Edited by: Marshall Abrams, The University of Alabama at Birmingham, United States
Reviewed by: Antonio González-Martín, Complutense University of Madrid, Spain; Natalia Martinkova, The Czech Academy of Sciences (ASCR), Czechia
This article was submitted to Evolutionary and Population Genetics, a section of the journal Frontiers in Genetics
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
History of East-Central Europe has been intertwined with the history of Turks in the past. A significant part of this region of Europe has been fallen under Ottoman control during the 150 years of Ottoman occupation in the 16–17th centuries. The presence of the Ottoman Empire affected this area not only culturally but also demographically. The Romani people, the largest ethnic minority of the East-Central European area, share an even more eventful past with Turkish people from the time of their migration throughout Eurasia and they were a notable ethnic group in East-Central Europe in the Ottoman era already. The relationship of Turks with East-Central European ethnic groups and with regional Roma ethnicity was investigated based on genome-wide autosomal single nucleotide polymorphism data. Population structure analysis, ancestry estimation, various formal tests of admixture and DNA segment analyses were carried out in order to shed light to the conclusion of these events on a genome-wide basis. Analyses show that the Ottoman occupation of Europe left detectable impact in the affected East-Central European area and shaped the ancestry of the Romani people as well. We estimate that the investigated European populations have an average identity-by-descent share of 0.61 with Turks, which is notable, compared to other European populations living in West and North Europe far from the affected area, and compared to the share of Sardinians, living isolated from these events. Admixture of Roma and Turks during the Ottoman rule show also high extent.
According to genome-wide studies based on sex chromosomes, Europeans have two main features in their genetic makeup regarding the paternal and maternal lineages. Investigations of Y haplogroups show that there is geographical pattern in the Y haplogroup characteristics of distinct European regional populations. East-Central Europe and the rest of East Europe separate clearly from the Northern and Western parts of Europe regarding some major haplogroups. The R1b haplogroup is primarily a characteristic of West Europeans and its significance strongly decreases toward East Europe, where R1a is the predominant R haplogroup. R1b is Celtic, Saxon, Basque and Frisian related, originating from the Neolithic. R1a is mainly a Slavic derived haplogroup (
Roma (Romani, Gypsies) are a traditionally itinerant, diverse in culture and dispersed ethnic group with a population counting at least 10–15 million individuals worldwide, with the largest population size in Europe (
Europe has a long common history with Turks, which dates back to the dawn of the Ottoman era, when Turks began to conquer the Eastern regions of Europe. The Ottoman Empire was founded by Oghuz Turks led by Osman I in Anatolia in the 13th century (
Researchers on the field of history show us that Romani people share an even more eventful history with Turks. Romani people share a common history with Turks not only because they lived in East-Central European countries under Ottoman occupation, but because their migration route included also Oghuz Turk occupying area. After their exodus from Northwest India, Roma reached the Caucasus region, the Middle East, including the Anatolian peninsula, in the 12th and the Balkans in the 13th centuries (
The conquering campaigns of the Ottoman Empire played an important role in the history of East-Central Europe. The effect of the Ottoman occupation left its mark on these parts of the European continent not only in a demographic and cultural but supposedly also in a genetic manner. Although, the genetic impact on populations living in the formerly Ottoman occupied area remained unrevealed and has not yet been studied using genome-wide autosomal marker data. Our aim was to investigate the relationship of Turks with the major ethnic groups of East-Central European countries and with the local Roma ethnic minority using genome-wide single nucleotide polymorphism (SNP) array data. We attempted to measure the Turkish ancestry in these populations. However, according to the history of East Europe, other Turkic ancestry can also be derived from several nomadic Turkic tribes, such as the Cumans, Kipchacks, Pechenegs and Bulgars invading and ruling West Eurasia (including parts of East Europe as well), and also the Khazars, of which a group called Khabars joined to the migrating ancestors of Hungarians (
In summary, our study concentrates on the effect of the Ottoman Empire on the formerly Ottoman occupied area of East-Central Europe in a genetic level, including major ethnic groups of regional countries and also Romani people, the largest ethnic minority of the area. We attempted to reveal their presumed admixture with Turks.
