Predictive power of molecular dynamics receptor structures in virtual screening

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Molecular dynamics (MD) simulations are crucial for understanding protein dynamics.
• Assessing MD-generated conformations for blind virtual screening (VS) using docking remains underexplored.
• Protein flexibility influences ligand binding and requires accurate conformational sampling.

Purpose of the Study:

• To critically analyze the predictive power of MD snapshots for blind VS.
• To evaluate the utility of MD-generated conformations in docking-based VS.
• To compare VS performance using MD snapshots versus crystal structures.

Main Methods:

• Utilized unrestrained MD simulations to generate protein conformations.
• Performed blind VS using docking on selected protein systems (ligand-bound and unbound).
• Compared VS results with experimentally determined structures.

Main Results:

• MD snapshots consistently improved VS predictive power over crystal structures.
• The enhanced performance is attributed to MD's ability to sample relevant receptor conformations.
• MD simulations made previously undockable conformations amenable to VS.

Conclusions:

• MD simulations provide valuable conformational ensembles for enhancing VS.
• Incorporating MD-generated conformations can significantly improve drug discovery pipelines.
• This approach expands the scope of targets amenable to structure-based drug design.