The Genetic Origins of the Roma Diaspora

The Roma, known as Gitanos in Spain, Tziganes in France, Zigeuner in Germany, are among the most genetically distinct populations in Europe. Despite a thousand years of European presence, their autosomal DNA continues to carry an unmistakable signature from the Indian subcontinent. This article uses G25 coordinates from the Moriopoulos 2025 modern collection to model the ancestry of nine Roma diaspora populations across the Balkans, Central Europe, and the Iberian Peninsula, and to quantify both their shared South Asian heritage and the variable degree of local European admixture accumulated along their long journey west.

I. Historical Migration Route

The Roma are not a monolithic group but a cluster of related populations who share a common origin in northwestern India, most likely in the region spanning today's Punjab, Rajasthan and Sindh. The earliest genetic divergence from South Asian sister populations is estimated between the 5th and 11th centuries CE, with most genomic studies converging on a departure from the subcontinent around the 9th, 10th century CE.

From the Indian subcontinent, ancestral Roma populations moved into Persia and the Caucasus, a passage clearly reflected in the autosomal signal. Byzantine Greek sources mention a group called the Atsingani in 9th, 11th century Anatolia, widely considered the first historical reference to Roma in the Mediterranean world. By the 13th century they were documented in the Balkans, and by the 15th century they had spread across most of Western Europe, reaching Spain (1425 CE), England (~1500 CE) and Scandinavia.

This route is not merely a historical narrative: it leaves a layered genetic trace that remains partially visible in modern Roma populations and that the G25 model below attempts to decompose systematically.

II. Haplogroup Signatures of South Asian Origin

Before examining the autosomal G25 results, it is worth noting the uniparental haplogroup evidence, which provides the clearest possible signal of South Asian ancestry and is entirely independent of admixture modeling.

III. G25 Admixture Results by Diaspora

The G25 coordinates below were extracted from the Moriopoulos 2025 Modern Population Collection (no-simulations averages). Admixture modeling was performed using NNLS (non-negative least squares) in the G25 scaled space, with three macro-level source categories:

• South Asian (proxy: Gujarati, Jat_Haryana, Punjabi_Hindu_India, Sindhi), representing the ancestral Indian source population.
• Balkan / SE European (Greek Peloponnese, Greek Macedonia, Bulgarian, Romanian, Macedonian), reflecting Byzantine and post-Byzantine host admixture.
• Iberian (Spanish Andalucia, Extremadura, Cataluña), relevant for the Gitano populations of Spain and Portugal.
• Iranian / MENA (Iranian Persian, Iranian Zoroastrian, Armenian), the Caucasus and Persian passage layer.
• Turkish / Ottoman (Turkish Rumeli), Ottoman Balkan admixture visible in some groups.

You can run your own G25 coordinates against these Roma population averages or model Roma populations on Calculator #186, The Migration Era Calculator (Age of the Huns, Scaled) available on this site.

Note on the Serbian and Turkey/Balkans results: Both populations are based on a single individual (n=1), meaning individual-level variance may distort macro-component proportions. They should be interpreted with greater caution than the multi-individual averages.

Iberian Roma, The Gitanos

The Iberian Roma, collectively known as Gitanos, represent one of the most studied Roma communities in Europe. Their entry into Spain is historically documented to 1425 CE, when King Alfonso V of Aragon granted a safe-conduct to a group of Roma pilgrims. Over the following centuries they spread throughout the Iberian Peninsula. The genetic results below show that six centuries of Iberian residence have significantly increased their local admixture, yet their South Asian ancestry remains clearly detectable and sharply distinguishes them from all non-Roma Spanish populations.

IV. Comparative Table, All Roma Populations

V. Interpretation: What the Results Tell Us

The South Asian Persistence

The most striking result across all nine Roma populations is the consistent presence of a South Asian genetic component ranging from approximately 18% to 42%. This is remarkable for a population that has been resident in Europe for up to a millennium. For comparison, no other non-South-Asian European population, including Sephardic Jewish communities whose presence in Iberia pre-dates the Roma, shows any detectable South Asian autosomal signal.