The populations investigated in this study were obtained from various sources. Our Roma data consists of upon request available data and in international collaboration collected and genotyped data (
The participants gave their written informed consent to participate in this study. They all got personal verbal information prior their signed consent, which was approved for this study by the Regional Research Ethics Committee. All samples were anonymized. The research was conducted according to the principles expressed in the Declaration of Helsinki.
The 19 Turkish samples, featuring 555,736 SNPs, were obtained from the free public repository of the Estonian Biocentre (
The Ottoman Empire in Eurasia. Territory of the Ottoman Empire at its greatest extent, tributary and vassal states of the Empire. Investigated regions and ethnic groups are also indicated.
To study the relationship of the investigated populations, we implemented two different methods. The software SMARTPCA from EIGENSOFT 6.01 Software package was applied to perform principal component analysis (PCA) on our data and to compute pairwise average allele frequency differentiation (Fst) values (
For population structure and ancestry analysis, we created a merged dataset containing Turks, Roma and OEC, EUR, OCA, OME, Turkmens, Punjabi, and Han Chinese populations (
To test if East-Central Europeans and Roma have Turkish ancestry, we used the algorithm implementing 4-Population Test included in the ADMIXTOOLS 4.1 Software Package (
For revealing admixture between East-Central Europeans and Turks, we created a dataset containing Turkish, OEC and EUR samples (
In order to reveal admixture between Roma and Turks, we merged the data of Roma, Turks, OCA, OME, Turkmens, and Onge (
In order to test the proportion of genome-wide Turkish ancestry in Roma, we used an F4 ratio estimation algorithm, which is also part of the ADMIXTOOLS package. We considered the Punjabi as an ancestral population according to recent findings, which points out that Roma people originate from the Northwestern part of India (
For calculating the genome-wide ancestry proportion of Turks in Roma compared to Punjabi, we used a merged dataset of Roma, Turkish, OCA, Turkmen, Punjabi, and Han Chinese samples (
We computed the ratio of F4 (Turkmensi, Han Chinesei; Romai, Punjabisi)/(Turkmensi, Han Chinesei; Turksi, Punjabisi) and standard errors were computed using block jackknife with block size of 5 centimorgan (cM).
To reveal the relationship of East-Central Europeans and Roma with Turks and with other formerly Ottoman-occupied populations, and to confirm also our findings regarding the admixture events with Turks, we applied IBD segment detection using the Refined IBD method of Beagle 4.1. We created a dataset containing OEC, Roma, Turkish samples and EUR, OCA and OME groups (
We used the output of Beagle to compute an average pairwise IBD sharing between populations I and J.
where IBDij is the length of IBD segment shared between individuals i and j and n, m are the number of individuals in population I and J (
In order to estimate the date of admixture of East-Central Europeans and Roma with Turks, and therefore confirm that admixture with Turks could occur between these populations during the Ottoman occupation of East-Central Europe, we applied the ALDER algorithm (
A merged dataset was created considering Turkish, OEC and EUR samples (
For an admixture date estimate between Roma and Turks, we created a merged dataset consisting of Onge, Turkish and Roma population samples with an SNP number of 83,459 and
Data filtering and curve fitting options were set to default values in both of the tests, and we applied the fast Fourier transform algorithm.
We applied PCA using SMARTPCA and the clustering software ADMIXTURE to reveal the relationship of East-Central Europeans, Roma and Turks to each other and to other Eurasian populations. We merged the Turkish and Roma population samples into a massive dataset containing also the OEC, OCA, OME, and EUR groups. We included also Turkmen, Punjabi and Han Chinese samples in this dataset. Turkmens are Turkic people in Central Asia next to Iran, lying outside of the territory of the former Ottoman Empire at its maximum extent. Punjabi, a Northwest Indian ethnic group, represents the population of the source area of Roma. It allows us to better understand the place of Romani people in the Eurasian supercontinent. Han Chinese was used to represent East Asian populations in the Eurasian perspective.
Principal component analysis results, plotted on the first four principal components, show the three main groupings of the analyzed populations, which are Europeans, populations of the Caucasus region together with Middle East populations and the South Asian ethnic group Punjabi. Roma people cluster mostly between the Caucasus populations and Punjabi. There are also several more admixed Roma individuals scattered throughout a straight line of which two ends are defined by Europeans and South Asians (
Relationship of Turks and Roma to distinct Eurasian Populations according to PCA. PCA results plotted on four principal components. Each symbol represents one individual.
