Ancient DNA | Medieval migrations | India to Europe

Where did the Roma come from?

For five centuries Europe guessed: Egypt, Nubia, Bohemia, Anatolia. The answer was written in a language nobody bothered to study, and confirmed by genomes nobody bothered to sequence. Here is what the DNA of eleven million Europeans finally says, and what the first Roma ever found in an archaeological excavation add to it.

31 to 37% South Asian ancestry in our Global25 models of Roma groups from Bulgaria to Finland, a remarkably stable figure
1502 to 1704 the dates of four genomes from Barc, Albania, the first Roma identified in the ancient DNA record
29 generations since the main admixture with West Eurasians, about 800 to 900 years, as published by Moorjani and colleagues

Hero image: AI generated illustration of a Romani coppersmith family travelling in the late medieval Balkans, with an Orthodox monastery on the ridge behind. It is not a portrait of any real person and does not depict a documented scene.

In 1417, a group of travellers appeared at the gates of German towns carrying letters of safe conduct and a story. They were pilgrims, they said, penitents from a land called Little Egypt, condemned to wander for seven years. The letters worked for a while. The name stuck for six hundred years: Egyptians, Egipcios, Gitanos, Gypsies. It was wrong.

They had also been called Atsinganoi in the Byzantine world, a word probably borrowed from a heretical sect, which became Zigeuner, Tsiganes, Zingari, tigani. That was wrong too, and in Romanian it became the legal term for a slave.

The people who used neither name for themselves called themselves Rom, and their word for the road was drom. Both words are the key to where they came from.

A language that was already an answer

In 1763, a Hungarian theology student named Istvan Valyi met three students from Malabar at the University of Leiden. He wrote down about a thousand words of their language and read them back to Roma near Gyor. They understood a large part of the list. Within a generation, comparative linguists had established what no chronicle had managed: Romani is an Indo-Aryan language, a sister of Hindi, Punjabi and Gujarati, as Indian as Sanskrit is.

The language carries the whole journey in its vocabulary. The core is Indic: pani for water, nakh for nose, trin for three. Then come Iranian words, from Persian or Kurdish: diz for fortress, zor for strength, baxt for luck. Then a thin Armenian layer: bov for oven, kotor for piece, grast for horse. Then, overwhelmingly, Greek: drom for road, foros for town, the numerals efta and oxto for seven and eight, and whole pieces of grammar. By the time Romani left the Byzantine world it was, in the phrase used by linguists, a balkanised Indic language.

Crucially, there is almost no Arabic. The route did not pass through the Arab world or across North Africa, which disposes of the Egyptian story and of the recurring claim that the Gitanos of Spain came from the Maghreb.

Where in India?

Genetics narrowed the question further. The Y chromosome did most of the early work. The lineage H1a1a-M82 is found at between 10 and 60% in various Roma populations and is otherwise a South Asian lineage; Rai and colleagues found that Roma haplotypes matched most closely those of northwestern Indian populations, and estimated the age of the Roma founder lineage at about 1,405 years, with a large error of plus or minus 688 years.

Genome wide data then arrived in a cluster of studies.

  • Mendizabal and colleagues, 2012. Thirteen Romani groups across Europe. A single origin in northern or northwestern India about 1,500 years ago, and a spread across Europe through the Balkans about 900 years ago.
  • Moorjani and colleagues, 2013. About 77.5% West Eurasian ancestry, with the main admixture dated to 29 generations, roughly 780 to 900 years ago, and a founder event about 27 generations back. The strongest sharing of long haplotypes was with northwestern Indian groups.
  • Melegh and colleagues, 2017. One hundred and seventy nine Roma samples against 51 Indian groups. The highest sharing of identical segments was with the northwest, and the abstract names Punjab, Rajasthan and Gujarat. Admixture at 29.9 generations, about 800 to 935 years.
  • Font-Porterias and colleagues, 2019. The most careful dissection so far, with 152 published Roma genomes plus 34 newly genotyped Iberian Roma. They put the South Asian source at about 35% and the West Eurasian at about 65%, and explain that the earlier 80% figure counted as European some ancestry the founders already carried in India. The main West Eurasian admixture dates to about 1270 to 1580 of our era.

