovo individuals, which presents a profile with an affinity with Western steppe rather than with a Central steppe with an affinity with East Asia (like Karasuk).

A Turkmens’ history

Despite speaking a Turko-Mongol language and having the same cultural practices as other Turko-Mongol ethnic groups53, Turkmens are genetically closer to Indo-Iranian populations than to Turko-Mongols54,55.

Indeed, Turkmens (TUR) fall into the Tajiks cluster and not in the Turko-Mongol cline in the PCA (Fig. 1) and in the ADMIXTURE analysis (Fig. 2), all Turko-Mongol populations from Central Asia except Turkmens show a significant (t-test, p-value < 2.10–16) high amount of Baikal (red component, mean 50%) and East Asian ancestry (pink component, maximized in the Han population). Turkmens, for their part, display a completely different pattern, with an amount of Baikal component (mean value: 22%) closer to the proportion in Tajiks (mean value: 15%) and almost no East Asian component. They do not show as much South Asia related ancestry as Tajiks, suggesting that the admixture with South Asian populations occurred or continued after Turkmens split from the remainder of the Indo-Iranian group.

We have established genetic affinity profiles with ancient populations for all Central Asia populations including Turkmens of the first dataset, based on f3-outgroup statistics of the form f3(Mbuti; Ancient pop, Present-day pop) (Fig. 5; Table S8). The f3-outgroup values comparing Turkmen to any ancient population are strongly correlated with the one comparing Tajiks to any ancient populations (Fig. 5a). On the other hand, f3-outgroups values comparing Eastern steppes and Baikal groups to a Turko-Mongol population (Kazakhs) are higher than those comparing these ancient populations to Turkmens (Fig. 5b). The Turkmens are more similar to Indo-Iranian populations than to any Turko-Mongol population on the amount of shared Siberian/East Asian ancestry.

Figure 5
figure5

Turkmens’ affinity with Tajiks rather than with Turko-Mongol groups shown by f3 statistics of the form f3(Mbuti; TUR/TJA/AKZ , Ancient population). (A) Outgroup f3-statistics for Turkmen and for Tajiks (TJA) plotted against each other. (B) Outgroup f3-statistics for Turkmen and for Kazakhs (AKZ), belonging to the Turko-Mongol group, plotted against each other. (Figure done with ggplot2 v. 3.3.3 R package https://cran.r-project.org/web/packages/ggplot2/index.html).

Finally, we modeled Turkmens as a mixture of Central Asia basal ancestry (represented by Yaghnobis) and East Asian ancestry (we obtained a negative value for f3(TUR; TJY, DevilsCave_N); f3 = −0.0025, Z = −5.266). qpAdm modelling for Turkmens produces a single nonrejected model (p-value = 0.048007) implying 6% of Golden Horde Asian and 94% of Tajiks (TAB) (with TJY, XiongNu, GoldenHordeAsian, TAB, Turkmenistan_IA as potential rotating left population) (Table 1). For this admixture event, we estimated a date of 687 ± 100 BP (23.7 ± generations) with DATES.

These results enlighten that Turkmens were an Indo-Iranian-like population not so long ago, who recently shifted language and culture without a substantial genetic change in population.

Source of this news: https://www.nature.com/articles/s41598-021-04144-4

Related posts:

