Leveraging Data Fusion Strategies in Multireceptor Lead Optimization MM/GBSA End-Point Methods

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• Accurate in silico modeling is crucial for efficient drug discovery lead optimization.
• End-point molecular mechanics/generalized Born surface area (MM/GBSA) calculations are widely used for affinity predictions.

Purpose of the Study:

• To evaluate data fusion strategies for leveraging MM/GBSA results across multiple receptors.
• To identify potent inhibitors from congeneric ligand sets using combined computational data.

Main Methods:

• Retrospective analysis of 13 congeneric ligand series from public data across seven biological targets.
• Application of MM/GBSA calculations and various data fusion metrics, including SumZScore.
• Comparison of SumZScore performance against single-receptor predictions and experimental binding affinities.

Main Results:

• MM/GBSA scores successfully identified potent inhibitor subsets in 90% of individual receptor structures.
• Data fusion strategies significantly enhanced prediction robustness.
• SumZScore proved to be a robust and physically meaningful metric for combining multi-receptor data.

Conclusions:

• Data fusion, particularly using the SumZScore metric, enhances the reliability of MM/GBSA predictions in drug discovery.
• SumZScore effectively prioritizes potent inhibitors even with modest correlations to experimental affinities.
• This approach improves the identification of promising drug candidates during lead optimization.