Measuring cancer behaviors: We study how tumors occupy a finite spectrum of gene expression states (“archetypes”), each associated with distinct treatment responses and metastatic potential. Our work aims to identify and explain these states within the framework of normal tissue biology.

Targeting cancer behaviors: We develop mathematical tools to identify state-specific cancer vulnerabilities. In particular, we use the principle of maximum entropy to generate explainable, data-driven cancer models. This enables the computational identification of key driver genes and potential drug targets.

Overcoming treatment resistance: We build tools to model the evolutionary dynamics of cancer. Our goal is to predict when therapy resistance emerges and how it may create secondary vulnerabilities, opening up new therapeutic windows over time.

Multimodal data integration: We are developing clinically accessible strategies to monitor treatment response and patient risk longitudinally. Our approach integrates radiomic data, liquid biopsies, and AI-driven forecasting to capture tumor evolution in real time.

Translation: We actively collaborate with clinical and experimental partners to move our discoveries toward practical applications in patient care.