Race, Ethnicity, clinical research and general life. A reminder/explainer. TL;DR: Uncouple "ethnicity" and "genetics" - there are threads that link them but they are not useful (certainly not useful for genetics) - and do not consider ethnicity as some proxy for genetics.

Genetics and ethnicity are two very different things: 1. the set of check boxes, sometimes census based, sometimes what the person designing the form decided on "ethnicity" (UK terms) or "ethnicity" split from "race" (US terms), to none at all due to the law (the French approach)

For largely historical reasons, rooted in the 18th+19th century thought and wrapped up in the justification of both colonisation and slavery, much of the implicit definition of who ticks these boxes is about skin colour.

2. the genetics of an individual - formally 3 billion letters of DNA which some 300,000 vary "relatively frequently" and another 1 million or so are rare / very rare changes (sometimes unique to you).

Skin colour genetics is a small subset of this variation - it is actually quite a few variants concentrated in a small set of genes (loci) - perhaps 30 or so places (loci; some ~100,000 letters) around the genome and many variants in this region

(This is a bit odd for a human trait; most traits with many variants have the variants spread across more genes. In the genetics parlance there is a lot of allelic heterogeneity).

Despite the obvious certainty of skin colour - Person A has dark skin, Person B has light skin - with gradations of tone which is largely genetic (~ sun exposure), skin colour is not a good proxy *at all* for genetics. It just... doesn't work this way.

The opposite is far more true - we can use genetics to predict skin colour (though it is far more complex than you think)

Furthermore, because in many western societies we've had recent enforced (eg, slavery) or voluntary (eg, post colonial) migration we can use recent genome ancestry *for a subset of the population* to predict which box they will tick (eg, "White British" or "Afro-Caribbean").

It's only a subset of people ticking the box "White British" we can do this recent ancestry inference on - some ~15% of people who tick "White British" box we can't do this inference on.

As groupings get bigger (just numbers) and each generation mixes more since these migrations, this ability will fall away, but more importantly the fact that we can go from genetics to predicting the box they tick on an ethnicity form *does not mean* we can do the reverse

This is most obvious in the "Hispanic" / "Latino" groupings where there are hugely different ancestry patterns which also just merge with many people who don't tick these boxes (as expected - humans are one species, and we ... fall in love and have sex in all sorts of ways).

We need to remind people again and again - despite the confidence we have in ticking our own "ethnicity" box, and the gestalt confidence we have in what boxes other people would tick (perhaps misguided confidence) due to skin colour - this tells us almost nothing about genetics.

For many people this feels weird. It feels weird because skin colour is clearly genetic (but in fact this is a small part of the genome). It feels weird because we've elevated this construct of the 18/19th century into an important part of our identities.

Thought experiment: we could have elevated hair colour as a salient and important thing. Blond haired people should hang out with other blond haired people. You would be surprised to see a blond haired child with dark haired parents.

Perhaps we'd ask people with red hair "where do they come from originally" or "which part of Ireland is your family from". It would be pretty trying for the red haired person in Italy endlessly fielding these questions, but ... if we decided this was important...

In medicine and clinical research we often get ethnicity terms and these are important, but they are important in my view as social labels - they represent the structuring of our society due to this world from the 19th century.

Do not mistake ethnicity terms as somehow equivalent to genetics. They are not "perfectly random" wrt to genetics - lactase deficiency *is* rare in people who tick "White British" - sickle cell disease is far more common in people who tick "Black African"

But it would be a category error to not diagnose sick cell because someone's skin colour was not "black". That's... not how this works!

And the business of moving, mixing and having sex is not a modern thing. All through history this has happened, most notably "modern Europeans" are a complex mixture of 3 groups which we can infer both from archaeology and genetics

One are the hunter-gatherers that hunted in Europe through the ice ages and as it thawed. Another are Anatolian farmers, who swept into Europe with this amazing new technology ("farming") - this group of people also went east into Levant and South down the eastern side of Africa

The third was the Yamnaya culture - noted for their "Corded ware" - a style of pottery - and likely brought pastoral farming schemes.

These migrations are not just about Europe; Anatolina farmers going across the Mahgreb (northern africa) and down the East African coast - Yamnaya going out to Siberia. And our knowledge continues to grow in particular as ancient genomics grows out from European sampling sites.

This is just one migration of many - of the complex population of the Americas; of the "Bantu migration" in Africa; of the ship based population of Polynesia; of the Portuguese colonies/trading posts at the Southern Cape, Goa and elsewhere; of slavery and indentured service

Uncouple "ethnicity" and "genetics" - there are threads that link them but they are not useful (certainly not useful for genetics) - and do not consider ethnicity as some proxy for genetics. It simply is not.