SynCraft: an integrated web server for ADMET-aware retrosynthesis and molecular design

Area of Science:

• Computational chemistry
• Drug discovery
• Medicinal chemistry

Background:

• Drug candidate design requires balancing synthetic feasibility and pharmacokinetic safety.
• Current computational tools often separate retrosynthesis and ADMET evaluation, leading to inefficient workflows.

Purpose of the Study:

• To present SynCraft, a unified web server integrating retrosynthesis and real-time ADMET evaluation.
• To streamline the drug design process by automating route planning and safety assessment.

Main Methods:

• Developed a web server (SynCraft) integrating template-based retrosynthesis with a multi-view learning ADMET prediction framework.
• Automated multi-step retrosynthetic planning with simultaneous safety and drug-likeness assessment of intermediates.
• Implemented interactive visualization and analysis tools for synthetic accessibility, complexity, and drug-likeness.

Main Results:

• SynCraft successfully integrates retrosynthesis and ADMET evaluation in a single workflow.
• The server automatically identifies potentially hazardous intermediates with colour-coded warnings.
• A case study on imatinib demonstrated SynCraft's ability to detect typically overlooked genotoxic intermediates.

Conclusions:

• SynCraft offers a practical solution for safety-aware drug candidate design by unifying synthetic route planning and ADMET evaluation.
• The tool enhances efficiency and safety in drug discovery by providing real-time assessment of intermediates.
• The web server is accessible at https://syncraft.denglab.org for broader use in the scientific community.