Introduction

Y-STRs and Y-SNPs on the non-recombining region of the human Y chromosome have been widely studied to understand the population history, origin, and migration of human populations (de Knijff 2000; Jobling and Tyler-Smith 2003). Such studies have greatly benefited from analyses of the non-recombining region of the Y chromosome, as this region is not altered by the conventional recombinational processes and is transmitted intact from one generation to the next, revealing the deep paternal lineages of a population. Populations share Y-chromosome lineages because of common origins or gene flow from the populations in neighbouring areas.

Hungarian Y-chromosome studies have been analysed (Völgyi et al. 2009; Bíró et al. 2015; Csányi et al. 2008), but only for 12 STRs and without a regional focus (except for Székely and Csángó populations from Romania). We have, therefore, analysed a specific Hungarian sub-population for 23 Y-STRs. The Bodrogköz region was chosen due to its isolated nature. Until the end of the 19th century, this area was a moorland encircled by the Tisza, Bodrog, and Latorca Rivers. Living in refugium, inhabitants of this part of Hungary escaped from both Tatar and Ottoman invasions, which decimated the post-Hungarian Conquest populations in many parts of the country. Furthermore, in the first half of the tenth century, this region served as the Palatial Centre and burial grounds of the Hungarian tribes (Révész 1996). It has thus been assumed that the present population in this area is likely to be more similar to the population that lived in the Hungarian Conquest period than to less isolated and protected populations in other areas of Hungary (Csányi et al. 2008).

Materials and methods

Materials

We collected 147 unrelated male samples from 24 small villages (Fig. 1) in Bodrogköz in north-eastern Hungary. Informed consent was obtained from all individual participants included in the study. The samples published herein were sent to the YHRD and the accession number is YA004249.

Fig. 1
figure 1

147 non-related Hungarian male individuals representing 24 different geographical regions in Bodrogköz of North-Eastern Hungary. The circle sizes are proportional to the sample sizes

Full size image

Methods

Testing of Y-STR and Y-SNP markers

Genomic DNA was extracted from buccal swab samples using the DNAIQ system (Promega, Madison, WI, USA), as described in the manufacturer’s instructions. The samples were quantified using the ABI 7500 Real-time PCR System (Life Technologies, Foster City, CA, USA).

DNA was amplified using the Promega PowerPlex Y23 amplification kit, including 23 Y-STR loci, according to the manufacturer’s instructions. Fragment sizes and allele designations were determined using a Genetic Analyzer ABI3130 (Applied Biosystems, Foster City, CA, USA) using the GeneMapper ID-X v.1.2 software. When testing Y-SNP markers, amplifications of 3–5 ng genomic DNA were performed in an ABI 7500 and in GeneAmp 9700 thermal cyclers, with TaqMan probes using the programs designed by ABI. The relative fluorescence of the PCR products was analysed on an ABI 7500 Real-time PCR System using its SDS software, as described in the ABI manufacturer’s manual. The SNP markers tested were E-M35, E-M78, E-M81, E-M123, G-M201, G2-L156, G1-M285, H1-M52, H1a-M82, I-M170, I1-M253, I1-L22, I2a1-M223, I2a2-P37, J-M304, J1-M267, J2-M172, J2-M67, J2b-M12, R1a-M198, R1a-M458, R1a-Z93, R1a-Z280, R-M207, R1-M173, R2-M124, R1b-M343, R1b-M269, R1b-U106, R1b-P312, R1b-M412, N-M231, N1c-L1034, N1c-VL29, N1c-Z1936, and Q-M242.

A list of primers and TaqMan assays for binary markers is used in the laboratory, as shown in Online Resource 1 (ESM_1).

Phylogenetic study

To examine the STR variation within the sub-haplogroups, relational networks were constructed using the Network 5.0.0.0 program (Bandelt et al. 1999). Repeats of the DYS389I locus were subtracted from the DYS389II locus. To put the results into a more extensive geographical context, we included haplotypes of 17 overlapping loci from other populations. Within the network programme, the rho statistic was used to estimate the time to the most recent common ancestor (TMRCA) of haplotypes within the compared haplogroups (Bandelt et al. 1999). We used the pedigree rate 2.5 × 10−3, as described by Goedbloed et al. (2009), for 17 loci included in the YFilter kit (Applied Biosystem, Foster City, CA, USA). Due to the expansion of Full Genomic testing, an SNP count-based TMRCA estimation method was also used and proved to be effective on closely related, as well as genetically distant individuals (Hallast et al. 2015; Underhill et al. 2015). In these recent studies, the pedigree rate based on large number STR loci (Goedbloed et al. 2009) tended to be significantly closer to the SNP-based time estimates.

