
PocketFlow
Generate ligands inside target binding pockets
PocketFlow generates drug-like molecules inside protein binding pockets with an autoregressive flow model that incorporates chemical knowledge, capturing inter-molecular and intra-molecular dependencies simultaneously. It produces 3D molecules with high binding affinity and valid geometries directly in the pocket context. Used for structure-based hit generation and lead optimization around a defined site.
PocketFlow is being onboarded — request access and be first in line.
At a glance
- Input
- Protein pocket (PDB)
- Output
- Generated ligands (SMILES, SDF)
- Developed by
- Sichuan University
- Published
- Jiang et al., Nature Machine Intelligence 2024 · 2023
#structure-based-design#molecule-generation#flow-model
Related tools
More methods in Ligand & Drug Design


