Junjie Zhu

DynaFold generates protein dynamic trajectories using a latent diffusion model, achieving high accuracy and efficiency in predicting conformational ensembles and transitions.

StruCloze reconstructs atomistic structures from coarse-grained models and inpaints missing regions in biomolecules, achieving state-of-the-art accuracy and transferability.

DynaRNA generates RNA conformation ensembles using a diffusion model with EGNN, outperforming MD simulations in accuracy and efficiency.

This paper designs 10,000 de novo GLP-1RAs using AI, with 60 passing stability, efficacy, and diversity criteria, and two showing extended half-lives and superior efficacy in vivo.

This paper introduces DeepReweighting, a deep learning framework for optimizing molecular force fields with improved efficiency and accuracy compared to traditional methods.

This paper introduces IDPFold, a diffusion model that generates conformational ensembles for intrinsically disordered proteins directly from sequences, without requiring MSA or experimental data.