Genetic structure

Pairwise Fst genetic distances were calculated based on haplogroup frequencies using the Arlequin 3.5 software (Excoffer and Lischer 2010). Y-STR-based pairwise Rst genetic distances were computed using YHRD.ORG’s online AMOVA program (release 52) (Willuweit and Roewer 2015). Kruskal’s non-metric multidimensional scaling (MDS) plots were constructed with the isoMDS() function from the R package ‘MASS’ (ver. 7.3-45) (Venables and Ripley 2002).

Results

The frequencies of the haplogroups are presented in Table 1. The 23 Y-STR haplotype and haplogroup results of the 147 males are shown in Online Resource 2 (ESM_2). The largest haplogroups were R1a-M458 (20.4%), I2a1-P37 (19%), R1a-Z280 (14.3%), and E1b-M78 (10.2%). Various R1b-M343 subgroups accounted for 15% of the Bodrogköz population. Haplogroup N1c-Tat covered 6.2% of the lineages, but most of it belonged to the N1c-VL29 subgroup, which is more frequent among Balto-Slavic speaking than Finno-Ugric speaking peoples. Other haplogroups had frequencies of less than 5%.

Table 1 Haplogroup distribution of the Bodrogköz population group
Full size table

Phylogenetic study

Based on Y-STR loci (haplotypes), networks were constructed within each of the haplogroups. To put the results into a wider geographical context, we included haplotypes of 17 overlapping loci from other populations. We included three networks (N1c-VL29, J2b-M12, and R1a-M458) in the main text of the study that could be potentially useful for unravelling the genetic history of Hungarians. Other networks belonging to non-frequent subgroups are not included in this study.

A median-joining network of 151 N1c-VL29 haplotypes was generated using Bodrogköz Hungarians (n = 5), Ukrainians (n = 33), Belarusians (n = 22), Central Russians (n = 14), Northern Russians (n = 6), Slovaks (n = 4), Lithuanians (n = 17), Estonians (n = 26), Finns (n = 5), Saami (n = 5), Karelians (n = 5), Vepsa (n = 1), Mari (n = 5), Komi (n = 1), and Chuvash (n = 3). DYS635 and DYS385a-b (3 loci) data were not available in the N1c-Tat network published by Ilumae et al. (2016). The remaining 14 loci had an estimated 657–788 year/1 molecular step (25 vs. 30 year/generation), based on pedigree data. We compared the data with SNP-based estimates available at YFull (https://www.yfull.com/tree/).

TMRCA estimates (with 30 years/generation):

All VL29 (n = 151): 2541 ± 473 years (YFull by SNPs: 3500 ± 500 years).

M2783 (“Balto-Slavic” branch): 2373 ± 561 years (YFull: 2800 ± 400 years).

The founder N1c-VL29 haplotype was shared by 11 samples (4 Estonians, 3 Karelians, 3 Northern Russians, and 1 Komi), which is in agreement with the postulated Finno-Ugric origin of the clade (Fig. 2, see cluster 1). We did not test our samples deeper than N1c-VL29; however, based on deep SNP data from public databases, we have further divided into subgroups. There is an N1c-M2783 (whose new name is N3a3-M2783) subgroup defined by DYS19 = 15 (instead of DYS19 = 14), which is the predominant N1c-VL29 subgroup among Balto-Slavic-speaking populations (Fig. 2, clusters 2, 3, 4, 6). The founder haplotype was a Lithuanian male (Fig. 2, haplotype 2). N1c-M2783 (xZ16981) modal (Fig. 2, cluster 6) was the second most populous haplotype, with 9 haplotypes (5 Ukrainians, 2 Lithuanians, 1 Belarusian, and 1 Hungarian sample from Bodrogköz). At one molecular step, the L551 (Fig. 2, cluster 4) founder haplotype (notably the clade of Gedeminids, i.e., Lithuanian Gedeminid and Polish Jagiellonian kings) was shared by two Lithuanians, two Ukrainians, one Belarusian, and one Central Russian. The Z16981 (Fig. 2, cluster 3) founder haplotype contained 1–1 Estonian, Slovak, Belarusian and Central Russian. L1022 (Fig. 2, cluster 5) consisted of predominantly Baltic-Finnic haplotypes, but no Hungarians. It should also be noted that no Hungarian belonged to the L551 clade, which means that we could not anchor the genetic descendants of Jagiellonians in Bodrogköz. To summarize, out of five Bodrogköz Hungarian N1c-VL29 samples, four may be of post-Hungarian Conquest Lithuanian or Slavic admixture and one of Finno-Ugric origin, with no recent matches.