The ADMIXTURE analysis produced similar results and visualizes the data similarly to PCA in a stacked column chart style. At
Relationship of Turks and Roma to various Eurasian populations according to ADMIXTURE analysis. ADMIXTURE analysis results with
We computed the genome-wide pairwise average allele frequency differentiation values (Fst) between the investigated populations. Examining Fst values between Turks and OEC, EUR populations show that OEC have lower Fst with Turks than the Fst between Turks and EUR populations (
Population relationships based on pairwise average allele frequency differentiation (Fst) estimations.
Turks | Roma | OEC | OCA | OME | EUR | Sardinians | |
---|---|---|---|---|---|---|---|
Turks | |||||||
Roma | 0.011 | ||||||
OEC | 0.006 | 0.014 | |||||
OCA | 0.004 | 0.014 | 0.009 | ||||
OME | 0.003 | 0.014 | 0.008 | 0.006 | |||
EUR | 0.009 | 0.017 | 0.005 | 0.013 | 0.016 | ||
Sardinians | 0.013 | 0.023 | 0.009 | 0.019 | 0.016 | 0.012 | |
Turkmens | 0.013 | 0.021 | 0.017 | 0.015 | 0.014 | 0.022 | 0.02 |
A formal test of admixture, the 4-population test, was applied to find evidence that East-Central Europeans and Roma have admixed with Turks (
In case of East-Central Europeans, accordingly to the PCA results, we used Russian individuals from European populations that were never affected by the Ottoman rule. The Turkic people Chuvash from East Europe was also considered in the test to try to rule out other possible Turkic ancestry. We tested if the phylogenetic tree, on which East-Central Europeans to East Europeans and Turks to Turkic groups are closely related. The 4-population test showed a violation of the expected phylogenetic tree, confirming that East-Central-Europeans are admixed with Turks.
In order to examine whether Roma and Turks are admixed, we considered Onge samples, which group is an accurate surrogate of the South Asian ancestry of Roma. Since recent admixture between Onge and West Eurasians did not occur, European ancestry components cannot influence our test (
We applied F4 ratio estimation to measure the proportion of Turkish ancestry in the Roma (
To detail further the relationship of East-Central Europeans and Roma to Turks, and assess the Turkish ancestry in East-Central Europeans and Roma, we estimated the average pairwise Identity-by-descent segment sharing between certain populations. We created a merged dataset of the Turks, Roma and groups living in the formerly Ottoman-ruled regions, as well as EUR populations. OEC populations have an average pairwise IBD share of 0.68 with Turks. This means that East-Central Europeans have similarly higher average IBD sharing with Turks than the populations of formerly Ottoman-occupied Caucasus and Middle Eastern regions, which was 0.75 and 0.60, respectively. The average sharing between other Europeans and Turks and between Sardinians and Turks were 0.51 and 0.42, respectively. These results suggest that Ottoman rule in East-Central Europe has a detectable impact on its populations (
Population relationships based on average IBD sharing estimations. We computed the genome-wide pairwise average shared IBD length between certain groups. OEC – Formerly Ottoman-occupied East Central Europe, OCA – Formerly Ottoman-occupied Caucasus region, OME – Formerly Ottoman-occupied Middle East, EUR – Europe free from Ottoman conquest.
In order to further investigate Roma admixture with Turks during the Ottoman rule, we compared the average IBD sharing with Turks and OCA. The migration route of Roma included Turkey and the neighboring Caucasus region, therefore the rate of admixture with Turks and populations from the Caucasus should be close to each other if an admixture between Roma and Turks did not occur during the Ottoman occupation. However, IBD analysis shows a higher average IBD sharing with Turks (0.71) than Caucasus populations (0.60), likely confirming the assumption that admixture between Roma and Turks occurred during the Ottoman occupation (
In order to further investigate Roma admixture with Turks, we used ALDER to estimate admixture dates. Preliminary ALDER analyses applying multiple reference populations showed that Portuguese samples are appropriate surrogates for representing a population not affected by the Ottoman era and Hungarians are an appropriate surrogate of populations from the formerly Ottoman occupied East-Central European region. According to the results of ALDER, admixture between East-Central Europeans occurred 18.09 ± 5.33 generations ago (
ALDER analysis results. Weighted LD curves calculated by the ALDER algorithm.
