Watching @AdamRutherford chat with @toadmeister on twitter about racism and genetics reminds me again of the communication problem around sensibly talking around genetics, genetic ancestry and race - which are, counterintuitively, surprisingly separate concepts.
Race is complex mixture of visible physical characteristics (skin colour, hair type, facial features) mixed with many cultural signals (dress, accents, gestures); most people will self classify themselves confidently and other people
Although most people will acknowledge the continuum of people ("mixed race") most people are confident in the major groupings and assign labels, use heuristics around them for good and bad (bad more common than good!)
The visible characteristics - skin colour, hair type, facial features - are largely genetic, and so children are usually a mixture of their parents. *BUT* - and this is the key point - it is a very small part of the genome which determine these features - perhaps 2 or 5%
Because humans *continuously* mix in human history, using this 2-5% of the genome which one can assess visually (and actually less good than one realises - the cultural stuff is also key) is just not a very good way to tell you about the bulk of the genome
This is most obvious in well mixed cases of human populations where we have documented history about the mixing - Latinos (European / Native American), Liverpool Black British (European / West African), African American (ditto), Afro-Caribbean (ditto + a tiny bit of Carib)
Many of the most recent examples of obvious mixing is due to the slave trade, which also has had a complex cultural evolution post the abolition of slavery, in particular in the US, post Civil war, and then post Civil rights movement.
But these are just the most recent cases of mixing, with a dark, documented history; there are many others - Cape coloured (Portuguese / South African); the complex mixing across the Indian subcontinent; 'white' Russians into Siberia
Furthermore even the current populations which we considered well mixed now were not previously; in particular "Europeans" is mixture of 3 quite different human groups which we can track through ancient DNA sequencing
So - the story of humans is endless, endless mixing. We operate as one species and when we meet other members - we nearly always have children which are mixed. We are still doing this now - most obviously in Brazil, in London, in New York and Paris but actually everywhere.
When a geneticist uses the phrase "genetic ancestry" they are looking for "well mixed" populations in terms of genetics - there is a whole bunch of theory+practice that relies on this "well mixed" property.
The first thing every human geneticist learns is that "self reported" ethnicity just doesn't mean "member of a well mixed population" - it is correlated but nowhere near enough for you to use as a proxy, and in any case, if you have the genetics, use that directly
The standard thing is to exclude people with more complex genetic mixtures - you might think these people are "mixed race" but many - most - are not. They classify themselves into ethnicities confidently. It just that "genetic ancestry" and "race" aren't the same-
The common mistake therefore is reading things about "genetics" (eg, genetics and heart attacks, or genetics and educational attainment) and thinking that one can swap "genetics" with "ethnicity". Just not true. Really not true
There are big, complex societal things to discuss about this research (see previous threads by myself) but oddly enough we can really very confident that genetics can't be part of the explanation of why we see different distribution of some ethnicities say in higher education
(Yup - I know this seems counter-intuitive). The two key points are (a) the number of loci involved in these studies - eg, educational attainment - is much much greater than skin colour/hair type
(b) Some of the populations with the most differential education attainment as a group - eg, Afro Carribeans in UK - are from a European/African admixed population. Both "parent" populations - west african and european have better aggregate educational attainment
The property of mixing here where the mixed population isn't in the middle of the two parents is absolutely not consistent with a genetic explanation. A far simpler explanation is cultural factors, eg, discrimination
This area is complex to communicate; For example, David Reich's piece in the New York times, written by one of the best population geneticists around managed to imply this correlation of ethnicity with ancestry meant it is ok to have these labels.
I've not talked this through with David, but I read this piece about "we shouldn't be afraid about talking about groups", not that "you can use ethnicity as a proxy for genetics" - indeed, I know David doesn't do the latter.
Personally I think we should talk about this more, and also do more studies of (say) the British Caribbean (there are good studies of the Spanish Carribbean) to help understand both history and do good genetics for healthcare
This goes with an overall aspect that health focused genetics will be far better - more equitable and better biology - as projects such as the great @H3Africa_CC and other more diverse human populations are studied.