Fig. 2
figure 2

Median-Joining Networks (MJ) of 151 N1c-VL29 haplotypes constructed. The circle sizes are proportional to the haplotype frequencies. The smallest area is equivalent to one individual

Full size image

A median-joining network of 175 R1a-M458 samples was generated with Bodrogköz Hungarians (n = 30), Slovaks (n = 57), Slovak Roma (n = 7), Poles (n = 23), Czechs (n = 6), Germans (n = 13), Dutch (n = 2), Danish (n = 1), Austrians (n = 1), Croatians (n = 14), Bosnians (n = 3), Slovenians (n = 1), Macedonians (n = 1), Ukrainians (n = 7), Russians (n = 5), Konda Mansi (n = 2), and Estonians (n = 2), taken from data published by Underhill et al. (2015). This network had two large clusters. We have not tested SNPs below M458; however, based on deep SNP data from public databases, they can be defined by SNPs L260 and L1029. The L1029 founder haplotype was shared by two Hungarians and three Slovaks (Fig. 3, see cluster 1). In the L260 area (Fig. 3, see cluster 2), two Hungarian and one Estonian formed the founder haplotype. Based on a pairwise comparison of all haplotypes, eight samples are ancient (Migration period) Slavic admixture, two are likely Finno-Ugric, six Bodrogköz samples may be a post-Hungarian Conquest Slovak/Polish admixture, two are likely Finno-Ugric, and eight are a more ancient Slavic admixture (from the Migration period), while 14 R1a-M458 samples are unlikely to be of recent Slavic admixture and are of unspecified origin. It is important to note that Mansi R1a-M458 samples do not have close Russian or other Slavic matches, but they may be of recent Slavic admixture. TMRCA calculations based on 17 STR loci (556–668 years/mutation) were very close to SNP-based estimates of YFull.

Fig. 3
figure 3

Median-Joining Networks (MJ) of 175 R1a-M458 haplotypes constructed. The circle sizes are proportional to the haplotype frequencies. The smallest area is equivalent to one individual

Full size image

TMRCA estimates (30 years/generation):

L1029 area: 2063 ± 327 YBP (YFull: 2100 ± 200 YBP).

L260 area: 2442 ± 368 YBP (YFull: 2600 ± 500 YBP).

A median-joining network of 65 J2b-M12 haplotypes was generated with Bodrogköz Hungarians (n = 5), Serbians (n = 3), Bulgarians (n = 12), Ukrainians (n = 2), Austrians from East Tyrol (n = 4), Northern Albanians/Gheg (n = 30), and Central Asians (n = 9) (Sarno et al. 2016; Mirabal et al. 2009; Karachanak et al. 2013; Mielnik-Sikorska et al. 2013; Di Cristofaro et al. 2013; Niederstätter et al. 2012). This was probably the most surprising of all the networks. While most Albanians and Bulgarians clustered together, Serbian, Hungarian, and Austrian J2b-M12 haplotypes were very distant from the South Balkan cluster and in some cases were closer to Asian haplotypes. Four Hungarian samples split from a common ancestor much earlier (as illustrated by the diversity and the molecular mutational steps shown in Fig. 4), and only one belonged to the Albanian–Bulgarian cluster.

Fig. 4
figure 4

Median-Joining Networks (MJ) of 65 J2b-M12 haplotypes constructed. The circle sizes are proportional to the haplotype frequencies. The smallest area is equivalent to one individual

Full size image

TMRCA estimates (30 years/generation):

Albanian–Bulgarian cluster: 1781 ± 381 YBP.

Serbian–Hungarian–MidEast branch: 5344 ± 1215 YBP.

YFull (SNP-based) all M12: 15.700 ± 1700 YBP.

