Ligand Efficiency Outperforms pIC50 on Both 2D MLR and 3D CoMFA Models: A Case Study on AR Antagonists

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Ligand efficiency (LE) is crucial in drug design, quantifying biological activity per molecular size.
• Traditional metrics like pIC50 are widely used but may not fully capture structural contributions.
• Surface efficiency index (SEI) has emerged as a promising alternative LE metric.

Purpose of the Study:

• To compare the performance of pIC50 and SEI as dependent variables in structure-activity relationship (SAR) studies.
• To evaluate the influence of endpoint selection on predictive modeling of androgen receptor antagonists.
• To determine if SEI offers advantages over pIC50 in drug discovery optimization.

Main Methods:

• Utilized 2D Multiple Linear Regression (MLR) and 3D Comparative Molecular Field Analysis (CoMFA).
• Investigated SAR of androgen receptor antagonists.
• Employed both pIC50 and SEI as dependent variables in the modeling.

Main Results:

• SEI demonstrated superior performance compared to pIC50 in both MLR and CoMFA models.
• Models using SEI exhibited higher stability and predictive ability.
• Analysis suggests SEI better reflects the inherent relationship between molecular structure and bioactivity.

Conclusions:

• SEI is a more rational parameter for optimization in drug discovery than pIC50.
• The choice of endpoint significantly impacts SAR modeling outcomes.
• SEI provides a more nuanced assessment of ligand efficiency for drug design.