The NNLS model consistently selects a Gujarati-like proxy to represent the South Asian component rather than more easterly or more purely South Asian populations (Bengali, Dravidian). This is consistent with what population genetics research suggests about the ancestral Roma homeland: a Northwest Indian population with relatively high Iranian-related ancestry, as found today among communities from Punjab, Rajasthan, and Gujarat, particularly among upper-caste and Rajput-related groups. Some researchers have specifically proposed Rajasthan as the most likely departure point.

The Balkan Layer, Byzantine Admixture

Across all Roma populations, including those now living in Iberia or Central Europe, a strong Balkan / Greek component forms the bulk of the non-South-Asian ancestry. This is not the result of recent mixing with Balkan populations, but rather reflects the extended period, roughly 200 to 400 years, that ancestral Roma spent within the Byzantine Empire before beginning their westward dispersal. The Greek Peloponnese and Greek Macedonia proxies being selected consistently is highly coherent with this history.

The implication is significant: most of what is “European” in Roma ancestry is Byzantine-Greek-Balkan, not the ancestry of the western European host populations they later settled among. A Roma individual in Seville carries far more Greek Peloponnese ancestry than Andalusian ancestry in the purely genetic sense.

The Iberian Gradient

Among Iberian Roma populations, a clear geographic gradient emerges. The Roma of Barcelona show the highest Iberian absorption (~49%), while those of Granada (~22%), Porto (~16%), and Bilbao (~17%) show lower but still significant Iberian components. This variation may reflect differences in the degree of endogamy practiced by different local Gitano communities, as well as possible differences in when or how they settled into particular regions.

Interestingly, despite being based in Andalucia, the region of Spain most associated historically with Gitano culture, Roma from Granada do not show the highest Iberian proportion. The Granada sample retains a notably large Iranian/MENA component (~10%), possibly reflecting greater preservation of the older pre-Iberian ancestry in this specific community.

The Iranian / Armenian Passage Layer

A Middle Eastern / Caucasian layer ranging from approximately 2% to 12% is detectable in most Roma populations. This “passage layer” is consistently captured by Armenian or Iranian Zoroastrian proxies, both of which genetically approximate the populations ancestral Roma would have encountered during their transit through the Caucasus and Persia. This component is most visible in Turkish Balkans (12.4%), Czechia (8.3%), Granada (10.1%), and Madrid (8.9%).

The survival of this signal after so many generations of European residence suggests either that it was large enough in the founding Roma population to remain detectable, or that some ongoing admixture with MENA-related populations (including Ottoman Turks and Middle Eastern traders) reinforced it during the Balkan centuries.

VI. G25 Coordinates, Roma Diaspora Populations

The following G25 coordinates are drawn from the Moriopoulos 2025 Modern Population Collection. Paste them as target populations in Vahaduo or use them as reference points when running your own analysis on Calculator #186.

VII. Key Takeaways

The genetic data paint a consistent and striking picture of Roma ancestry. Across all nine populations examined, spanning the Balkans, Central Europe, and the Iberian Peninsula, a set of converging signals emerges that is entirely coherent with the historical narrative of migration from South Asia through the Middle East into Byzantine Europe and then westward.

The South Asian component is the most diagnostic: present in every Roma population, ranging from ~18% in a single Serbian individual to over 40% in the least-admixed Balkan groups, and pointing consistently to a Northwest Indian source population with affinities to modern Gujarati, Punjabi, and Sindhi communities. This is entirely in line with what haplogroup studies have shown for decades: Y-DNA H-M82 and mtDNA M5a1 leave no room for ambiguity about the Indian origin of the Roma people.

The Balkan / Byzantine layer, consistently the largest non-South-Asian component in almost every population, tells the story of a long pause in the Byzantine world before the westward dispersal. Roma arrived in Western Europe already carrying the genetic imprint of Greece and the Balkans, not as pristine descendants of the Indian subcontinent but as a population that had already absorbed significant Balkan admixture over several centuries.

Finally, the Iberian variation, from ~17% in Porto to ~49% in Barcelona, documents the ongoing, variable process of local admixture that continues to differentiate Roma communities even within a single country. Yet even at the high end of this range, the South Asian signal never disappears: Barcelona Roma are still ~20% South Asian, a proportion that exceeds what any admixture calculator would ever return for any non-Roma Western European population.