Genetic structure

Microsatellite genetic markers, such as STRs, are based on the variation in the number of repeats of DNA canonical sequences and serve as a potential tool to address population genetics and demographic history. STR loci are typically characterized by high mutation rates and high levels of polymorphisms, as well as by mutational processes that cause preferentially stepwise changes in the number of repeats and allele size. The larger the difference between two different STRs, the higher the number of mutation events is postulated to have occurred since common ancestry (Hardy et al. 2003).

Rst is an analogy of Fst based on allele size differences, so it is defined as the correlation of allele sizes between tested markers within populations. Fst is more efficient when there are high levels of gene flow, whereas Rst better reflects population differentiation under low gene flow (Balloux and Goudet 2002). This is the reason we used Rst-based MDS analysis for population comparison.

Fst genetic distances and p values between 58 populations, including Bodrogköz Hungarians and Eurasian populations from published data, were calculated as shown in Online Resource 3 (ESM_3) and presented as an MDS plot (see Fig. 5). Affinities between populations may result from their common origin or from recent admixture due to geographic proximity. Genetic distances between populations, in particular, can generally be related to geographic distances, according to a model of isolation by distance (Cavalli-Sforza and Bodmer 1977). As shown in Fig. 5, the Bodrogköz Hungarian population group, along with two Hungarian, Ukrainian, Western Slovakian, Slovenian, and Moldavian populations, formed a compact cluster based on close Fst genetic distances with non-significant p values (p > 0.05).

Fig. 5
figure 5

Multidimensional scaling (MDS) plot constructed on Fst genetic distances of Y haplogroup frequencies of 58 populations compared. A three-letter coded population name is presented in Online Resources 2

Full size image

Out of the six populations, five populations were geographic neighbours of Hungarians, except for Moldavians, a finding which fits with the criteria mentioned above. These negligible genetic distances can be seen among those populations, which can be considered a single population from a genetic point of view (such pairs being Norwegians and Swedes; Bavarians, East Tyrolean Austrians and Northeast Italians; Croatians and Bosnians; Uzbeks and Tajiks in Afghanistan, and Georgians and Circassians). Apart from the East-Central European cluster, a Southeast European cluster with non-significant Fst genetic distances is visible among Macedonians, Bulgarians, Serbians, Romanians, and Gagauz Turks. Moldovans may form a link between the Southeast and East-Central European clusters. Interestingly, Hungarian-speaking Székely in Romania exhibited a high similarity to not only other Székely and Hungarians, but also to Greeks, who may derive from pre-Slavic genetic interactions between the two populations. Romanians are most similar to Macedonians, which could support the origin of Romanians in the Macedonian region as shown by the presence of Aromun populations in that area. Hungarian-speaking Csángó in Romania were closest to the Gagauz sample.

We also constructed non-metric MDS based on Y-chromosomal haplotypes that consisted of 23 STR loci available from 22 populations (http://www.yhrd.org). As shown in Fig. 6, the Bodrogköz Hungarian population showed a non-significant Rst genetic distance from Ljubljana Slovenians, which is in line with the haplogroup-based Fst result (Slovak, Ukrainian, and Moldovan STR data were not available). Similarly, a smaller than 0.05 Rst distance was observed among other Hungarian, Czech, Upper Bavarian, Bosnian and Herzegovian, and Croatian and Polish populations. Rst genetic distances and Rst p values of the studied populations are presented in Online Resource 4 (ESM_4).

Fig. 6
figure 6

Multidimensional scaling (MDS) plot constructed on Rst genetic distances of 23 STR-based Y haplotype frequencies of 22 populations compared

Full size image

The genetic distances between 35 populations, including the Bodrogköz Hungarians, and other Eurasians (http://www.yhrd.org), were calculated based on minimum haplotypes, and were again presented as an MDS plot (Fig. 7). The Bodrogköz Hungarian population showed a close genetic affinity with non-significant Rst p values with surrounding populations, such as Slovenian, Ukrainian, Croatian, Slovakian, Hungarian, Ukrainian, and Romanian Csángó. Rst genetic distances and Rst p values of the studied populations are presented in Online Resource 5 (ESM_5).

