AI-driven precision oncology (G2P/G2K frameworks). Building interpretable multi-omics and agentic-AI systems — iGenSig-AI for mechanism-driven in silico drug screening, and BRACE, an autonomous multi-agent framework that reasons as a cancer biologist to perform end-to-end genomic discovery from a dataset and a research goal. Best suited to students with strong computational and machine-learning interests.
Precision immuno-oncology biomarkers. Discovering genomic predictors of immunotherapy response for the many patients missed by PD-L1/TMB criteria — intragenic rearrangement (IGR) burden, tumor-associated antigen (TAA) burden, and immune-privileging regulon signatures — and developing them into clinical-grade assays. Combines computational discovery with clinical validation.
Uncharted cancer genetics. Mapping recurrent gene fusions (e.g., ESR1-CCDC170, BCL2L14-ETV6) and intragenic rearrangements as protein-altering driver events and defining their roles in cancer progression and immune evasion, with matched genotype-directed therapies advancing toward investigator-initiated trials. Best suited to students with both computation and experimental background.
Actionable kinase targets in refractory breast and ovarian cancer. Characterizing structural mutations in kinases such as EPHA3 and defining therapeutic vulnerabilities in aggressive, therapy-resistant disease. Best suited to students with experimental focus.