LibINVENT: Reaction-based Generative Scaffold Decoration for in Silico Library Design

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Lead optimization relies on screening focused, core-sharing chemical libraries due to the structure-activity relationship.
• Existing in silico methods primarily focus on molecule generation, not focused library design.

Purpose of the Study:

• To introduce LibINVENT, a novel computational tool for de novo drug design.
• To enable the rapid generation of focused, core-sharing chemical libraries for lead optimization.

Main Methods:

• LibINVENT designs virtual chemical libraries by proposing compounds with a shared structural core.
• The tool maximizes desirable compound properties and allows user-specified synthetic reactions.
• It ensures library compounds are similar, possess desired properties, and are synthesizable under similar conditions.

Main Results:

• LibINVENT rapidly proposes chemical libraries tailored for lead optimization.
• The tool facilitates the design of focused libraries by incorporating user-defined synthetic constraints.
• Shared core structures ensure compound similarity and synthetic feasibility.

Conclusions:

• LibINVENT offers a flexible solution for generating virtual chemical libraries in drug discovery.
• The tool supports efficient lead optimization by creating focused, synthesizable compound libraries.
• LibINVENT is freely available, promoting broader accessibility in drug design research.