Boltz-ABFE: Free Energy Perturbation without Crystal Structures.

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• Free energy perturbation (FEP) is the gold standard for binding affinity estimation.
• FEP accuracy relies on precise protein-ligand complex structures, often unavailable experimentally early in drug discovery.
• Existing methods are limited by the need for experimental crystal structures.

Purpose of the Study:

• To develop a robust pipeline (Boltz-ABFE) for absolute binding free energy (ABFE) estimation without experimental crystal structures.
• To assess the utility of predicted protein-ligand complex structures for FEP simulations.
• To enable structure-based affinity estimation in early-stage drug discovery.

Main Methods:

• Integration of the Boltz-2 structure prediction model with an absolute FEP protocol.
• Development of automated methods to refine predicted structures for molecular dynamics simulations.
• Validation of the Boltz-ABFE pipeline using four protein targets from the FEP+ benchmark set.

Main Results:

• Boltz-2 successfully predicts protein-ligand complex structures suitable for FEP.
• Automated structure refinement improves the quality of predicted models for simulations.
• The Boltz-ABFE pipeline accurately estimates ABFE for multiple protein targets without experimental structures.

Conclusions:

• Boltz-ABFE demonstrates the feasibility of performing FEP simulations using predicted structures.
• This approach significantly broadens the applicability of FEP in drug discovery.
• Boltz-ABFE accelerates early-stage drug discovery through accurate, structure-based binding affinity estimation.