Systematic identification of target set-dependent activity cliffs.

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Activity cliffs (ACs) represent critical data points in drug discovery, highlighting subtle structural changes that significantly impact compound potency.
• Understanding ACs is essential for optimizing lead compounds and predicting potential off-target effects.
• Existing methods for AC identification often do not fully account for compound set-dependent potency distributions or confirmed inactive compounds.

Purpose of the Study:

• To establish a comprehensive knowledge base of novel activity cliffs (ACs).
• To define ACs based on compound set-dependent potency distributions, incorporating confirmed inactive compounds.
• To facilitate the identification and analysis of structure-activity relationships.

Main Methods:

• Rationalized and applied various AC definitions, representations, and search criteria.
• Analyzed medicinal chemistry and biological screening data for nearly 100 different target proteins.
• Identified target set-dependent ACs using computational and data-mining approaches.

Main Results:

• Successfully identified numerous target set-dependent activity cliffs across diverse protein targets.
• Generated a freely accessible knowledge base containing over 20,000 target set-dependent ACs and associated data.
• Demonstrated the feasibility of incorporating confirmed inactive compounds into AC identification.

Conclusions:

• The developed knowledge base provides a valuable resource for researchers in drug discovery and medicinal chemistry.
• Continued data acquisition for new targets will enable the ongoing search for target set-dependent ACs.
• Future work will focus on systematic structure-activity relationship analysis of second-generation ACs.