Fig. 7
figure 7

Multidimensional scaling (MDS) plot constructed on Rst genetic distances of Y-chromosomal minimum haplotype frequencies of 35 populations compared

Full size image

Discussion

The main objectives of this study were to estimate the level of various admixtures in the Bodrogköz population and to determine the most likely arrival period of the analysed paternal lineages on 17 Y-STR loci comparisons with neighbouring populations. We also accounted for the pedigree mutation rate of the 17 loci (YFilter kit), as described by Goedbloed et al. (2009), considering that one mutation occurs every 668 years on average per 17 loci (2.5 × 10−3/668 years at 17 loci) or 705 year per 15 loci if we exclude locus DYS385. For simplicity, we defined a single mutational event (one STR molecular step) as one genetic distance (GD). This made it possible to classify the Y-haplogroups into four large groups: pre-Roman/Roman era ancient lineage (genetic distance to closest non-Hungarian match, i.e., GD >2, approximately 2000 years or more), original Finno-Ugric arriving into the Carpathian Basin (closest match from Finno-Ugric-speaking population), Migration period (admixture, GD 2, approximately 1300–1400 years), and post-Hungarian Conquest admixture (GD <2, i.e., less than 1000 years). According to historical sources, the pre-Roman period (up to 1 year BC), Roman period (between first and fourth centuries AD), Migration period (between third and ninth centuries AD), Hungarian Conquering period (tenth century AD), and post-Hungarian Conquest period (from the 11th century AD) lasted in Carpathian Basin, Hungary (Mócsy 1987; Bóna 1987; Győrffy 1987; Bóna 1987). These historical periods are consistent with TMRCAs calculated by the pedigree mutation rate and reported in this article.

Of the analysed samples, 29.3% (n = 43) were of the Migration Era admixture (2 molecular steps to the closest non-Hungarian match) and another 43 samples were likely remnants of pre-Roman times (3 or more molecular steps from the closest non-Hungarian matches). This implies that on the eve of the Hungarian Conquest, half of the population of the Carpathian Basin was Roman/pre-Roman, while the other half consisted of early Slavs, Germanic peoples, and Sarmatians/Alans from the Migration Era.

We identified nine samples (6.2%) as likely Finno-Ugric lineages, with genetic matches closer to those of Uralic ethnic groups, such as Mansi and/or Estonian, than to non-Uralic groups (5 R1a-Z280, 2 R1a-M458, 1 N1c-Z1936, and 1 N1c-L1034). While haplogroup R1a-M458 has been considered “West Slavic”, our analysis showed that over half of the 30 M458 samples were not of recent Slavic admixture. For example, the TMRCA of R1a-M458, with 4700 ± 800 years based on SNPs, predates the proto-Slavic period of 300–600 AD, and this is also true of the biggest subclades: L260 2500 ± 500 YBP and L1029 2200 ± 200 YBP.

In case of R1a-Z280 and R1a-M458, deep subgrouping is needed to verify the possibility of the existence of Finno-Ugric clusters. Before this study, the predominant haplogroup of Finno-Ugric speaking peoples, namely, N1c-Tat, was not found among recent inhabitants of Hungary (Semino et al. 2000; Lahermo et al. 2000; Csányi et al. 2008). However, two out of four ancient samples from Hungarian Conquest period proved to be belonged to haplogroup N1c-Tat (Csányi et al. 2008). In our recent study, one N1c-L1034 sample was found in Bodrogköz, which has proven to be a paternal link among Hungarians and Western Siberian Mansi (Fehér et al. 2015), and N1c-Z1936 (xL1034) was found to likely be early Ugric. Since this sample showed close genetic relationship with a sample of a warrior from a Hungarian Conquest period cemetery (our unpublished data), we could anchor the genetic descendants of a Hungarian Conqueror men. In addition, based on Y-STR analysis, one R1b-M343* sample was recent Turkic admixture from the R1b-M73 Kypchak (DYS390 = 19) cluster; however, the time of its arrival in Carpathian Basin is unknown, though it could be of Avar, Onugor, Hungarian, or Cumanian origin. Based on the Bodrogköz population, the arriving Finno-Ugric-speaking paternal lineages likely made up no more than 10% of the population of Hungary in the tenth century AD (this does not mean that all of such lineages arrived during the Hungarian Conquest or that all Conquerors were Finno-Ugric speakers). The share of Finno-Ugric-looking paternal lineages shrank to 6.2% due to continuous influx and assimilation of neighbouring peoples in the following centuries.

