Spatial transcriptomics is a molecular profiling method that uses mRNA expression data to assign cell types to their locations in histological sections. This allows researchers to measure all gene activity in a tissue sample and map where that activity is occurring. The goal of this project is to apply spatial transcriptomics to a clinical trial being run on ovarian cancer patients. There has been variability in the outcomes of patients: some experience complete remission while others experience recurrence. We believe that, by encoding spatial information in a graph and combining it with expression data and using a graph neural network to get an output representation, we can shed light on this variability by detecting patterns differentiating the two groups of patients.

I am participating in SuperUROP because I want to gain further experience with computational biology, especially from a machine learning lens. I am excited to apply the knowledge I gained from 6.047 and 18.418 to my project and to leverage a method as powerful and relevant as spatial transcriptomics.