Thanks to advances in big data, health experts and researchers can glean genetic insights from large populations of people. To do this, they conduct what are called genome-wide association studies (GWAS). GWAS rely on databases that house huge amounts of information. However, most of the information in these databases comes from people of European ancestry — a feature that doesn’t translate well if you’re studying other populations.
Researchers from the Broad Institute of MIT and Harvard and Massachusetts General Hospital demonstrated this phenomenon in a recent paper in Nature Genetics. The team analyzed information from one of the largest publicly available genetic data sets, the UK Biobank. Specifically, they looked at polygenic scores, which can predict someone’s risk for developing conditions like breast cancer or Type 2 diabetes. They found that the scores had a higher prediction accuracy for people of European descent than those of African, Asian or Hispanic/Latino backgrounds. This find, the authors say, highlights the importance of increasing the genetic diversity of these databases.
[This article originally appeared in print as "The Power of Representation."]