Advances in genetics research have improved our ability to understand disease risk, but current prediction models often fail to capture the diversity of human populations. Polygenic scores, which estimate genetic risk, are usually trained on data from individuals of European ancestry and perform poorly for others. My project evaluates the inclusive polygenic score (iPGS) model, a machine learning approach designed to address this gap. iPGS has shown superior performance for non-European individuals in the UK Biobank. In the project’s first phase, I will extend this work by testing the model in the U.S.-based All of Us dataset, comparing performance across traits and ancestry groups. The second phase will be to refine the model, contributing to advancing equitable genomic medicine.

I am participating in SuperUROP to build on my previous work in statistical genetics and bioinformatics and to gain more experience working with large-scale data and computational research methods. I am looking forward to learning more about the publication process and contributing a paper based on this project. What excites me most is that my project advances genetic risk prediction in a way that is more inclusive and clinically useful.