Ring system-based chemical graph generation for de novo molecular design

Area of Science:

• Computational Chemistry
• Cheminformatics
• Drug Discovery

Background:

• Generating novel chemical structures is essential for virtual combinatorial chemistry.
• Ensuring generated structures are both exhaustive and non-redundant is a key challenge.

Purpose of the Study:

• To develop novel algorithms for generating chemical structures by combining ring systems and atom fragments.
• To create a structure generator (Molgilla) implementing these algorithms.

Main Methods:

• Developed a three-part algorithm for canonical structure generation, treating ring systems as reduced graphs.
• Implemented a rule-based algorithm for generating diversified structures.
• Enabled accurate estimation of the number of structures without exhaustive generation.

Main Results:

• Successfully implemented the algorithms in the Molgilla structure generator.
• Demonstrated practical application by generating structures mimicking rosiglitazone's pharmacophore.
• Algorithms proved simple, flexible, and capable of accurate structure enumeration.

Conclusions:

• The developed algorithms offer an efficient and flexible approach to in silico chemical structure generation.
• Molgilla provides a powerful tool for virtual combinatorial chemistry and drug discovery.
• The algorithms' simplicity and flexibility allow for broad applicability in various structure generation programs.