Cheminformatic Analysis and Machine Learning Modeling to Investigate Androgen Receptor Antagonists to Combat Prostate Cancer | JoVE Visualize

Area of Science:

Medicinal Chemistry
Computational Chemistry
Pharmacology

Background:

Prostate cancer (PCa) is a significant cause of cancer mortality in males.
Androgen receptor (AR) antagonists are crucial therapeutic targets for PCa treatment.
Understanding the chemical space and structure-activity relationships (SARs) of AR antagonists is vital for drug discovery.

Purpose of the Study:

To systematically analyze the chemical space, scaffolds, and SAR landscape of human AR antagonists using cheminformatics and machine learning.
To identify favorable scaffolds and key structural features for potent AR antagonists.
To provide insights for hit identification and lead optimization in developing novel AR antagonists.

Main Methods:

Systematic cheminformatic analysis of 1678 AR antagonist molecules.
Physicochemical property visualization and principal component analysis (PCA) for chemical space exploration.
Murcko scaffold analysis to assess scaffold diversity and identify representative scaffolds.
Quantitative structure-activity relationship (QSAR) modeling and structure-activity landscape visualization.
Identification of activity cliff (AC) generators.

Main Results:

Potent AR antagonists generally exhibit smaller molecular weight, logP, nHA, nRot, and TPSA compared to inactive molecules.
Low scaffold diversity was observed, with potent molecules showing even lower diversity, highlighting the need for novel scaffolds.
Sixteen representative Murcko scaffolds were identified, with scaffolds 1, 2, 3, 4, 7, 8, 10, 11, 15, and 16 being highly favorable.
A QSAR classification model achieved high accuracy (0.756 on test set) for predicting AR antagonist activity.
Seven significant activity cliff generators were identified, providing valuable SAR information.

Conclusions:

The study provides a comprehensive analysis of the chemical space and SAR landscape of AR antagonists.
Favorable physicochemical properties and specific scaffolds are associated with potent AR antagonist activity.
The identified scaffolds and activity cliffs offer valuable guidance for the rational design and optimization of novel AR antagonists for prostate cancer therapy.