Structural Biology & Drug Design

Molecular Docking: Predict Protein–Ligand Binding
Poses with AutoDock Vina and GLIDE

BioMate automates the full molecular docking pipeline — protein structure preparation from AlphaFold or PDB, grid definition, docking with AutoDock Vina or GLIDE, and interaction analysis — returning ranked binding poses with docking scores and interaction fingerprints. No manual file preparation required.

Try free → See all structural biology services

Pipeline steps

From sequence or PDB to ranked binding poses

Every step is automated and containerized. Provide a protein identifier and ligand structures — BioMate handles the rest.

Structure preparation BioMate fetches the protein structure from PDB by accession, or runs AlphaFold 2/3 if no experimental structure exists. Side chains are completed, protonation states set at physiological pH, and waters removed.
Binding site definition BioMate identifies the active site from known ligand positions, literature annotations, or fpocket cavity detection. The docking grid is centered automatically on the binding pocket.
Ligand preparation SMILES strings or SDF files are converted to 3D structures with RDKit, tautomers enumerated, and partial charges assigned. Up to thousands of ligands processed in parallel on AWS Batch.
Docking and scoring AutoDock Vina or GLIDE runs exhaustive pose sampling. Top-ranked poses are rescored with interaction fingerprinting (hydrogen bonds, hydrophobic contacts, pi-stacking).
Results and visualization Ranked pose table (docking score in kcal/mol), 3D binding pose (SDF/PDB), interaction fingerprint heatmap, and a downloadable methods report with all parameters cited.

Input	Accepted formats
Protein structure	PDB accession, PDB file, UniProt ID (triggers AlphaFold), mmCIF
Ligand structures	SMILES string(s), SDF file, compound name, ChEMBL ID
Binding site	Residue list, known ligand coordinates, or auto-detected by fpocket

Output	Description
Binding poses	Top-N poses in SDF format, ranked by docking score
Docking score	Estimated free energy of binding (ΔG, kcal/mol)
Interaction fingerprint	Per-residue H-bond, hydrophobic, electrostatic contacts
Methods report	DOCX with software versions, parameters, and citations

Virtual screening at scale

Screen thousands of compounds in a single run

BioMate parallelizes docking across AWS Batch workers, enabling high-throughput virtual screening against compound libraries. ADMET pre-filtering removes high-liability compounds before docking to reduce compute cost.

Fragment screening

Dock fragment libraries (e.g. Enamine REAL Fragment, DSI-Poised) against a novel target pocket to identify starting points for fragment-based drug design.

Lead optimization

Dock analogue series to rank substituent effects on binding affinity. Pair with ADMET profiling to identify compounds with improved potency and drug-like properties.

Allosteric site screening

Use fpocket to identify cryptic or allosteric pockets, then dock against multiple binding sites simultaneously to find allosteric modulators.

Covalent docking

Screen covalent warheads against reactive cysteines or lysines using CovDock. BioMate handles reactive group enumeration and covalent binding geometry validation.

FAQ

Common questions about molecular docking in BioMate

What docking software does BioMate use?

BioMate supports AutoDock Vina (open-source, Scripps Research) and GLIDE (Schrödinger) for rigid and flexible docking. The tool is selected based on throughput requirements and licensing.

Can I dock against an AlphaFold-predicted structure?

Yes. BioMate runs AlphaFold 2 or AlphaFold 3 to predict the target structure, then feeds it directly into the docking pipeline — no manual file preparation required.

What input do I need for molecular docking?

A protein structure (PDB file, AlphaFold model, or UniProt ID) and ligand structures (SMILES strings, SDF file, or compound name). BioMate handles all format conversions.

What does the docking output include?

Predicted binding pose (SDF), docking score (ΔG in kcal/mol), interaction fingerprint (hydrogen bonds, hydrophobic contacts), and pLDDT-weighted confidence for the binding site residues.

Molecular Docking: Predict Protein–Ligand BindingPoses with AutoDock Vina and GLIDE