Our lab focuses on developing computational methods to understand complex biological systems. We combine machine learning, systems biology, and experimental validation to tackle challenging problems in biology and medicine.
Research Directions
Machine Learning for Biology
How can we develop new machine learning methods tailored for biological data?
We develop deep learning architectures and algorithms specifically designed for biological sequences, structures, and networks. Our methods handle the unique challenges of biological data including noise, missing values, and complex dependencies.
Key areas:
- Graph neural networks for molecular property prediction
- Transformer models for biological sequences
- Generative models for molecular design
- Interpretable AI for biological insights
Systems Pharmacology
How do drugs interact with biological systems at multiple scales?
We use computational approaches to understand how drugs affect cellular networks and pathways. By modeling drug-target interactions and their downstream effects, we can predict both therapeutic outcomes and side effects.
Key areas:
- Drug-target interaction prediction
- Polypharmacology analysis
- Adverse effect prediction
- Drug repurposing
Single-Cell Analysis
What can we learn from analyzing biology at single-cell resolution?
We develop methods for analyzing single-cell sequencing data to understand cellular heterogeneity, differentiation trajectories, and cell-cell communication.
Key areas:
- Cell type identification and annotation
- Trajectory inference
- Multi-modal data integration
- Spatial transcriptomics
Precision Medicine
How can we personalize treatments based on individual molecular profiles?
We combine genomic, transcriptomic, and clinical data to predict treatment responses and identify biomarkers for patient stratification.
Key areas:
- Biomarker discovery
- Treatment response prediction
- Patient stratification
- Clinical decision support
Current Projects
Project 1: Deep Learning for Drug Discovery
We’re developing new deep learning models that can predict drug-target interactions and optimize molecular properties. This project combines graph neural networks with chemical knowledge to improve drug design.
Project 2: Multi-omics Integration
This project focuses on integrating multiple types of biological data (genomics, proteomics, metabolomics) to build comprehensive models of cellular states and disease mechanisms.
Project 3: Interpretable AI for Biology
We’re working on making machine learning models more interpretable for biological applications, allowing researchers to understand not just predictions but also the biological mechanisms behind them.
Collaborations
We actively collaborate with experimental labs, clinicians, and industry partners. If you’re interested in collaboration, please contact us.
