Bo Sun

The paper develops an interpretable deep learning model to predict cancer patient survival using drug prescriptions and personal transcriptomes, achieving high accuracy and identifying pathways that influence survival.

The paper identifies genetic variants influencing NMD regulation across human tissues and finds they are tissue-specific, enriched in gene regions and regulatory elements, and colocalize with disease SNPs.

The paper identifies transcriptional trajectories in human kidney disease progression using machine learning and validates findings in a mouse model.