Admixture between Roma and Turks occurred 30.56 ± 3.77 generations or approximately 777–996 years ago (
In order to better understand the relationship of Roma, East-Central European populations and Turks to each other and in a Eurasian perspective, we applied the population structure and ancestry analysis methods PCA and ADMIXTURE. These methods helped to place the investigated populations in a Eurasian context, which showed that East-Central Europeans and Turks belong to two well-separated groups. Roma, due to their nomadic nature, scattered severely between Europe and South Asia. However, most Roma individuals clustered somewhat more tightly to each other, scattering between South Asia and the populations of the Caucasus, Middle East and Central Asia. PCA and ADMIXTURE results reflected the actual geographical positions of East-Central Europeans and Turks. Ancestry analysis and Fst calculations showed that East-Central Europeans have significant Turkish ancestry, even compared to other formerly Ottoman-occupied East-Central European populations. These analyses also revealed that Roma might have remarkable Turkish ancestry compared to the populations of neighboring regions of Turkey, which also fall on the migration route of the ancestors of recent Roma.
Based on the results of PCA and ADMIXTURE, we investigated the proposed admixture events with formal test of admixture between Roma, East-Central Europeans and Turks, which could provide evidence of gene flow between these populations. In case of East-Central Europeans, we used East European Slavic and Turkic population samples to find evidences of admixture with Turks. The 4-population test results showed that gene flow between East Europeans and Turks occurred. This result suggests that the test revealed Turkish ancestry from the Ottoman occupation of East-Central European territories. Estimating the average IBD sharing of East-Central Europeans with Turks showed that East-Central European populations have higher average IBD sharing with Turkic people than other formerly Ottoman-occupied populations. The 4-population test showed also that Roma and Turks are admixed, and as we did in case of East-Central Europeans, we confirmed with average IBD sharing estimation that this admixture could originate from the former Ottoman rule of East-Central Europe, since Turkish ancestry in Roma shows a higher degree than the ancestry of populations living adjacent regions of Turkey.
To assess the significance of average IBD sharing difference values between investigated populations or groups, we also calculated the average IBD sharing of Turks with Sardinians, which group lives separately from the continent on the Sardinia Island, therefore they are largely isolated from European demographic events. This investigation suggested that average IBD sharing differences of investigated populations with Turks are significant.
We estimated also the date of the admixtures in order to further investigate and provide evidence for the proposed admixture of East-Central Europeans and Roma with Turks during the Ottoman occupation. In case of East-Central Europeans, the obtained admixture date interval corresponds to the time interval when Ottomans were present in East-Central Europe. Ottoman presence in the region solidified approximately with the Battle of Kosovo (1389), and fall of Ottoman rule in East-Central Europe is associated with the successful campaign led by the Habsburgs, which resulted in the Treaty of Karlowitz signed in 1699 (
We confirmed that the expansion of the Ottoman Empire into East-Central Europe left its mark on local populations, contributing significantly to their Turkic ancestry. Using genome-wide autosomal SNP array data, we were able to find a similar geographical pattern in the genetic makeup of European populations, as studies based on Y haplogroups found. Population structure showed that Western and Eastern European populations are well-separated from each other. Ancestry analysis, testing for admixture and IBD segment analyses further separated OEC groups from the rest of Europe, showing that ancestry derived from the Middle Eastern area is also observable in autosomal data. As the Middle Eastern region derived Y haplogroups E3b and J were observed in high extent in East-Central Europeans (
This study was carried out in accordance with the recommendations of the Hungarian Human Genetic Law (act XXI/2008) with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Research Ethics Committee in Pécs (REKEB).
ZB, BIM, and KS conceived and designed the investigations. ZB and BIM evaluated and interpreted the results. KH, KS, and BM contributed in the data collection and genotyping. ZB, BM, AM, and MK cowrote the manuscript. BM, AM, and MK revised the initial draft of the manuscript. All authors were involved in this work.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The present scientific contribution is dedicated to the 650th anniversary of the foundation of the University of Pécs, Hungary. Detailed information about the methods and sample collection for the Population Reference Sample (POPRES) data are described in
The Supplementary Material for this article can be found online at: