Predicting the Binding Affinity of ERβ Ligands Based on a Novel Variable Selection Method.

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Pharmacology

Background:

• Estrogen receptor β (ERβ) plays a crucial role in various physiological processes.
• Developing selective ERβ ligands is important for therapeutic applications.
• Understanding structure-activity relationships is key to designing effective ligands.

Purpose of the Study:

• To develop a robust quantitative structure-activity relationship (QSAR) model for ERβ ligands.
• To identify key molecular descriptors influencing ERβ binding affinity.
• To establish a predictive model for novel ERβ ligand discovery.

Main Methods:

• Characterization of 128 ERβ ligands using molecular descriptors.
• Development of a QSAR model employing a variable selection method based on variable interaction.
• Internal and external validation of the QSAR model.

Main Results:

• A QSAR model with five descriptors was successfully developed.
• High determination coefficient (R²) of 0.8272 and cross-validated correlation coefficient (Q²) of 0.8041 were achieved.
• External validation yielded an estimated correlation coefficient of 0.8255, confirming model reliability.

Conclusions:

• The developed QSAR model accurately predicts the binding affinity of ERβ ligands.
• The model adheres to OECD principles, enabling its use for predicting novel ERβ derivatives.
• Mechanistic interpretation of descriptors provides insights into ligand-target interactions.