The debate over racial bias in genomics rages on, but are we avoiding a larger problem?

Genomics is failing on diversity. That’s certainly the case put forward by Popejoy and Fullerton in their recent op-ed in Nature. If you haven’t read the article yet, it’s worth a look. The authors look at the phenomenal bias towards European populations in genome-wide association studies. The exceedingly short version is: Europeans represent over 80% of participants in these studies, while Asian representation has increased over the past 7 years. This European bias also extends to direct-to-consumer genetic tests, as Manuel Corpas and Richard Shaw recently found.

Globe world

Image via Luke Price/Flickr

From a purely scientific perspective, this isn’t great. In essence, non-European data is treated as an anomaly. This presents two damaging scenarios. First, while any analysis is representative of Europeans, the results could be used to make healthcare decisions for non-Europeans. This is a particular problem when using pharmacogenetic testing to influence drug efficacy, side effects, and dosages.

The second scenario presents a huge missed opportunity. By not looking at the full breadth of human genetic variation, we could be missing some very useful insights simply because certain variants are uncommon in European populations.

Those are both things that people far more qualified than myself have written about. But what is the scientific community not talking about?

The answer comes in the solutions being proposed. By and large, everyone keeps saying that ‘we’ need to look at focusing ‘our’ attention on minority populations and incentivising that attention through grants.

Where in the world

Let’s dig a little deeper. Who here can list the 10 most populated cities on Earth?

world most populous countries racial bias

via upyernoz/Flickr

10) Istanbul – 14,160,467

9) Dhaka – 14,648,000

8) Kolkata – 15,552,000

7) Shanghai – 16,575,000

6) New York City – 18,406,000

5) Mexico City – 19,460,000

4) Mumbai – 20,041,000

3) Sao Paulo – 20,262,000

2) Delhi – 22,157,000

1) Tokyo – 30,000,000

I was surprised to see Tokyo top that list. And to be extremely honest, I’ve committed a crime I always warned my students against – I’ve used Wikipedia as a reference for those numbers. Wikipedia also has helpful demographics for the global population breakdown. Only 11.4% of the world’s population is based in Europe.

“Hang on, there are plenty of people in the USA from European descent!” I hear you cry. Correct. North America accounts for 7.6% of the world’s population. But the USA does not account for all of North America. A large chunk of that proportion is taken up by Mexico.

But you have to allow for immigration across the continents, as we are talking about ethnicity not location. These demographics are a useful guide, but not great for critical analysis. When you consider that the Han Chinese represent about 18% of the global population, the point still stands – those of European descent are a minority.

Building a truly global community

It’s time to take a global look at genomics, rather than a local look. Historically, people of European descent have taken a ‘paternalistic’ (to put it mildly) approach to their relationships with other races and ethnicities. I would argue that adopting this same approach will not solve the problem we now face, and will not resolve the bias. Rather than taking the ‘we know best’ approach to the problem of racial bias in genomics, we should be trying to understand what the underlying social and economic problems are that are driving this disparity.

Is there really a mistrust of medical testing at a community level? There are fantastic research institutions outside of the US and Western Europe, groups of people conducting innovative and exciting science, and very far from being sceptical of healthcare technology. So perhaps, the solution lies in improving cross-border, and cross-culture, collaboration. In every nation on Earth you will find the privileged and the under-privileged. In a predominantly Caucasian population, there will also be poorly educated Caucasians; lack of education is not a uniquely ‘minority’ problem.

So if we, all of us regardless of race, want to see that bias in the data corrected, let’s help researchers around the world find their counterparts in countries where Caucasians are a minority. And if they don’t have counterparts, find out why not and tackle that problem appropriately. Because as the population data reveals, really there is no such thing as a minority; we are one global population. Just as healthcare has often been accused of being too paternalistic, let’s not fall into the same trap when we discuss other people and their cultures.  

More on these topics