Receptor prep alone used to take — half a day per target
Without receptor prep scripts, box definition, job submission, or manual PLIP analysis after the fact. Submit SMILES and a PDB ID. The platform handles everything — and validates every pose against strain energy.
“Dock this lead series against EGFR (PDB: 1M17) and flag any poses with high strain energy.”
How it works
Submit SMILES and a PDB ID
Provide up to 100 molecules and a protein target. The platform handles receptor preparation, protonation state (configurable pH), box definition, and job submission.
Docking with pose validation
AutoDock-GPU runs on Azure GPUs. Every pose is analyzed by PLIP for binding contacts — hydrogen bonds, hydrophobic contacts, pi-stacking. Strain energy via GFN2-xTB flags artifact poses (>5 kcal/mol).
Ranked results, ready for MD
Binding affinities, contact residues, distances, and strain energy per molecule — rendered inline. Top candidates feed directly into molecular dynamics without file conversion.
Proof
AutoDock-GPU on Azure Container Apps. Reference ligand co-docking for box definition. Configurable protonation pH (1–14, default 7.4).
PLIP protein-ligand interaction profiling. dock_with_strain GFN2-xTB validation — high strain (>5 kcal/mol) indicates the docking score may be an artifact.
Two-phase workflow: Phase 1 returns cost estimate + confirmation token. Phase 2 executes after user approval. Max 100 molecules per batch.
Binding, validated
AutoDock-GPU. PLIP contacts. Strain energy. Sign up and dock your first molecule.