The ability to predict gene function has far-reaching applications, and recent research shows that gene function can be inferred from experimentally derived associations of other genes. These networks of associations are formed via metrics such as physical binding and genetic interaction. The Berger lab recently detailed a novel method to combine the information of many networks by learning gene embeddings based on patterns in individual networks, and then applying standard machine learning methods to infer gene function. The goal of my project is to improve this method by embedding the functional labels of gene networks in hyperbolic space. With hierarchical representations, I am to create a tool that is valuable for biological and translational research.

“My excitement about computational biology research at MIT is what initially inspired me to apply to this school, and with SuperUROP, I actually have the opportunity to pursue interdisciplinary research with guidance from some of the leading experts in the field. I’ m looking forward to being able to use cool math and machine learning techniques to hopefully make an impact in the world of health technology.”