Like many others I’ve been curious about DNA matching, especially with regards to my own ancestry. Could a simple swab tell me where my ancestors lived?

Where am I from?

I asked this question of FamilyTreeDNA, one of many such services. Using a 12 marker test, it turns out that I’m 4.6% Irish and 2.5% New Zealander. But I’m also 2.4% from Trinidad and Tobago and, surprise… 12.5 % Faroe Islands. All of these numbers are higher than for the US (1.1%), even when including Native American (.6%).

If I wanted to have a simple narrative of my ancestry, these numbers don’t seem to help. I suppose I could say that my main origin is Faroese, with a little Irish, but my ancestors must have cavorted in New Zealand and the Caribbean as well.

The situation worsens when I use a different number of markers, Y-DNA, mitochondrial DNA, haplotrees, or other measures. It’s messy.

The jigsaw puzzle of ancestry

It may be disappointing to hear, but for anyone who has similar questions, the underlying science is deeply flawed. That’s true even for someone whose data seems more definitive than mine.

These tests are interpreted to fill in a jigsaw puzzle of who we are; 12.5 % Faroe Islands would be a big chunk of mine. I’d be a little bit of this and a little bit of that, an admixture. This is a term used to describe the process of gene flow between distinct populations. It can be fun to see the scores, but as applied to human populations, the whole idea is wrong.

It relies on assumptions of homogeneity among people living within national boundaries, mapping of ethnicities onto countries and of races onto continents, and racialist assumptions about pure categories. The idea that we can just fill in the jigsaw puzzle of our ancestry doesn’t work.

We’re not admixtures

Kostas Kampourakis and Erik Peterson discuss the science of this in a recent article published in Genetics: The racist origins, racialist connotations, and purity assumptions of the concept of ‘admixture’ in human evolutionary genetics (2023).

The article shows how admixture of the kind an ancestry detective would like requires the existence of pure or unadmixed categories. 

There are two big assumptions:

(1) particular populations existed before the colonization era, have not undergone significant admixture, and have maintained their genetic variation.

(2) the people included in the reference datasets have four grandparents all born in the same country, implying they are somehow pure representatives of that country.

But these assumptions are ahistorical and regularly violated. They don’t account for migrations due to colonization, wars, or plagues. And archeological evidence shows that humans were migrating around the globe long before the age of colonization. 

In addition to being a problem in the field of genetics, admixture leads to people saying things like “I just learned that I’m a big part Faroese, with a little Irish and Caribbean thrown in!” which is nonsense scientifically.