Sangseon Lee

spHOT identifies phenotype-associated spatial biomarkers in spatial transcriptomics by localizing disease-relevant niches across multiple diseases.

CAREPath improves drug repurposing by combining semantic path embeddings with mechanism-context augmentation in biomedical knowledge graphs, achieving better performance than existing methods.

This study introduces a context-aware hierarchical fusion model for drug relational learning, showing improved performance in predicting drug-drug interactions and side effects across different contexts.

CLOVER integrates multi-omics data with whole-slide images to generate molecular-aware slide representations, outperforming existing methods in cancer subtype classification and biomarker prediction.

EnsDTI improves drug-target interaction prediction by combining multiple models and providing confidence scores, achieving a 2.7% accuracy improvement over baselines.

This paper introduces a multi-layer guilt-by-association approach for drug repurposing by integrating clinical knowledge into biological networks, improving drug-disease association prediction accuracy.