Viking DNA test: what a DNA test can and cannot tell you about Viking ancestry

A Viking DNA test sounds like a simple product: send in a raw file, get back a Viking percentage. The trouble is that no such number can exist, because the people the word describes were not a genetic population. What a DNA test can honestly do is measure how much your genome resembles published Viking Age individuals, model it from Viking Age Scandinavian sources, and show the stretches of DNA you share with specific people from the period. This guide explains what each result means, and what none of them can mean.

"Viking" was a job, not a people#

In Old Norse, a víkingr was someone who went on a víking, a season or a career of raiding and trading by sea: something a person did, not something a person was born as. Ancient DNA has now confirmed the consequence. Margaryan and colleagues sequenced 442 people from the Viking Age and the centuries around it, and the individuals buried with Scandinavian rites turned out to be genetically varied. Two men in Orkney, buried in Scandinavian fashion, had ancestry indistinguishable from present-day Irish and Scottish people. Individuals in Denmark and Sweden carried substantial southern European ancestry; some in Norway carried Sami-related ancestry. Scandinavia itself was regionally structured, with related but distinguishable Danish-like, Norwegian-like and Swedish-like profiles. Our full reading is in Viking DNA: what 442 ancient genomes reveal.

The consequence for testing is blunt. There is no Viking allele, no Viking haplogroup and no Viking reference population. There are Viking Age Scandinavian reference genomes, which is a different and far more useful thing, and a test that sells a "Viking percentage" is selling a label for a Scandinavian-related ancestry estimate. Read the estimate, not the label.

What a Scandinavian-like signal means#

Suppose a report finds part of your genome best explained by a Viking Age Scandinavian source. Three readings are open, and a genome cannot decide between them.

The first is descent from Scandinavians who left during the Viking Age: Danish settlers in the Danelaw, Norwegian settlers in Orkney, Dublin, Iceland and Normandy, Swedish traders on the eastern rivers. The second is descent from the same North Sea populations at a different time. The Angles, Saxons and Jutes who settled England from the fifth century came from a coast running from the northern Netherlands through northern Germany into Denmark, and Gretzinger and colleagues found their ancestry overlaps heavily with what Danish settlers brought later. Separating the two by autosomal ancestry alone is at the edge of what any method can do. The third is plain geography: Scandinavians, the Dutch, northern Germans and the eastern English form one continuum, so a northern European genome resembles Viking Age Scandinavians partly because it resembles their neighbours.

Rodríguez-Varela and colleagues added a further twist in 2023: the Viking Age brought British-Irish, eastern Baltic and southern European ancestry into Scandinavia, and modern Scandinavians carry less of that inflow than the Viking Age population did. The reference genomes are not purer versions of today's Swedes and Danes; the deeper history is in Scandinavian DNA: ancient origins. A result is a resemblance to sampled people, not a certificate of membership.

Three methods, three kinds of answer#

Ancestrify runs three analyses on one raw file, each asking a different question of the Viking Age.

qpAdm modelling. A formal admixture model offers your genome a set of ancient sources and asks whether it can be written as a mixture of them, with a p-value for the fit and a standard error on every weight. For a British or Irish customer the natural model contrasts a Viking Age Scandinavian source with an Iron Age British source and an early medieval continental one. The honest output is often a weight with a wide error bar, or a nested model in which the Scandinavian source is not needed at all, because the three sources are close relatives. That is the method telling the truth about how similar those populations were. The qpAdm analysis prints the p-value and every standard error so the trade-off can be seen.

Global25 distance. With a coordinate row, you can measure the distance from your position to published Viking Age individuals and population averages, labelled in our catalog by region and period: Danish Viking, Norwegian Viking, Swedish Viking, English Viking, Iceland Norse-Gaelic and others. Distance is a similarity measure, not an ancestry proportion. A modern Norwegian sits close to the Norwegian Viking Age samples because little has changed since; a modern English person sits at a moderate distance from all of them. Read the cluster, not the single winner, in the Global25 analysis.

Ancient Matches. The third analysis skips populations altogether. It scans your raw file against every individual in the ancient panel, including the Viking Age individuals published by Margaryan and later studies, and reports the stretches you share with each, painted on your chromosomes with the evidence on every row. Sharing a segment with a man from the Salme ship burial or a woman buried in Viking Age Denmark is identity by state: your genome and theirs read the same across that stretch. It is the most concrete of the three results and the most easily over-read, which is why every row says what it is and why Ancient Matches never names anyone as your ancestor.