That last point matters for anyone reading percentages. The founding group was not a sample of some pure Indian population. Northwestern Indians themselves carry a large share of ancestry related to Iran and the steppe. A Roma genome is therefore not 35% Indian and 65% European. It is 35% derived from a South Asian source that was itself part West Eurasian, which is why the honest summary is closer to: about half the genome traces to the people who left India, and about half to the people they met afterwards.

The first Roma in the ancient DNA record

Until recently, not a single Roma individual had been identified in an ancient DNA study. In 2026, Davranoglou and colleagues published a large study of Albanian origins in Nature Human Behaviour. Among more than six thousand genomes, three individuals from the cemetery of Barc, in the Korca basin of southeastern Albania, had previously been filed as being of Central Asian or Turkic origin. They are not. Their autosomal profile carries clear South Asian ancestry, around 20% membership in the ADMIXTURE clusters that are modal in India. The sole male belongs to the Y lineage J2a-Y18404, which is found almost exclusively among Roma. One of the women carries the mitochondrial lineage U3b1c2. The authors describe this as the first record of Roma people in the ancient DNA record.

Four of these genomes are available in Global25 coordinates, dated between 1502 and 1704 of our era. They are the material for the most interesting test in this article: were the Roma of the early modern Balkans already, genetically, the Roma of today?

Our Global25 analysis

We used four sources throughout: a South Asian population, a Balkan population, a West Asian population, and, where relevant, the local non-Roma population of the country concerned. All the models and their diagnostics are listed in the methods note at the end.

1. First, does the test work at all?

Before reporting any percentage we checked what the estimator does when the true answer is known. We built synthetic genomes by mixing European populations with known quantities of South Asian ancestry, and ran them through the same model.

Validation: known amounts of South Asian ancestry added to European genomes 0% 5% 10% 15% 20% 25% 30% 35% 40% 0% added 0.0% 2% added 0.9% 5% added 3.8% 10% added 8.9% 20% added 19.1% 35% added 34.5% Mean over 15 synthetic mixtures (5 European hosts x 3 South Asian sources). Below 5% the estimate is unreliable.
Figure 1. Validation of the South Asian estimator. Fifteen synthetic mixtures per level, combining five European host populations with three South Asian sources. A 35% spike-in reads back as 34.5%, a 20% spike-in as 19.1%, and a 10% spike-in as 8.9%. Below 5% the estimate becomes unreliable, which is our honest detection floor.

We also ran real non-Roma populations through the model as controls. Albanians, Bulgarians, Serbians, Romanians, Greeks, Hungarians, Slovaks and Andalusians all come out at 0.0% South Asian. A single medieval individual from the same Barc cemetery, who is not Roma, comes out at 0.2%. Turks of Kayseri come out at 10.3%, which is not an artefact: Anatolian populations do carry a real South and Central Asian component, and this is a useful reminder that the estimator measures ancestry, not identity.

2. A South Asian core that barely moves

South Asian ancestry in Roma groups (Global25, NNLS) 0% 5% 10% 15% 20% 25% 30% 35% 40% Barc, Albania, 1502 to 1704 (ancient) 34.4% Bulgaria 37.1% Serbia 33.1% Kosovo 34.5% Bosnia and Herzegovina 35.2% Greece 32.1% Romania 35.5% Turkey 36.7% Czechia and Slovakia 36.7% Russia 33.5% Finland (Kale) 34.2% Spain, Granada 34.2% Spain, Madrid 31.9% Spain, Bilbao 32.6% Spain, Barcelona 19.7% Portugal, Porto 30.8% England (Romanichal) 10.3% Model: a South Asian source + Balkan + West Asian + the local non-Roma population. Detection floor about 5%.
Figure 2. South Asian ancestry estimated for each Roma group. The green bar at the top is the four ancient genomes from Barc in Albania. Everything else is a modern population average.

The striking result is how flat this is. From Bulgaria to Finland, from Turkey to Portugal, across eight hundred years and four thousand kilometres, most Roma communities carry between 31 and 37% South Asian ancestry. Our Balkan figures agree closely with the 35% published by Font-Porterias and colleagues using completely different methods, which is the sort of agreement that makes a number trustworthy.

