Mixed Solvent Molecular Dynamics for Drug Discovery

What is it?

This protocol simulates the target of interest (GPCR, kinase, ion channel, etc.) in a solution of water and a co-solvent, like benzene or any other fragment-like molecule.

During the simulation, the co-solvent molecules interact with the surface of the protein, reveling binding hotspots. These hotspots have been proven to correlate very well with actual pockets [1].

Furthermore, binding modes for the probes can be extracted, which can prove very valuable, as the binding mode of bigger, drug-like molecules could mirror that of the probe [2].

We can run multiple, different fragments in parallel to conduct a virtual screening campaign.

If you are interested, you can ask us for more information or request a quote here .

What is the value of this service?

  • Binding site prediction: Identify the most likely binding sites of your target of interest, including orthosteric, allosteric and cryptic sites.
  • Binding mode prediction: Obtain binding mode predictions and pharmacophoric insights from your co-solvent molecule, which can be used to precisely guide a docking campaign.
  • Pocket ensemble: For each pocket, we will select the snapshots from the simulation where the pocket is in a "binding-ready" state. This ensemble can be used do model protein flexibiliy while docking.
  • Binding pathway: For each pocket, we will provide a trajectory linking the unbound state to the bound state.
  • One CrypticScout job typically takes only one week to finish (~10 replicas of 80 nanoseconds each). We know speed is key in your drug discovery project.

Pockets move, but molecular docking software ignores -for the most part- that flexibility. That's why it is worth it to dock your library into an ensemble of pocket conformations. The ensemble that we will provide you only includes pocket conformations with the co-solvent bound, so the pocket is always "opened".

What will you obtain?

Deliverables include:
  • A PyMol or VMD scene with binding hotspots and binding mode predictions.
  • The full simulations (.xtc and .pdb files).
  • The cloud of hotspots as a .cube file (orange wireframe in the image).
  • An extensive report summarizing the structural insights obtained.
  • Conference calls with the team to discuss the results.
  • For each identified pocket:
    • Binding mode prediction, which can be used to do pharmacophoric docking or scaffold-docking. (.pdb + .sdf)
    • Ensemble of druggable conformations, to model protein flexibility in docking (.pdb files)
    • Binding pathway of the co-solvent molecule from the bulk into the pocket (.xtc file)

Results of a CrypticScout job for a Ras protein (PDB code:4Q21). At high isovalues, the orange wireframe only highligths the pockets which the probe visited the most (or the longest). One of the identified pockets is the actual binding pocket of the endogenous ligand. The extracted pose for the chosen cosolvent (imidazole) overlaps very well with the actual ligand.


  1. Martinez-Rosell, G., Lovera, S., Sands, Z. A., & De Fabritiis, G. (2020). PlayMolecule CrypticScout: Predicting Protein Cryptic Sites Using Mixed-Solvent Molecular Simulations. Journal of Chemical Information and Modeling, 60(4), 2314–2324.

  2. Arcon, J. P., Defelipe, L. A., Modenutti, C. P., López, E. D., Alvarez-Garcia, D., Barril, X., Turjanski, A. G., & Martí, M. A. (2017). Molecular Dynamics in Mixed Solvents Reveals Protein-Ligand Interactions, Improves Docking, and Allows Accurate Binding Free Energy Predictions. Journal of Chemical Information and Modeling, 57(4), 846–863.

alejandroMixed Solvent Molecular Dynamics