Rapid, Accurate, Ranking of Protein-Ligand Binding Affinities with VM2, the Second-Generation Mining Minima Method

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• Structure-based drug discovery relies on methods to rank small molecule drug affinities for proteins.
• Existing methods include fast but less accurate docking and slow but accurate explicit solvent free energy methods.

Purpose of the Study:

• To benchmark the accuracy and computational speed of the mining minima technology, specifically the VeraChem Mining Minima Generation 2 (VM2) code.
• To compare VM2's performance against established computational methods on protein-ligand benchmark datasets.

Main Methods:

• Systematic benchmarking of VM2 on two protein-ligand datasets.
• Analysis of VM2's performance with varying run settings, including crystallographic water molecules.
• Evaluation of computational costs on Amazon Web Services (AWS) instances.
• Assessment of the impact of a generalized Born to finite-difference Poisson-Boltzmann correction.

Main Results:

• VM2 achieves accuracy comparable to explicit solvent free energy methods at a significantly lower computational cost.
• The study identified optimal run settings and quantified computational expenses.
• The generalized Born to finite-difference Poisson-Boltzmann correction consistently improves VM2's accuracy.

Conclusions:

• VM2 offers a distinctive technology for early-stage drug discovery.
• VM2 provides a powerful combination of computational efficiency and predictive accuracy.
• This method represents a valuable advancement in computational drug design.