The clearest post-Hungarian Conquest admixture types were the genetically Roma (2 H1a-M82) and Ashkenazi Jewish samples (J1-M267, Q-M242). Overall, according to our estimate, 35.4% of all samples (n = 52) were likely to be the result of post-Hungarian Conquest (896 AD) admixture in the Bodrogköz Hungarians, meaning that they have either a zero or a single molecular step to non-Hungarian haplotypes. The admixture originated from neighbouring Germanic (G2a-L156, I1-M253, I2-M223, R1b-P312, R1b-U106) and Balto-Slavic (E-M78, I2a1-P37, N1c-VL29, R1a-M458, R1a-Z280) populations. Two samples could be identified as Albanian admixture (1 E-M78 and 1 J2b-M12).

The relatively high rate of N1c-VL29 haplotypes, as well as the close genetic similarity between the recent Moldavian, Ukrainian and Hungarian populations of Bodrogköz, may be related to the escape of a Lithuanian prince to Hungary. Teodoras Karijotaitis, grandson of the Lithuanian Grand Prince Gediminas, had been defeated in a feud for the throne in 1393 AD, escaped to Hungary and lived under the protection of his relative, King Sigismund, in the adjacent areas of Bodrogköz. According to historical sources, 40,000 people came to Hungary from Moldavia and Podolia (part of present-day Ukraine), along with the Lithuanian prince (Войтович 2000).

Our study highlights the important issue that haplogroups cannot be easily linked to specific ethnic groups or admixture types without deep phylogenetic analysis and that the paternal genetic legacy of Hungarians and other neighbouring peoples is complicated. In most cases, significant migrations/demographic growth predated not only the Roman period, but also the migration waves of Germans, Slavs, and Hungarians. This is also underlined by our MDS analysis (Fig. 5), which revealed that Eastern European populations (e.g., Hungarians, Slovenians, Slovaks, Ukrainians, and Moldovans) represented close genetic affinities and a similar haplogroup composition, which, in most cases, precludes the ethnic identification of the paternal lineages.

Non-metric MDS plots constructed based on Fst genetic distances (Fig. 5) and Rst genetic distances (Fig. 6) using 23 Y-STRs showed that Bodrogköz Hungarians, as well as two other contemporary Hungarian samples, were genetically nearly identical to Slovenians, Western Slovaks, Ukrainians and Moldovans, because they were all part of an “East-Central European” genetic cluster, regardless of their linguistic differences.

The MDS plot (Fig. 7) compares 35 population groups with Hungarian, Hungarian-speaking Székely, Csángó groups in Romania and other Eurasian populations, using minimum Y haplotypes (9 Y-STRs). The general structure of the distance matrix was depicted by an MDS plot, where the Bodrogköz Hungarian population showed a close genetic affinity (with non-significant p values) with the surrounding populations, such as Slovenian, Ukrainian, Croatian, Slovakian, Hungarian, Ukrainian, and Romanian Csángó.

Other population groups that we compared were clearly separated from the Bodrogköz Hungarian population. This segregation was statistically significant between most pairs of population groups. Based on Rst genetic distances, the Bodrogköz Hungarians and Székely in Romania (Korond) were close to each other (0.0479) and statistically significant, suggesting close genetic relationships. Finally, we would like to highlight several Asian populations (e.g., Kazakh, Mongolian, Georgian, and Turkish in Central Anatolia) that are close to the Bodrogköz Hungarians based on statistically significant Rst genetic distances, indicating an overall similarity of genetic structuring of Y-chromosome haplotypes.

In conclusion, present-day Hungarian speakers are genetically very similar to neighbouring populations, the Bodrogköz Hungarians and to isolated Hungarian-speaking groups (Csángó and Székely in Romania). We also conclude that they have a relatively higher presence of genetic elements that are missing from present-day neighbouring populations. However, we could classify the Y-chromosomal haplogroups into four large groups based on STR mutation events: pre-Roman/Roman ancient lineage, Finno-Ugric speakers arriving into the Carpathian Basin, Migration period admixture, and post-Hungarian Conquest admixture. Our study highlights the crucial finding that haplogroups cannot be easily linked to specific ethnic groups or admixture without deep phylogenetic analysis, as well as the fact that the paternal genetic legacy of Hungarians and other neighbouring peoples is very complicated. Our initial hypothesis that the present-day population in Bodrogköz is likely to be more similar to the population that lived in the Conquest period, was partly confirmed. It is clear that a significantly larger database with deep haplogroup resolution, including ancient DNA data, is required to strengthen this research.