Chengsheng Mao

This paper introduces gANCHOR, a T-cell foundation model that predicts CAR T-cell immunotherapy response by integrating gene-level structure with clinical outcomes, achieving high F1 score and biological insights.

The paper develops an AI framework to analyze blood scans and predict patient outcomes by characterizing tumor-immune ecosystems in circulation.

SOAR provides a curated spatial transcriptomics database and analysis tools to facilitate disease mechanism studies and drug discovery.

This study introduces DeepCues, a deep learning model that improves cancer type classification using DNA sequencing data, achieving 77.6% accuracy on TCGA data and showing significant improvement over conventional methods.