Investigators

Summary

In my role as leader in AI for Cancer Research, I work closely with Hillman PIs and Hillman leadership to develop the AI infrastructure and capability to advance clinical operation, clinical research, and basic cancer research. I envision building a robust AI capability at Hillman that can meet the AI needs for cutting-edge cancer research and be adapted to address new challenges. My current cancer-related research includes:

1. Mechanism of infection and oncogenesis by KSHV
Goals: Using a combination of bioinformatics/machine learning, high throughput profiling (scRNA-seq, in-situ-seq, 16s-seq, etc), and bench experiments to delineate the mechanism of KSHV-induced cellular transform and oncogenesis, elucidate the pathogenesis of KSHV-associated cancers, and identify effective therapeutic targets and prognostic biomarkers. Through a long-term collaboration with Dr. SJ Gao, we have identified targets and pathways regulated by KSHV miRNAs (Nat. cell Biology, 2010), delineated the addicted cellular genes and networks by genome-wide CRISPR-Cas9 screening (MBIO 2019), performed the first genome-wide viral and cellular m6A profiling in multiple KSHV-infected systems (Nat. Microbiology 2018), and identified the signatures of oral microbiome in HIV-infected individuals with oral KSHV (PLOS PATHOGENS, 2019).

2. m6A mRNA modification/epitranscriptome and cancer
Goals: 1) Develop bioinformatics/machine learning tools to facilitate the functional study of m6A mRNA methylation and cancer. Using a combination of bioinformatics/machine learning and high throughput profiling technologies to 1) understand the mechanisms by which m6A regulates cancer and viral infection; 2) identify m6A-related clinical makers. m6A/epitranscriptome is a new and rapidly advancing area that studies modifications in mRNAs. My lab leads the development of computation tools for analyzing m6A profiling data and predicting m6A functions. The analysis pipeline for m6A sequencing, exomePeak, has been used widely and cited > 400 times (Google Scholar) by many high impact papers in Cell, Cell Stem Cell, Nature, Nature Cell Biology, Nature Neuroscience, Nature Genetics, and Cancer Cell. Using these tools, we have uncovered the reprogramming of viral and cellular m6A epitranscriptome during the life cycle of Kaposi sarcoma-associated herpesvirus (KSHV) (Nature Microbiology, (2018) 3(1):108-120) and reported common and distinct m6A regulation of innate immune response during bacterial and viral infection (Cell Death & Disease 2022, 13(3)). Besides, I have collaborated with other cancer biologists to uncover a cross-talk among m6A writers, erasers, and readers that regulates cancer progression (Science Advances, 2018 4(10)) and linked the increased activity of ALKBH5 with dysregulation of histone ubiquitination in cancers (Cancer Research, 2022/5).
 
3. Single-cell spatially-resolved transcriptomics (scSRT) analysis
Goals: 1) Develop the analysis and visualization pipeline for scSRT data(Nanostring CosMx, 10X Xenium, and Vizgen MERSCOPE), 2) Develop AI/machine learning tools for modeling tissue structure and disease pathology, and 3) apply scSRT to study immune microenvironment associated with tumor and viral infection. We have recently applied scSRT to reveal the molecular and immune signatures as well as pathological trajectories of fatal COVID-19 lungs (JMV, 2023)

4. Functional interpretable deep learning models and large language models for cancer genomics
Goals: 1) Develop novel deep learning models capable of cancer phenotype predictions and explain the underlying mechanisms. 2) ChatGPT-based prompt engineering for extracting knowledge about molecular regulatory mechanisms from literature. We have developed several genomics-based deep learning/AI tools for cancer prognosis and survival analysis, drug response prediction, and gene dependence prediction (Science Advances, 2021; Cancers, 2023).