I’m going out of my comfort zone on this one, not being a population geneticist, I would welcome being educated on this.

I’m having interesting/challenging conversations about definitions of race/ethnicity/ancestry (REA) in medical genomics as part of @hail_CSER @NYCKidSeq 1/

2/ Specifically, how REA information might help us to improve genetic diagnostics in a project designed to study rare diseases and cancer.

Possibly surprisingly, it’s difficult to make a strong case for the value of REA information for de novo/somatic mutations

3/ Plus there’s the complexity of defining REA as a single category per patient when people have such complex ancestries, especially in a diverse county like the #Bronx

And we do not want to end up with REA information in some way promoting bias or doing harm.

4/ Puzzling through this in my own head, our typical REA categories (Black, white, Hispanic etc) are pretty useless when making a diagnosis. Sub-Saharan African ancestry inherently means incredibly diverse, why lump this all together in one group?

5/ So I started thinking like a clinical geneticist. What matters in making a diagnosis in a clinical setting? Not REA as much as family information. If a distant cousin also has a similar phenotype, you start wondering about something shared within the family.

6/ Which brings me to an area of inexpertise for me, population genetics. Having learned that IBD doesn’t necessarily stand for inflammatory bowel disease, but also identity by descent, you can see how IBD can reconstruct populations as large families. https://www.nature.com/articles/ncomms14238

7/ What’s the Utopian future for REA information in genetics diagnostics? Can we abandon gross, kind of useless categories in favour of IBD information? How would the IBD information be collected and presented when needed? What risks would it carry?

8/8 I am learning about this stuff, but there should be a value in naive questions, so I will welcome getting taught by colleagues. It’s timely that we think about the value of REA in medical genomics, so we use this kind of information only for positive purposes.