All Services

Molecular Docking

Predict how your compounds bind to a protein target: prepare the receptor and ligands, dock with three engines, see every interaction, and make the ranking robust with replicate, multi-engine and ensemble docking — then hand the best complexes to molecular dynamics.

3docking engines
AutoDock-GPU · Vina · smina
3 × A30GPUs
1.5 Åre-docking accuracy
example below
PDFreports with every interaction

From a structure to a binding hypothesis

One job folder carries the receptor, the ligands and all results from step to step.

Receptor
PDB download or your model; chains, ligands, docking box and grid maps
Ligands
files, PubChem, screening hits or peptide sequences → 3D, minimised
Docking
AutoDock-GPU on GPUs; AutoDock Vina and smina on CPUs
Analysis
poses, clusters, complexes, interactions, images, reports
Beyond
molecular dynamics and MM/GBSA binding free energies

What we do

Preparation

Receptor & Ligand Preparation

  • Structures from the PDB by ID, or your own models (homology, AlphaFold)
  • Choice of chains and co-crystallised ligands; waters removed; hydrogens and charges added
  • Docking box around a bound ligand, chosen residues or the whole protein; AutoGrid4 / AGFR grid maps
  • Ligands from SDF / MOL2 / PDB, PubChem (names, SMILES, CIDs) or virtual-screening hits; 3D coordinates and MMFF94 minimisation
  • Peptide ligands built from their sequence
Receptor editor: chains, co-crystallised ligands and 3D view
Receptor editor for cathepsin S (PDB 2FQ9): chains, the bound inhibitor and the grid options.
The docking box around the bound ligand
The docking box: ligand extent plus 10 Å.
Review before docking
Readiness check before docking, for each engine.
Docking

Three Docking Engines

  • AutoDock-GPU — AutoDock4 scoring, many independent runs per ligand clustered into binding modes, on three NVIDIA A30 GPUs
  • AutoDock Vina and smina — empirical scoring on the CPUs, reproducible with a recorded seed
  • All engines search the same box, so results can be compared and combined
  • From a handful of ligands to large libraries
Docking engine selection and settings
Choosing the engine and its settings for a docking job.
Docking scores of six ligands with two engines
Six ligands docked into cathepsin S with AutoDock-GPU and AutoDock Vina.
Analysis

Poses, Interactions & Reports

  • Pose selection by cluster population or lowest energy
  • Protein–ligand complexes for every ligand and engine
  • Interactions with PLIP: hydrogen bonds, hydrophobic contacts, π-stacking, salt bridges, halogen bonds, water bridges
  • 3D images of the complex and the pocket, 2D interaction diagrams, interaction tables
  • Excel summaries and a publication-style PDF report
Binding pocket and 2D interaction diagram of a docked ligand
Relacatib in the cathepsin S pocket: 3D view and 2D interaction diagram.
Complexes and batch analysis
Complexes ranked by score, with batch 2D, 3D and interaction analysis.
Docking PDF report builder
The PDF report: images and interactions of every complex.
Robust ranking

Consensus Docking

  • Statistical consensus — AutoDock-GPU at two search depths (SP: 50 runs, XP: 100 runs with 10× longer searches), each in replicates: mean ± SD, weighted consensus score, t-test and effect size, plots
  • Multi-engine consensus — AutoDock-GPU, Vina and smina ranked together by mean rank, rank sum or exponential consensus ranking
  • Complexes and interactions of the consensus top hits; PDF reports
Statistical consensus ranking
Statistical docking: best SP and XP scores and the consensus per ligand.
Multi-engine consensus results with interactions
Multi-engine consensus: interactions of each ligand and engine, with a 3D view.
Receptor flexibility

Ensemble Docking

  • Several receptor conformations: representatives of an MD simulation, crystal / NMR structures, or an uploaded trajectory
  • All conformations superposed, one common box
  • Rankings by best, mean, population-weighted and robustness scores; binding modes compared across conformations
Ensemble docking ranking
Ranking over three cathepsin S structures.
Score heatmap ligands × conformations
Score of every ligand in every conformation.

Example: cathepsin S

A worked example from our documentation with public data (PDB 2FQ9, inhibitors from PubChem).

1.47 Å
the crystal ligand re-docked into its own structure by AutoDock Vina (heavy-atom RMSD; AutoDock-GPU 2.59 Å)
6 ligands
four cathepsin inhibitors, the crystal ligand and a non-binder control, docked with three engines
rank 1
relacatib ranked first by AutoDock-GPU, Vina and smina alike; the non-binder control last in the consensus
3 structures
ensemble docking into 2FQ9, 1NPZ and 1MS6 confirmed the ranking

What you receive

  • Docking scores of every ligand, cluster and engine (Excel / CSV)
  • Protein–ligand complexes (PDB / PDBQT) of the best poses
  • 3D and 2D interaction images and interaction tables
  • PDF reports — docking, statistical consensus, multi-engine, ensemble
  • A Methods & parameters record — software versions, boxes, settings and seeds
  • Selected complexes ready for molecular dynamics on the same platform

Infrastructure

  • 3 × NVIDIA A30 GPUs (AutoDock-GPU)
  • 2 × Intel Xeon Gold 6342 (96 threads) for Vina / smina
  • 384 GB RAM
  • AutoDock-GPU · AutoDock Vina · smina · AGFR · MGLTools · Open Babel · PLIP

Ready to dock your compounds?

Register and request a docking study — or contact us to discuss a research collaboration.

Get Started Contact Us