TamGen: drug design with target-aware molecule generation through a chemical language model

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Generative drug design offers potential for novel therapeutics but often yields molecules with limited practical viability.
• Existing methods frequently overlook crucial drug-related properties, hindering downstream drug discovery success.

Purpose of the Study:

• To develop TamGen, a method for target-aware molecule generation and refinement using a GPT-like chemical language model.
• To enhance the molecular quality and viability of generated compounds for drug discovery.

Main Methods:

• Utilized a GPT-like chemical language model for generative drug design.
• Integrated TamGen into a drug discovery pipeline for target-specific compound identification.
• Employed target-aware molecule generation and compound refinement strategies.

Main Results:

• Compounds generated by TamGen exhibited improved molecular quality and viability.
• Identified 14 compounds with inhibitory activity against Tuberculosis ClpP protease.
• The lead compound demonstrated a half maximal inhibitory concentration (IC50) of 1.9 μM.

Conclusions:

• TamGen enhances the practical applicability of generative drug design.
• The method shows real-world potential for discovering effective therapeutic agents.
• Findings pave the way for future advancements in computational drug discovery.