SQM2.20: Semiempirical quantum-mechanical scoring function yields DFT-quality protein-ligand binding affinity...

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate protein-ligand binding affinity estimation is crucial for computer-aided drug design.
• Existing methods may be computationally expensive or lack accuracy.

Purpose of the Study:

• To introduce SQM2.20, a novel, universal, physics-based scoring function.
• To provide a computationally efficient yet accurate tool for binding affinity prediction.

Main Methods:

• Utilized semiempirical quantum-mechanical methods to model binding free energy terms.
• Developed the SQM2.20 scoring function incorporating recent advancements.
• Compiled the PL-REX benchmark dataset with high-resolution structures and experimental affinities.

Main Results:

• SQM2.20 demonstrates superior performance compared to other scoring methods.
• Achieved excellent correlation (average R² = 0.69) with experimental data on the PL-REX dataset.
• Showed accuracy comparable to computationally intensive DFT calculations but with significantly reduced runtime.

Conclusions:

• SQM2.20 offers a computationally efficient and accurate solution for binding affinity prediction.
• Its speed makes it suitable for practical applications in hit identification and lead optimization.
• The PL-REX dataset serves as a valuable resource for validating scoring functions.