Evolution of Molecular Generation Models in Drug Discovery.

Area of Science:

• Computational Chemistry
• Medicinal Chemistry
• Drug Discovery

Background:

• Molecular generative models (MGMs) are increasingly vital in drug discovery.
• Thousands of MGMs exist, aiding in designing molecules with specific properties.
• Effective classification of MGMs is crucial for understanding their applications.

Purpose of the Study:

• To categorize common MGMs into three primary types: ligand-based molecular generation (LBMG), structure-based molecular generation (SBMG), and pharmacophore-based molecular generation (PBMG).
• To discuss the principles, evolution, and applications of these MGMs in drug discovery.
• To explore the advantages, limitations, and future research directions of MGMs.

Main Methods:

• Literature review and synthesis of existing research on MGMs.
• Classification of MGMs based on their underlying principles and application context.
• Analysis of representative examples and their success in drug discovery.

Main Results:

• MGMs are classified into LBMG, SBMG, and PBMG.
• Each category demonstrates unique principles and applications in designing molecules.
• Successful applications highlight the potential of MGMs in accelerating drug discovery.

Conclusions:

• MGMs, categorized as LBMG, SBMG, and PBMG, are powerful tools in modern drug discovery.
• Understanding their principles, advantages, and limitations is key to their effective application.
• Continued development of MGMs promises to further advance the field of drug discovery.