Best Free Proxy Sites and Services to Hide your Web - BollyInside
This list is about the Best Free Proxy Sites and Services to Hide your Web. We will try our best so that you understand this list Best Free Proxy Sites and Services to Hide your Web. I hope you like ...
Points Reasons Why Your Company Should Use Proxy Servers – Occasions when Square Chronicles - Occasi...
More often than not, many people find it difficult to be familiar with use of proxy servers learn company. As a result, they stay away from them as they consider regarding an unnecessary ...
Using Microsoft's YARP project to proxy web-based microservices - OmniChannel Media
Inner source is the idea of using open source techniques to develop internal tools, using platforms such as GitHub for collaboration. Engineers across a company identify common issues and technologi...
30 thoughts on “How To Get Discord to Work at Your School or College” - Alphr
When you are at a school, college, or governmental institution, chances are that your access to certain websites is limited. This is especially true for social platforms or content sharing websites t...
Open Text : What's new in OpenText Exceed TurboX 12.0.4 - Marketscreener.com
As remote work has become the new normal, remote access to business-critical software has become the norm and needs to be supported on any platform and any device. With OpenText™ Exceed TurboX (ET...
Dallas Invents: 129 Patents Granted for Week of March 2 » Dallas Innovates - dallasinnovates.com
Dallas Invents is a weekly look at U.S. patents granted with a connection to the Dallas-Fort Worth-Arlington metro area. Listings include patents granted to local assignees and/or those with a N...
What Are Cookies? And How To Clear Them? - Fossbytes
If you have been on the Internet long enough, you might have heard about cookies once in a while. But what exactly is a cookie? Do they our data? How can we delete them? Here we will try to answer ea...
Form 425 D8 Holdings Corp. Filed by: D8 Holdings Corp. - StreetInsider.com
News and research before you hear about it on CNBC and others. Claim your 1-week free trial to StreetInsider Premium here. Filed by D8 Holdings Corp. pursuant to Rule 425 under the Securities Act o...
Everything you need to know about data extraction - Flux Magazine
words Alexa Wang Data is being generated more than ever. The main reasons for that are the development of digital technologies and the internet, and it’s an excellent opportunity for businesses worl...
How can you unblock websites in UAE? - Daily Host News
UAE is a business hub acknowledged for its sleek skyscrapers. At the moment, UAE’s new political environment is getting influenced because of new cybersecurity threats.As per gulf businesses, four ou...
Using international SIM cards in your Australian phone - CHOICE
If you're travelling overseas and want to stay connected with family and friends or use your phone to get from A to B, you'll need to work out whether you'll use your Australian SIM and pay for inter...
The correct way keyboard warriors made a fabulous killing on PS5s and as a consequence designer trai...
South imon Clements produced a promise last year that he or she has come to regret deeply. Afterward telling his son he could have a PlayStation 5 intended for his 18th birthday, Clements scour...
Comparative host-coronavirus protein interaction networks reveal pan-viral disease mechanisms - Scie...
How lethal coronaviruses engage hostsSevere acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is closely related to the deadly coronaviruses SARS-CoV-1 and Middle East respiratory syndrome corona...
Achronix Announces First Quarter 2021 Financial Results and Business Highlights - Yahoo Finance
Achronix Semiconductor Corporation, a leader in high-performance field-programmable gate arrays (FPGAs) and embedded FPGA (eFPGA) IP, today announced financial results for the first quarter of 2021, ...
Can be the difference between a VPN and a proxy? - TechRadar
So you are looking to add an extra layer pertaining to privacy online, and have discovered the words VPN associated with proxy being thrown around? Both allow you to browse the world anonym...
ESET takes part in global operation to disrupt Zloader botnets - We Live Security
ESET researchers provided technical analysis, statistical information, and known command and control server domain names and IP addresses ESET has collaborated with partners Microsoft’s Digital Cri...
Trojan Shield: FBI punks crims with faux app—and international help - TechBeacon
Police forces around the world are arresting more suspects of organized crime. They’re unsealing evidence gathered over the past two to three years via a private-messaging app, Anom (styled ΛNØM).Age...
How to open and use Windows 11 Settings to tweak your PC - TWCN Tech News
Windows 11 is here now! If you have installed Windows 11 on your computer and are feeling overwhelmed by it, then don’t you worry – This easy-to-understand tutorial will help you sort everything! Win...

IP Rotating Proxy Onsale

SPECIAL LIMITED TIME OFFER

00
Months
00
Days
00
Hours
00
Minutes
00
Seconds
First month free with coupon code FREE30