Control of Synthetic Feasibility of Compounds Generated with CReM

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• De novo compound generation methods often overlook synthetic feasibility, limiting practical applications.
• Controlling the synthetic complexity of generated molecules is crucial for drug discovery pipelines.

Purpose of the Study:

• To investigate the synthetic feasibility of compounds generated using the chemically reasonable mutations (CReM) approach.
• To explore methods for indirectly controlling synthetic complexity in CReM and its impact on target scores.
• To evaluate CReM's performance against established methods on the Guacamol benchmark.

Main Methods:

• Utilized the chemically reasonable mutations (CReM) approach for de novo compound generation.
• Implemented strategies for indirect control over the synthetic complexity of generated molecules.
• Assessed compound performance on Guacamol benchmark tasks, analyzing the trade-off between complexity and target scores.

Main Results:

• A clear inverse relationship was observed between synthetic complexity and achieved target scores.
• CReM-based solutions demonstrated competitive performance compared to reference approaches.
• The study highlights the balance between synthetic accessibility and molecular property optimization.

Conclusions:

• The chemically reasonable mutations (CReM) approach offers a viable strategy for generating synthetically feasible drug candidates.
• Indirect control over synthetic complexity can be effectively integrated into de novo design workflows.
• CReM shows promise for improving the efficiency and success rate of drug discovery efforts.