The haplogroups sold as Viking#

Three Y-chromosome lineages carry the label in marketing copy: I1, R1a-Z284 and R1b-U106, and each is genuinely common in Scandinavia. I1 reaches roughly a third of men in Sweden and around a quarter in Norway and Denmark; R1a-Z284 is the Scandinavian branch of R1a, concentrated in Norway and Sweden; R1b-U106 is the main R1b branch of the Germanic-speaking world, from the Netherlands and northern Germany to Denmark and England.

None of them is diagnostic, because a haplogroup is a frequency, not a proof. I1 is carried by around one man in seven in England and is common across northern Germany and the Low Countries. Gretzinger and colleagues found I1 absent from sampled Iron Age and Roman Britain and present in more than a third of early medieval English men, so an English I1 lineage most likely arrived with the Anglo-Saxon migration, three centuries before the first raid on Lindisfarne. The same holds for U106. R1a-Z284 is the most Scandinavia-specific of the three, and even it says only that a paternal line once lived in Scandinavia, not when it left.

The free Clade Finder reads your Y and mitochondrial branches from the raw file and shows where each is common today. It will not, and cannot, tell you that anyone on it was a Viking.

Reading a result from where you are#

• England and Scotland. Expect a Scandinavian-related component that trades off against the Anglo-Saxon-era continental one. Leslie's study of 2,039 people found no clear signal of the Danish occupation in its regional clusters, while Orkney stood out with roughly a quarter Norwegian-related ancestry.
• Ireland. Norwegian-related ancestry entered through Dublin and the other ports; the genome-wide share is small, and a match to a Viking Age individual from Ireland is a resemblance to a mixed, mobile port population.
• Normandy. The Norse settlers of 911 were absorbed within a few generations; the modern Norman genome is French with a faint northern tilt. A large Scandinavian weight in a Norman file is more likely a modelling artefact than a memory of Rollo.
• Iceland. The founders were Norse and Gaelic together, and the Norse share rose after settlement. An Icelandic result is Viking Age ancestry almost by definition, and still says nothing about occupation.
• The United States and the diaspora. With mixed northern European ancestry the Scandinavian-related and British-related weights blur. A Swedish great-grandparent produces a measurable signal; a Viking one does not, because that is not a category the genome carries.

A checklist before you buy#

1. Does the report name the ancient individuals and populations it compared you with, or only a label? Named samples can be checked; "Viking" cannot.
2. Does every ancestry weight come with a standard error or another statement of uncertainty?
3. Are the sources ancient genomes from the Viking Age, or modern Scandinavians relabelled?
4. Does the test say plainly that a haplogroup is a frequency? If it calls your haplogroup Viking, stop reading.
5. Where is your raw file processed? Ancestrify runs in the EU, one-time, and does not sell a percentage it cannot define.

Frequently asked questions#

Can a DNA test tell if I have Viking ancestry?#

It can measure resemblance to published Viking Age individuals, model your genome with Viking Age Scandinavian sources and show shared segments with specific people from the period. It cannot show that an ancestor went raiding, because Viking describes an occupation and the Viking Age population that practised it was genetically mixed.

Which haplogroups are Viking?#

None. I1, R1a-Z284 and R1b-U106 are common in Scandinavia and marketed as Viking, but each is found across northern Europe, and in England I1 and U106 arrived mainly with the Anglo-Saxon migration. A haplogroup tells you where a single paternal line is common today, not what any of its bearers did.

How much Viking DNA do English people have?#

A Scandinavian-related contribution exists, concentrated in the old Danelaw counties, and the summary of the 442-genome study put the Viking-related share of the UK population at up to about six percent. Gretzinger's early medieval genomes show why the figure is soft: the continental ancestry that reached England with the Angles and Saxons overlaps with what Danish settlers brought later, and no autosomal method separates them cleanly.

What is the best Viking DNA test?#

One that names its ancient reference individuals, publishes the uncertainty on every number and refuses to call any of it Viking. Ancestrify's three analyses do that from one raw file: a formal model with error bars, a distance to Viking Age samples, and segment matches to the individuals themselves.

Terms used here are defined in the glossary.

Sources and further reading#

1. Margaryan, A. et al. (2020). Population genomics of the Viking world. Nature, 585, 390 to 396.
2. Rodríguez-Varela, R. et al. (2023). The genetic history of Scandinavia from the Roman Iron Age to the present. Cell, 186, 32 to 47.
3. Gretzinger, J. et al. (2022). The Anglo-Saxon migration and the formation of the early English gene pool. Nature, 610, 112 to 119.
4. Ebenesersdóttir, S. S. et al. (2018). Ancient genomes from Iceland reveal the making of a human population. Science, 360, 1028 to 1032.
5. Leslie, S. et al. (2015). The fine-scale genetic structure of the British population. Nature, 519, 309 to 314.