Our lab is interested in developing sophisticated machine learning approaches to extract useful information from the large-scale multi-omics data to understand the complex disease such as cancer. Our research covers several important topics in cancer transcriptome, spanning from technique-driven research that aims at developing graph-based learning models for cancer transcriptome analysis with prior knowledge (e.g., isoform quantification, biomarker identification, cancer outcome prediction, drug sensitivity prediction), to hypothesis-driven investigation of specific biological problems (e.g., changes of transcriptome upon mTOR hyper-activation). Our development leads to novel computational models and molecular signatures, which could be used in early detection, diagnosis, and prognosis of specific tumors.

List of Projects