The two exceptions are informative. The Romanichal of England come out at 10.3% and the Roma of Barcelona at 19.7%. These are not different origins. They are the signature of heavy intermarriage with the surrounding population, which is also visible in the pale fourth segment of the next figure and in the fact that the English sample sits 0.128 away from the Barc Roma, as far as a Bulgarian does.

3. Four layers, country by country

What the four sources contribute, group by group South Asian Balkan West Asian local host Barc, Albania, 1502 to 1704 (ancient) 34% Bulgaria 37% Serbia 33% Kosovo 34% Bosnia and Herzegovina 35% Greece 32% Romania 35% Turkey 37% Czechia and Slovakia 37% Russia 34% Finland (Kale) 34% Spain, Granada 34% Spain, Madrid 32% Spain, Bilbao 33% Spain, Barcelona 20% Portugal, Porto 31% England (Romanichal) 10% Each bar sums to 100%. The number on the right is the South Asian share. Condition numbers 3 to 51.
Figure 3. The composition of each group under the same four source model. The local host component is whatever is left after the South Asian, Balkan and West Asian layers are accounted for.

The West Asian layer, between 13 and 35% in most groups, is the genetic counterpart of the Armenian and Persian words in the language. We tested whether it is an artefact. When synthetic Roma are built with no West Asian ancestry at all, the model returns between 0 and 4% for most Balkan host populations, and up to 13% when the host is a Greek population, which is itself partly West Asian in composition. The 21 to 28% we measure in the Balkan and Central European Roma is well above that. The exact figure, however, is not reliable: it moves by several points depending on which West Asian population is used as the pole, and part of it is ancestry the founders already carried out of northwestern India.

The Balkan layer behaves exactly as Font-Porterias and colleagues described: it is the largest West Eurasian contribution everywhere, including in Spain, Portugal, Russia and Finland, long after those communities left the Balkans. Every Roma community in Europe passed through the Byzantine and Balkan world, and carries it.

4. The Roma of 1502 were already the Roma of today

Genetic distance from the Barc Roma of 1502 to 1704 0.000 0.025 0.050 0.075 0.100 0.125 0.150 Roma of Serbia 0.0154 Roma of Greece 0.0185 Roma of Bulgaria 0.0200 Roma of Czechia and Slovakia 0.0204 Roma, Balkan average 0.0209 Roma of Romania 0.0210 Roma of Kosovo 0.0215 Gitanos of Granada 0.0245 Roma of Turkey 0.0254 Ciganos of Porto 0.0313 Kale of Finland 0.0512 Turks of Kayseri 0.0865 Armenians 0.1200 Punjabi Jat Sikhs 0.1240 Greeks of Thessaly 0.1256 Bulgarians 0.1261 Romanichal of England 0.1276 Albanians 0.1294 Euclidean distance in Global25. Green: Roma groups. Gold: non-Roma reference populations.
Figure 4. Euclidean distance in Global25 between the average of the four Barc genomes and other populations. Green bars are Roma groups, gold bars are non-Roma reference populations.

This is the result we find most remarkable. The four individuals buried at Barc between 1502 and 1704 sit closer to the Roma of Serbia today, at 0.0154, than two populations of the same country usually sit to each other. They are closer to the Gitanos of Granada, 2,000 kilometres away, than to the Albanians among whom they were buried, 0.0245 against 0.1294. Individually they carry 31.3, 32.4, 33.6 and 39.0% South Asian ancestry, which is the modern range.

In other words, the genetic profile of the Roma was already fully formed by 1500. The Indian ancestry, the Balkan ancestry and the West Asian ancestry were already combined in the proportions we see today, and four centuries of European history since then have added local host ancestry to some communities without changing the core.

5. What Global25 cannot tell you

We tried to go further and identify which region of India the founders came from. We could not, and the reason is worth showing.

We scanned 466 South Asian population averages as the South Asian source, scoring each by how well it reproduces the target. On a clean synthetic Roma built from a known northwestern source, the method works perfectly: the true source comes back ranked first out of 466, and the tenth best fit is 0.0056 worse. On real Roma, the ranking collapses: the best and the tenth best fits differ by 0.0009, and the top of the list is a mixture of populations from the northwest, the centre and the south of the subcontinent that are statistically indistinguishable from each other.

