qpAdm vs G25: the genetics reopening Iberia's map

qpAdm vs G25: the genetics reopening Iberia's map

A debate on a population genetics forum has peine a discussion that seemed buried: how much North African ancestry there is in the Iberian Peninsula, when it arrived, and why the northwest comes off worst in the identity narrative. The thesis circulating is that Galicians and Portuguese have the highest North African contribution in Europe, that this contribution is not medieval but Roman, and that the northeast of the peninsula remains more "pure". The problem is that the tool doesn't say that. It says what the model allows it to say.

What qpAdm measures and what G25 measures

qpAdm is a tool from the admixtools suite that models ancient populations with standard errors and p-values. G25 is a hobbyist calculator that returns whatever you ask for depending on the sources you feed it. The technical difference is real: one gives confidence intervals, the other gives pretty decimals. But neither turns a statistical model into a certificate of purity.

The argument repeated most often is that the North African contribution in Iberia is largely pre-Sarracinic. Olalde 2019 already pointed to North African flow in the imperial era, especially along the Mediterranean arc, through slavery, auxiliaries and trade. It's a trickle, not a flood. The big peak, argue the less militant analyses, is still al-Andalus. Anyone who doesn't want to see that is choosing the model that suits them.

The Celtic component doesn't exist in DNA

This is where half the narrative falls apart. A genetically separable Celtic component does not exist. There is Atlantic Bronze Age ancestry shared by Galicians, Asturians, Bretons and half of Portugal. Celtic is a language and a culture, not a chromosome. And the North African gradient in Iberia doesn't draw a clean border between northwest and northeast, but a map with more nuance than fits in a tavern headline.

The discussion soon drifted to the Moriscos who were not expelled from Murcia, the Xuetes of Mallorca and the role of the Inquisition. The Xuetes are the example that shatters the clean assimilation narrative: fifteen surnames, forced conversion and brutal endogamy for centuries. Adams et al 2008, in the American Journal of Human Genetics, found in them paternal lineages from the Levant and North Africa that don't match the Mallorcan surroundings.

Roman North Africa was not today's

One of the arguments that appears in the debate is that North Africa in Roman times was genetically more similar to Europe than it is today. With the Sarracinic expansion from the 7th century onwards and the trans-Saharan slave trade, the region received sub-Saharan and Arab flow it hadn't had before. The North African DNA that reached Galicia in Roman times was the original Berber component, free of later mixtures.

That doesn't turn Galicians into Scandinavians, but neither into Rifians. The Iberomaurusian base already had a sub-Saharan component, less than now, and a Berber from those centuries would look more like a present-day Canary Islander than a Segarro of today. The discussion about phenotypes —blondes in Biscay, dark-haired in A Coruña— is background noise: eye colour doesn't measure ancestry.

What the model doesn't settle

The conversation ended where these things usually end: with a table of reference populations still pending and a claim nobody has demonstrated. That Galicians exceed 10% non-ancient North African contribution is a figure thrown out without a source. That the Portuguese are the most North African in Europe, another one. qpAdm can give p-values, but it doesn't give permission to turn a gradient into a purity ranking.

With all this, the population genetics of Iberia still hasn't resolved the question everyone asks and nobody formulates: how much of what we call identity is ancestry and how much is narrative. The data is in the papers. The conclusion, in whoever reads them.