The objective of the project is to construct a pan-cancer atlas of immune cellular neighborhoods and their spatial assembly. Previous multiplexed imaging studies have developed frameworks to map the spatial organization of individual tumors in smaller cohorts. In addition, foundation models have yielded machine learning models that can extract histology features that predict clinical outcomes and molecular characteristics. We will apply a foundation model to extract image features from whole-slide imaging data from the Cancer Genome Atlas (TCGA), and apply the tissue schematics framework to find cellular neighborhoods and map their spatial assembly. By leveraging the clinical, gene expression and mutation data in TCGA, we can decipher the functional mechanisms and impact of this assembly.

I am participating in this SuperUROP because I am interested in research that has the intersection of computer science and biology. This project excites me because of its medical applications to cancer. Taking on this project will not only help me understand this specific subject better, but also develop my research and communication skills.