QMQSAR: utilization of a semiempirical probe potential in a field-based QSAR method

Area of Science:

• Computational Chemistry
• Medicinal Chemistry
• Drug Discovery

Background:

• Quantitative Structure-Activity Relationship (QSAR) models are crucial for drug discovery.
• Existing QSAR methods often require complex calculations or lack detailed physical interpretation.
• Developing efficient and interpretable QSAR models remains an active area of research.

Purpose of the Study:

• To develop a semiempirical quantum mechanical approach for creating molecular field-based QSAR models.
• To utilize probe interaction energy (PIE) values for characterizing ligand structures.
• To establish a computationally efficient method for predicting ligand binding affinity.

Main Methods:

• Employed a semiempirical quantum mechanical approach using PM3 calculations.
• Characterized aligned ligand structures by computing probe interaction energy (PIE) values at grid points.
• Built multilinear regression models using PIE values as independent variables, refined by simulated annealing.

Main Results:

• Developed a method to generate interpretable 3D physical models of ligand binding affinity.
• Successfully validated the approach against three literature datasets.
• Demonstrated the ability to model critical binding interactions in diverse chemical systems.

Conclusions:

• The described semiempirical quantum mechanical approach provides an effective strategy for QSAR modeling.
• The method offers an interpretable 3D physical model for understanding ligand-receptor interactions.
• This technique shows promise for accelerating drug discovery and development.