Target Best fit overall Gap to the tenth best fit
Synthetic Roma, known northwestern source the true source, rank 1 of 466 0.0056
Balkan Roma, real data no single population stands out 0.0009
Czech and Slovak Roma, real data no single population stands out 0.0012
Barc Roma, 1502 to 1704 no single population stands out 0.0011
Table 1. A method that recovers the correct source from a clean simulation cannot do so on real Roma. Eight hundred years of founder effect, drift and admixture have blurred the regional signal that Global25 would need.

This is why we do not publish a Punjabi or Rajasthani percentage, and why you should be suspicious of any calculator that gives you one to two decimal places. The regional attribution in the literature rests on identical by descent segments and on Y chromosome phylogeography, which use information that a 25 dimensional summary throws away. On that evidence, the northwest of the subcontinent remains the best answer. Our own data neither confirm nor contradict it.

If you have Roma ancestry and want to test it: we built Drom, a report that models your own Global25 coordinates against the South Asian, Persian and Byzantine Anatolian stages of this route and compares you with nine modern Roma communities. The free compatibility check tells you, before you pay anything, whether your profile is a fit for the model.

Eight hundred years of European history, written in the host layer

The differences between Roma groups in Figure 3 are not differences of origin. They are a record of how each society treated them.

In Wallachia and Moldavia, Roma were enslaved. The first document is from 1385, when Prince Dan I assigned forty households of atigani to the monastery of Tismana. Slavery lasted until the 1840s and 1850s, ending for privately owned slaves in 1855 in Moldavia and 1856 in Wallachia, with compensation paid to the owners. Estimates put the enslaved population at the time of abolition at between 200,000 and 250,000 people, around 7% of the population of the principalities.

In Spain, the pragmatic of 1499 ordered Gitanos to settle or leave, and the Great Round-up of 1749 imprisoned thousands in a single coordinated operation. In Habsburg lands, Maria Theresa and Joseph II banned the language, the name and the marriages. In Germany and across occupied Europe, the Porajmos killed somewhere between 200,000 and 500,000 Roma and Sinti.

Against all of that, a group that is intensely endogamous keeps its genetic core intact, and the host percentage rises only where the social barrier was lower. The 75% English ancestry of the Romanichal and the 55% Catalan ancestry of the Roma of Barcelona are, read this way, the measurable trace of local histories of integration and intermarriage. The flat 35% South Asian core everywhere else is the measurable trace of everything that prevented it.

What remains

The Roma left northwestern India around a thousand to fifteen hundred years ago. They spent a long time in a Persian and Armenian speaking world, and a long enough time in Byzantine Anatolia and the Balkans for Greek to reshape their grammar. They entered Western Europe in the fifteenth century with a cover story about Egypt, and by then they were already genetically what they are now, as the Barc graves show.

Eleven million people in Europe descend from that journey. It is the largest minority on the continent, and until a few years ago its genetic history was among the least studied, for reasons that have more to do with marginalisation than with science. Four skeletons from an Albanian hillside have now been correctly identified, five hundred years after they were buried. It is a small correction to the record, and a reminder of how much of this history is still underground.

Methods note

All coordinates are Global25 scaled coordinates. Population averages were taken in priority from the Global25 ancient averages, then the Global25 modern averages, then the Moriopoulos 2026 collection, then the dated ancient individual file. Scale compatibility was verified by comparing independent averages of the same groups across files, for example the Porto Roma, which agree to 0.0000015, and the Granada Roma, which agree to 0.0106, the larger gap reflecting a different sample composition between the two files rather than a scale mismatch. The four source model is a South Asian population with a scheduled caste profile from the Lahore region, a Balkan population, a West Asian population, and the local host where relevant; source separations are 0.329 between the South Asian and Balkan poles, 0.318 between the South Asian and West Asian poles and 0.126 between the Balkan and West Asian poles. Condition numbers range from 3.1 for the three source Balkan models to 50.9 for the Romanian model, and are reported in full in our working files. NNLS fits use a sum to one constraint applied as a weighted extra equation, weight 1000, followed by normalisation. The spike-in validation used five European host populations crossed with three South Asian sources at 0, 2, 5, 10, 20 and 35%. The West Asian spike-in control used eight Balkan and Southern European hosts crossed with three South Asian sources and two West Asian poles. The source region scan covered 466 South Asian population averages. Published admixture dates, haplogroups and qpAdm style proportions are taken from the cited studies and are not recomputed here.

