Accurate Binding Free Energy Predictions in Fragment Optimization.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Predicting protein-ligand binding free energies is crucial for structure-based drug design (SBDD).
• Accurate binding free energy predictions can reduce drug development costs and timelines.
• Fragment-based drug design requires precise prediction of binding affinities for small molecules.

Purpose of the Study:

• To assess the suitability of the free energy perturbation protocol, FEP+, for calculating relative binding strengths of fragment-sized compounds.
• To validate FEP+ performance against experimental data for pharmaceutically relevant targets.

Main Methods:

• Utilized the FEP+ protocol, employing explicit solvent molecular dynamics and physics-based scoring.
• Applied the method to over 90 fragments across several pharmaceutically relevant targets.
• Compared FEP+ predictions with experimental binding affinity data.

Main Results:

• FEP+ accurately predicted relative fragment binding affinities.
• Achieved R(2)-values greater than 0.5 when compared to experimental data.
• Obtained overall RMS errors of approximately 1.1 kcal/mol.
• Demonstrated significant improvement over docking and MM-GBSA methods.

Conclusions:

• The FEP+ methodology is suitable for predicting relative binding affinities of fragment-sized compounds.
• FEP+ shows the predictive power necessary to impact fragment-based drug discovery and affinity optimization projects.
• This computational approach offers a reliable tool for accelerating lead discovery and optimization in SBDD.