Our Vision
We aim to create a computational bridge between genome, proteome, and immunome — to reveal the molecular complexity of tumors and guide the next generation of personalized cancer therapies.
Research Focus
Advancing computational proteogenomics for precision oncology
Proteogenomic Integration for Tumor Antigen Discovery
We explore how somatic mutations, alternative splicing, and non-coding region translations give rise to novel peptides that may serve as tumor-specific antigens. By integrating whole-genome sequencing, RNA-seq, and mass spectrometry-based proteomics, we systematically identify and prioritize candidate neoantigens. Our recently developed tool, pXg, enables precise and scalable detection of variant peptides.
Personalized Protein Databases
To better capture individual tumor heterogeneity, we construct customized protein sequence databases directly from patient-derived genome (VCF) and transcriptome (GTF) data. This allows comprehensive detection of non-reference peptides arising from mutations, fusions, or non-canonical translation events — enabling patient-specific proteomic analysis.
Computational Methods and Machine Learning
Our lab leverages computational modeling, statistical learning, and AI-driven inference to interpret large-scale multi-omics data. We develop algorithmic frameworks for peptide-spectrum matching, FDR control, and feature-based ranking. We are also exploring deep learning approaches for peptide identification and immunogenicity prediction.
Translational Proteogenomics
By bridging molecular signatures with clinical phenotypes, we aim to translate computational findings into actionable insights for cancer therapy. This includes identifying druggable targets, biomarkers, and immune-relevant neoantigens that can inform immunotherapy and vaccine development.