Global25 coordinates used in this article

The four Barc genomes, their average, the modern Roma groups and the source populations used in our models. They can be pasted directly into the Vahaduo tool.

Roma_Barc_Albania_I13836_AD1502,0.094473,0.071087,-0.055814,0.004522,-0.007694,0.014781,0.002350,0.001846,0.017589,0.013303,0.002273,-0.002698,0.003568,0.007019,-0.019815,-0.005834,0.005085,0.000507,0.000628,-0.002376,-0.002246,0.007543,0.005669,-0.000361,-0.002155
Roma_Barc_Albania_I13837_AD1562,0.087644,0.069056,-0.064865,0.007106,-0.019388,0.007809,-0.001880,0.008538,0.015339,0.024420,0.002761,-0.004796,-0.000892,0.008257,-0.008415,-0.016574,-0.004563,-0.003801,0.000880,-0.007879,0.008360,0.003833,0.001356,0.006145,0.002754
Roma_Barc_Albania_I13833_AD1625,0.085367,0.054839,-0.069013,0.022610,-0.024004,0.014223,0.003525,-0.001385,0.005318,0.018224,0.000812,0.005395,0.007136,0.004817,-0.014929,-0.019358,0.004042,0.000380,0.005405,0.001751,-0.003619,0.001484,0.000863,-0.003012,0.000958
Roma_Barc_Albania_I13835_AD1704,0.092197,0.072103,-0.053174,0.015504,-0.010771,0.008367,0.001645,0.000923,0.014930,0.014032,-0.001624,-0.004046,-0.001933,0.009221,-0.014929,-0.008618,0.010300,-0.001014,-0.001383,0.003252,0.007112,0.002473,-0.000863,0.007471,-0.005029
Roma_Barc_Albania_average_n4,0.089636,0.066517,-0.061942,0.013162,-0.014772,0.011365,0.002468,0.003750,0.012834,0.017722,0.003451,0.000337,0.001486,0.007431,-0.015371,-0.011303,0.004010,0.000823,0.002640,-0.003314,0.001778,0.004823,0.002218,0.003826,0.000389
Roma_Bulgaria_n5,0.083474,0.057759,-0.063210,0.021355,-0.018377,0.015794,0.002085,0.005641,0.016159,0.016478,-0.002352,0.000161,-0.000068,0.009971,-0.012673,-0.005916,0.007245,-0.000698,0.000091,0.000577,-0.000664,-0.000033,0.001407,-0.000901,-0.000143
Roma_Serbia_n7,0.090039,0.063856,-0.057386,0.016209,-0.016760,0.011876,0.001005,0.003464,0.010909,0.014220,-0.000459,0.000562,-0.000458,0.005360,-0.010797,-0.002312,0.006717,0.000584,0.001658,0.000675,-0.000801,0.002254,0.001603,0.001371,-0.002712
Roma_Greece_n6,0.086600,0.071685,-0.056156,0.010231,-0.015786,0.005660,0.000850,0.003877,0.015610,0.014190,0.000710,-0.001732,0.004366,0.003550,-0.010382,-0.001045,0.002681,0.000196,0.001546,-0.000643,-0.002405,0.002484,0.001897,0.001453,-0.001028
Roma_Romania_n2,0.089920,0.058901,-0.062979,0.017442,-0.014311,0.011713,0.001293,0.004615,0.017589,0.017039,-0.000650,0.006144,0.000149,0.004404,-0.009501,-0.001127,0.012582,-0.001014,-0.000189,0.002126,-0.001373,0.001484,0.003204,0.005060,-0.000598
Roma_Turkey_n4,0.084986,0.057860,-0.071102,0.015747,-0.029131,0.016841,0.001857,0.003268,0.008808,0.013384,-0.000121,0.001367,-0.003658,0.002726,-0.008204,-0.004418,0.004017,0.000193,0.001330,-0.000436,0.000012,0.002457,0.001421,0.001464,-0.000579
Roma_Czechia_n84,0.087484,0.056969,-0.058950,0.024900,-0.015169,0.016491,0.003100,0.004751,0.013472,0.014291,-0.000380,-0.000677,0.000265,0.008601,-0.009318,-0.008578,0.003937,0.000956,0.000594,0.001419,-0.000869,0.000227,0.002250,0.000498,-0.001813
