Jintao Zhu

DiffTernary is a diffusion model for predicting molecular glue ternary complex structures, achieving competitive performance on benchmark datasets.

CellChem learns cell-guided molecular representations that integrate structure and function, improving drug discovery by capturing cellular response similarities.

This paper evaluates co-folding methods for predicting ternary complexes involving molecular glues, finding AlphaFold3 performs best but relies on memorization rather than accurate modeling.

BioMiner automates extraction of protein-ligand bioactivity data from literature, achieving an F1 score of 0.32 and improving downstream model performance by 3.9%.