Summary
The convergence of the “$100 genome” and the AI revolution makes it possible to reimagine omics-driven precision care. Today’s NGS panels still leave 85–90% of cancer patients without an actionable finding. Our research program is conceived to close this gap—unlocking new discoveries from the same genomic data through AI-driven precision oncology. Our research program is distinguished by its unusual breadth and translational depth, spanning from AI-Genomics, Uncharted Cancer Genetics, Precision genetic markers, Immuno-Oncology Markers, to Omics-based Precision Oncology, which bridges the gaps between computational discovery, experimental validation, and clinical translation.
Our research program advances five synergistic pillars of innovation:
• Genomics to Precision care (G2P) AI: We pioneered an integral genomic signature framework (iGenSig, iGenSig-Rx; Nature Commun. 2022; BMC Bioinformatics 2024) for explainable multi-omics modeling, with iGenSig-AI extending this to mechanism-driven in silico drug screening for individualized therapy. We are also developing a G2P Agentic-AI system that autonomously assembles evidence-graded therapeutic options, clinical context, and biomarker-driven predictions to guide individualized care.
• Genomics to knowledge (G2K) agents. We built EnSEMBLE, an agent for enhancer-anchored pathway analysis that locks in enhancer-RNA–corroborated pathways from transcriptome sequencing data for biological validation, and BRACE-AI, an autonomous multi-agent framework that accepts omics data and a research goal, designs and executes iterative analyses, reasons as a cancer biologist would, self-validates code and statistics, grounds claims in PubMed, and delivers publication-grade reports.
• Precision Immuno-Oncology Biomarker Panels. We discovered intragenic rearrangement (IGR) burden as a predictor of ICB response in TMB-low tumors such as breast and ovarian cancers (Cancer Immunology Research 2024), the tumor-associated antigen (TAA) burden algorithm as a predictor of ICB response for PD-L1-negative tumors (Cancer Research 2012; Cancer Immunology Research 2024), and the IMPREG (immuno-privileging regulon) signature for predicting immunotherapy resistance, validated across 40 transcriptomic trial datasets (Nature Commun., 2026). These are being consolidated into our precision immuno-oncology panel to deliver actionable insights for the majority of patients overlooked by PD-L1 or TMB criteria.
• Uncharted Cancer Genetics for Precision Therapy. My lab identified the only canonical recurrent gene fusions described in common breast cancers—ESR1-CCDC170, BCL2L14-ETV6, and RAD51AP1-DYRK4—each matched to an effective genotype-directed targeted therapy, with DOD-funded translational work now advancing toward investigator-initiated clinical trials (Nature Commun. 2014; PNAS, 2020; Clin Cancer Res. 2021). In parallel, we are mapping intragenic rearrangements (IGRs) as a previously overlooked, second-most-prevalent class of protein-altering cancer aberration and defining their roles in breast cancer progression and immunotherapy resistance.
• Actionable Kinase Targets for refractory breast and ovarian cancers. We have characterized TLK2 as a key kinase target in aggressive luminal breast cancer (Nature Commun. 2016) and developed a dual p38/NLK inhibitor to target NLK-driven endocrine-resistant breast cancer (Clinical Cancer Res. 2021). Funded by a NCI R21 award, our ongoing work explores novel structural mutations in EPHA3 and other actionable kinases for refractory breast and ovarian disease.
Cancer Biology