Roma_Russia_n3,0.090794,0.066384,-0.053496,0.023625,-0.016773,0.015066,0.001591,0.004340,0.012358,0.014127,-0.001760,0.000331,-0.000842,0.007022,-0.009788,-0.004811,0.004820,-0.000419,0.000843,0.000292,-0.002081,0.000630,0.002686,0.000196,-0.000855
Roma_Finland_Kale_n11,0.094571,0.058026,-0.035437,0.046869,-0.013211,0.022725,0.002908,0.006597,0.012893,0.007572,-0.002552,0.000273,-0.001152,0.001973,-0.001614,0.000200,0.002799,-0.000059,0.000499,0.002650,0.001013,0.001922,0.002828,0.004863,0.000140
Roma_England_Romanichal_n3,0.121210,0.115564,0.031483,0.042968,0.022349,0.019374,0.004094,0.006322,0.006160,0.005903,-0.004475,0.004689,-0.010737,-0.009000,0.013178,0.001586,-0.004386,0.000352,0.000118,0.002124,0.003722,0.002223,0.000529,0.007648,-0.000252
Roma_Spain_Granada_n7,0.082278,0.068040,-0.053551,0.018088,-0.010771,0.014343,0.001007,0.006066,0.021504,0.019421,-0.000394,0.000792,-0.005394,0.004561,-0.008647,-0.005398,0.002049,-0.003837,-0.002011,0.000143,-0.001604,-0.001696,-0.000933,-0.002995,0.004858
Roma_Spain_Madrid_n4,0.089067,0.072610,-0.049780,0.020026,-0.006770,0.010389,0.005699,0.002077,0.014316,0.021778,0.003613,-0.004159,-0.004125,0.002305,-0.002171,-0.002983,0.003455,0.000570,0.000691,0.000532,0.000905,0.003926,0.004437,-0.001084,0.003113
Roma_Portugal_Porto_n5,0.088555,0.076774,-0.045481,0.020866,-0.005786,0.016231,0.002914,0.004661,0.015953,0.022962,-0.002923,0.003028,-0.007195,0.003688,-0.009663,-0.009573,0.000261,-0.001495,0.000000,0.006678,-0.001722,-0.003734,0.001997,0.001639,-0.003257
Dom_Iraq_n4,0.029550,0.029225,-0.100300,0.029600,-0.053225,0.021850,-0.003475,0.007575,0.013125,-0.002175,-0.001625,-0.002600,0.005475,-0.004975,0.003825,0.009550,-0.005775,-0.004450,-0.003525,-0.006975,0.004200,-0.009325,0.007925,-0.007750,0.008175
Source_SouthAsian_Punjab_Lahore_Dalit_n42,0.047291,-0.073578,-0.153740,0.110413,-0.069742,0.061110,-0.001662,0.012379,0.042069,0.024363,-0.005316,0.001285,-0.001108,0.002585,-0.003988,-0.004192,0.001999,0.000305,-0.000494,0.001879,0.001168,0.000403,0.000267,0.002450,-0.003718
Source_Balkan_Bulgarian,0.124750,0.134862,0.025644,0.005362,0.024620,-0.000837,0.005499,0.003046,-0.000286,0.006998,-0.003345,0.000899,0.002111,0.008643,-0.014305,-0.002228,0.007093,0.002002,0.003142,-0.003702,-0.004617,-0.000544,0.002810,-0.002482,-0.001078
Source_WestAsian_Armenian_n19,0.104583,0.137746,-0.054989,-0.060475,-0.029358,-0.015262,0.003533,-0.007079,-0.030132,-0.003317,0.002838,0.000725,-0.003098,0.003112,-0.004878,-0.002903,-0.000969,0.000739,0.001581,-0.001969,0.001386,0.002432,-0.003391,-0.001484,-0.000047
Source_Albanian,0.122240,0.143591,0.017566,-0.015768,0.026394,-0.007339,0.002499,0.001469,0.001507,0.017312,-0.000863,0.003573,-0.006240,0.006284,-0.019308,-0.002097,0.011999,0.000537,0.007400,-0.006184,-0.008088,0.001409,0.002835,0.003358,-0.003032
Source_Spanish_Andalucia,0.108132,0.146076,0.035707,-0.003434,0.045256,-0.005857,-0.001930,0.002660,0.024715,0.031162,-0.002154,0.007272,-0.013207,-0.011408,0.010243,-0.002498,-0.006176,-0.001100,-0.004261,-0.002133,0.002929,-0.001302,-0.001972,-0.005334,0.001885

