AutoGrow 3.0: an improved algorithm for chemically tractable, semi-automated protein inhibitor design

Area of Science:

• Computational Chemistry
• Drug Discovery
• Bioinformatics

Background:

• Automated ligand optimization is crucial for drug discovery.
• Existing methods may lack chemical intuition and synthesizability considerations.

Purpose of the Study:

• To introduce an improved version of AutoGrow (3.0) for automated ligand optimization.
• To enhance synthesizability and druglike properties of candidate molecules.
• To integrate chemical intuition into the evolutionary drug design process.

Main Methods:

• AutoGrow 3.0, an evolutionary algorithm, was developed.
• It utilizes click chemistry rules to guide ligand optimization.
• Ligands with non-druglike properties are discarded during the process.

Main Results:

• AutoGrow 3.0 successfully generated druglike molecules with high predicted binding affinities.
• Inhibitors for three drug targets (Trypanosoma brucei RNA editing ligase 1, PPARγ, DHFR) were predicted.
• The algorithm demonstrated enhanced synthesizability and chemical feasibility.

Conclusions:

• AutoGrow 3.0 offers a valuable tool to supplement medicinal chemists' efforts in drug design.
• The enhanced algorithm improves the efficiency and success rate of identifying potential drug candidates.
• AutoGrow 3.0 represents a significant advancement in automated, chemically intuitive drug discovery.