Research and training Program
Our multidisciplinary research and training program is designed to bridge the computational and experimental aspects of cancer genetics and precision oncology research. I would like to summarize our key achievements and highlight our current frontline research below:
- Recurrent Gene Fusions as Genetic Markers for Breast Cancer Precision Oncology: Through an integrative bioinformatics approach (Nature Biotech 2009, Cancer Discovery 2012), we identified ESR1-CCDC170, BCL2L14-ETV6, and RAD51AP1-DYRK4 as precision biomarkers for aggressive breast cancer subtypes matched with targeted therapies (Nature Commun., 2014, PNAS 2020, Clinical Cancer Res. 2020). To date, these are the only canonical gene fusions identified in common breast cancer forms.
- Revealing an Uncharted Area in Cancer Genetics: Our most recent research suggests that intragenic rearrangements as the dark matter of cancer genetics, could be the second most frequent class of genetic aberrations, following simple mutations. We have identified a substantial number of recurrent intragenic rearrangements (IGRs) across various cancer types, which could substantially enhance next-generation sequencing (NGS) panels in precision oncology. We are exploring the oncogenic roles of these IGRs in breast cancer progression, therapy resistance, and immune evasion.
- Integral Genomic Signature-based Precision Oncology Modeling: We developed a novel class of machine learning methods tailored for multi-OMIC data, known as "integral genomic signature analysis," which predicts responses to targeted therapy and chemotherapy (Nature Commun., 2022). We are now advancing this with iGenSig-AI, a method that can perform in silico drug screening to identify the most effective drugs for individualized cancer therapy.
- Precision Genomic Markers for Immuno-Oncology: Another central objective of our research is the development of precision genomic biomarkers aimed at predicting the efficacy of immune checkpoint blockade (ICB) in tumors that exhibit low tumor mutation burden (TMB) and/or reduced PD-L1 expression. Among these biomarkers, the intragenic rearrangement burden (IRB) has emerged as a particularly promising candidate for forecasting ICB responses in tumor types with low TMB infiltration (Cancer Immunology Res., 2024). Supported by funding from a Department of Defense Breakthrough Award, we are advancing a whole-genome sequencing (WGS)-based IRB assay specifically designed to predict ICB responsiveness in cases of triple-negative breast cancer (TNBC).