References

  1. Ancient DNA Davranoglou LR, Lauka A, Aristodemou A, Maroti Z, Bojaxhi G, Muhaj A, Mikerezi I, Wesolowski D, Joseph BD, Heraclides A, et al. (2026). Ancient DNA evidence for the history of the Albanians. Nature Human Behaviour 10(7): 1371 to 1391. doi.org/10.1038/s41562-026-02462-z
  2. Preprint Davranoglou LR et al. (2023). Ancient DNA reveals the origins of the Albanians. bioRxiv. The Barc individuals, their South Asian ancestry, the Y lineage J2a-Y18404 and the mitochondrial lineage U3b1c2 are described here. doi.org/10.1101/2023.06.05.543790
  3. Population genetics Font-Porterias N, Arauna LR, Poveda A, Bianco E, Rebato E, Prata MJ, Calafell F, Comas D (2019). European Roma groups show complex West Eurasian admixture footprints and a common South Asian genetic origin. PLoS Genetics 15(9): e1008417. doi.org/10.1371/journal.pgen.1008417
  4. Population genetics Moorjani P, Patterson N, Loh PR, Lipson M, Kisfali P, Melegh BI, Bonin M, Kadasi L, Riess O, Berger B, Reich D, Melegh B (2013). Reconstructing Roma history from genome-wide data. PLoS ONE 8(3): e58633. doi.org/10.1371/journal.pone.0058633
  5. Population genetics Mendizabal I, Lao O, Marigorta UM, Wollstein A, Gusmao L, Ferak V, Ioana M, Jordanova A, Kaneva R, Kouvatsi A, Kucinskas V, Makukh H, Metspalu A, Netea MG, de Pablo R, Pamjav H, Radojkovic D, Rolleston SJH, Sertic J, Macek M, Comas D, Kayser M (2012). Reconstructing the population history of European Romani from genome-wide data. Current Biology 22(24): 2342 to 2349. doi.org/10.1016/j.cub.2012.10.039
  6. Population genetics Melegh BI, Banfai Z, Hadzsiev K, Miseta A, Melegh B (2017). Refining the South Asian origin of the Romani people. BMC Genetics 18: 82. doi.org/10.1186/s12863-017-0547-x
  7. Y chromosome Rai N, Chaubey G, Tamang R, Pathak AK, Singh VK, Karmin M, Singh M, Rani DS, Anugula S, Yadav BK, Singh A, Srinivasagan R, Yadav A, Kashyap M, Narvariya S, Reddy AG, van Driem G, Underhill PA, Villems R, Kivisild T, Singh L, Thangaraj K (2012). The phylogeography of Y-chromosome haplogroup H1a1a-M82 reveals the likely Indian origin of the European Romani populations. PLoS ONE 7(11): e48477. doi.org/10.1371/journal.pone.0048477
  8. Linguistics Matras Y. The structure and history of Romani: the Indic base, the Iranian and Armenian loanwords, and the Greek layer of Early Romani. Romani Project, University of Manchester. romani project
  9. History Slavery in the Danubian Principalities: the 1385 charter of Dan I, the categories of enslaved Roma, and the abolitions of 1855 and 1856. en.wikipedia.org/wiki/Slavery_in_Romania
  10. History Council of Europe. Factsheets on Romani history: the arrival in Europe and the journey from India. rm.coe.int
  11. Data Global25 PCA coordinates by Davidski, and the Moriopoulos 2026 collection of Global25 coordinates.