Hierarchical Reaction Logic Enables Computational Design of Complex Peptide Syntheses

Area of Science:

• Chemical synthesis
• Computational chemistry
• Medicinal chemistry

Background:

• Computer-assisted synthesis planning traditionally relies on vast reaction databases.
• Existing algorithms struggle with complex targets like advanced peptides.
• Current methods often fail to incorporate strategic logic effectively.

Purpose of the Study:

• To develop a novel hierarchical logic for computer-assisted synthesis planning.
• To enable efficient route planning for complex peptide targets.
• To mimic human expert strategies in automated synthesis design.

Main Methods:

• Implementing hierarchical logic to constrain reaction selection at different planning stages.
• Integrating protecting-group strategies and pathway pricing.
• Developing algorithms for both solid-state and solution-phase peptide synthesis.

Main Results:

• Algorithms successfully planned complete synthesis routes for complex peptides, including vancomycin and semaglutide, within minutes.
• Designed routes mirrored strategies employed by human experts, despite no literature precedent training.
• The hierarchical approach proved effective for advanced and large peptide targets.

Conclusions:

• Hierarchical logic significantly enhances computer-assisted synthesis planning for complex molecules.
• This method offers a more efficient and expert-like approach to designing peptide synthesis routes.
• The developed strategy is applicable to various peptide synthesis modes and extension directions.