Impact of geometry optimization methods on QSAR modelling: A case study for predicting human serum albumin binding...

Area of Science:

• Computational chemistry
• Medicinal chemistry
• Drug discovery

Background:

• Quantitative structure-activity relationship (QSAR) modeling is crucial for predicting biological activity.
• Understanding the influence of quantum chemical calculation methods on QSAR model performance is lacking.
• Molecular descriptors are key inputs for QSAR models, but their calculation can vary.

Purpose of the Study:

• To comprehensively analyze the quantitative effects of different geometry optimization methods on molecular descriptors.
• To investigate the influence of descriptors derived from various calculation methods on QSAR models for drug binding affinity to HSA.
• To propose a validated QSAR model and a rational approach for selecting calculation methods.

Main Methods:

• Employed semi-empirical, ab initio Hartree-Fock, and density functional theory methods for geometry optimization.
• Calculated molecular descriptors based on optimized geometries.
• Developed and validated QSAR models using experimental binding affinity data to human serum albumin (HSA).

Main Results:

• Different geometry optimization methods significantly impact molecular descriptor values.
• QSAR models built with descriptors from various methods showed varying predictive performance.
• A 4-descriptor QSAR model predicting log K_HSA was developed, adhering to OECD validation principles.
• The model demonstrated predictive capability on an external dataset.

Conclusions:

• The choice of quantum chemical calculation method critically influences QSAR model development and predictive accuracy.
• A rational, activity-independent approach for selecting geometry optimization methods is recommended for improved QSAR modeling.
• The proposed QSAR model serves as a valuable tool for drug development, aiding in predicting drug binding affinity to HSA.