Thread by @societyforepi

#SERjournalclub paper 3 is all about who gets included in our research.

@SarahTishkoff and colleagues discuss the biased sampling of human genetic studies and how this can lead us to miss important variation & harm public health.

cell.com/cell/fulltext/…

This paper was a bit challenging of a read for me, because it’s targeted at a genetics audience & includes a lot of specific genetic examples.

But the main crux of the argument is that by not including the full range of human diversity in genetic research we do harm.

So first, what diversity is missing?

Practically all of it! This figure from the paper shows the ancestry category distribution of people with catalogued genetic information: 78.4% European.

Reminder, only about 12% of the world’s total population is “white”.

So why is this bad?

Well for starters, we are probably missing huge amounts of information by not sampling more widely.

There’s more genetic diversity *within* ethnic groups than between them!

But more than that, genes dont act alone. Instead, a person’s other genes & their environment can interact in unexpected ways!

So, even when we identify a genetic variant in white people that’s also relevant for other groups, the *effect* of the variant might differ.

Another problem? Genetic databases are increasingly being used to develop polygenic risk scores (PRS) with the aim of identifying people are high risk to health issues *before* they get sick...

...but risk scores built using genetic databases will almost by definition work best for white people, since that’s whose information is used to build them. So, even if these scores work for their intended purpose, they can *worsen* health disparities & lead to *more* inequality.